Southern Cross University ePublications@SCU Theses 2004 Disappearing dividends: the case of Thai listed firms Malinee Ronapat Southern Cross University Publication details Ronapat, M 2004, 'Disappearing dividends: the case of Thai listed firms', DBA thesis, Southern Cross University, Lismore, NSW. Copyright M Ronapat 2004 ePublications@SCU is an electronic repository administered by Southern Cross University Library. Its goal is to capture and preserve the intellectual output of Southern Cross University authors and researchers, and to increase visibility and impact through open access to researchers around the world. For further information please contact [email protected]. DBA Thesis: Malinee Ronapat DISAPPEARING DIVIDENDS: THE CASE OF THAI LISTED FIRMS MALINEE RONAPAT BBA (Finance and Banking), Assumption University MSc (International Financial Markets), University of Southampton A THESIS SUBMITTED TO PARTCIALLY FULFILL THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF BUSINESS ADMINISTRATION JUNE 2004 21 DBA Thesis: Malinee Ronapat Certificate I certify that the substance of this thesis has not been submitted for any other degree and is not currently being submitted for any degree. I also certify that to the best of my knowledge any help received in preparing this thesis, and all the sources used, have been acknowledged in this thesis. ……………………………………….. Malinee Ronapat 22 DBA Thesis: Malinee Ronapat ACKNOWLEDGEMENTS I owe a debt of gratitude to a number of people who helped, supported and encouraged me during my candidature for this degree. I am very grateful to my supervisor, Associate Professor Dr. Michael David Evans, who inspired and encouraged me. This thesis could not have been completed without the help, suggestions, guidance, patience and wisdom he has given to me whilst preparing this thesis. I also acknowledge Dr. Mark Manning for his support with statistical techniques and suggestions. The staff of the International Office and Graduate College of Management are also sincerely thanked, particularly Sue White, Heike Kerber and Rosemary Graham. Lastly, I would like to express deepest gratitude to my parents, Narong and Maria Ronapat, for providing me with the priceless opportunity of undertaking my DBA degree at Southern Cross University (Australia) and for their patience, endless support, love and caring throughout. 23 DBA Thesis: Malinee Ronapat ABSTRACT The Stock Exchange of Thailand (SET) is an important source of funds for firms and provides opportunities for investors. However, the economic boom of 1990-1996, the Asian Economic Crisis and the recession of 1997-2002 have affected the performance of firms listed at SET. The dividend policies of listed firms have also been influenced by these fluctuations in the business cycle. This study investigates the phenomenon of disappearing dividends in the developing capital market of Thailand. It adopts a similar methodology to Fama and French (2001) by classifying listed firms in line with changes in their dividend polices over the period 1990 to 2002. More specifically, the study explores the characteristics of firms which pay dividends, non-payers, former payers and firms which have never paid dividends. These characteristics include profitability, investment opportunities and firm size. The analysis uses firm characteristics for predicting the dividend policies of listed firms. Changes in firm characteristics and the propensity to pay dividends are identified in this process. The analysis suggests that firms which pay dividends tend to be large and highly profitable, although they possess low investment opportunities. The study also suggests that the characteristics of firms which paid dividends changed slightly before the crisis of 1997 and changed markedly during the crisis. However, after the crisis (1998-2002) the characteristics of firms are similar to those observed before the crisis. This result is attributed to the fact that some firms have resumed paying dividends after briefly ceasing this payment during the crisis. More importantly, when firm characteristics are held constant, the propensity to pay dividends of listed firms declined slightly before the crisis and declined strongly after the crisis. Consequently, the majority of new firms and many mature firms do not pay dividends. The findings of this study are consistent with the results of Fama and French (2001), particularly with regard to the characteristics of firms and changes in the propensity to pay dividends. However, this study extends the knowledge on the phenomenon of disappearing dividends by focussing on a developing economy, Thailand. Finally, 24 DBA Thesis: Malinee Ronapat this study suggests that investors should consider the characteristics of firms, changes in these characteristics and the propensity to pay dividends when identifying opportunities for investment. Key Words: dividends, capital market, investment 25 DBA Thesis: Malinee Ronapat GLOSSARY OF ABBREVIATIONS At ADB AEC AMEX ANOVA APT AUD BEt BOI BOT BSE CAPM DDM Et EMH EPS FTSE GDP GNP IMF IPO Lt MEt MOE MOF MUA NASDAQ NESDB NPL NYSE P/BV P/E SEC SET SIMS SP SPSS S&P’s THB TRIS UK US USD Vt Yt Total Assets Asian Development Bank Asian Economic Crisis The American Stock Exchange Analysis of Variance Arbitrage Pricing Model Australian Dollar Book Common Equity Board of Investment (Thailand) Bank of Thailand Bangkok Stock Exchange Capital Asset Pricing Model Dividend Discount Model Earnings Before Interest but After Tax Efficient Market Hypothesis Earnings per Share London Stock Exchange Gross Domestic Product Gross National Product International Monetary Fund Initial Public Offering Total Liabilities Market Value of Common Equity Ministry of Education Ministry of Finance Ministry of University Affairs National Association of Securities Dealers Automated Quotation System National Economic and Social Development Board Non-Performing Loan The New York Stock Exchange Price to Book Value Price to Earning Security Exchange Commission The Stock Exchange of Thailand The Stock Exchange of Thailand’s Information Management System Trading Suspended Statistical Package for Social Sciences Standard and Poor’s Rating Services Thai Baht Thai Rating Information Service United Kingdom of Great Britain and North Ireland United States of America American Dollar Market Value of Firm After-Tax Earnings to Common Stock 26 DBA Thesis: Malinee Ronapat TABLE OF CONTENTS Certificate .......................................................................................................................... ii Acknowledgements .......................................................................................................... iii Abstract ............................................................................................................................ iv Glossary of Abbreviations............................................................................................... vi Tables of Contents........................................................................................................... vii List of Tables .................................................................................................................... xi List of Figures................................................................................................................. xiii CHAPTER 1 – INTRODUCTION TO THE STUDY 1.1 Introduction........................................................................................................ 1 1.2 Background to the Research ............................................................................. 3 1.3 Research Issues................................................................................................... 5 1.3.1 Research Problem ................................................................................... 5 1.3.2 Research Questions................................................................................. 5 1.3.3 Research Hypotheses .............................................................................. 6 1.3.4 Research Objectives.............................................................................. 10 1.4 Justification for the Study ............................................................................... 10 1.4.1 Disappearing Dividends and Investors ............................................... 11 1.4.2 Regulators.............................................................................................. 11 1.4.3 The Research Community.................................................................... 12 1.5 Data and Methodology Overview ................................................................... 12 1.5.1 Key Variables Used in the Study ......................................................... 12 1.5.2 Research Model and Data Analysis Procedure .................................. 16 1.5.3 Testing Periods and Data Sampling .................................................... 18 1.6 Structure of the Thesis..................................................................................... 18 1.7 Conclusion ........................................................................................................ 20 CHAPTER 2 – COUNTRY REVIEW (THAILAND) 2.1 Introduction...................................................................................................... 21 2.2 Country Profile................................................................................................. 23 2.2.1 Demographic ......................................................................................... 24 2.2.2 Social Highlights ................................................................................... 27 2.2.3 Economic Highlights............................................................................. 27 2.2.4 Trading Environment in Thailand ...................................................... 30 2.3 Asian Economic Crisis..................................................................................... 30 2.3.1 Asian Economic Crisis Review ............................................................ 31 2.3.1.1 Financial Sector Weaknesses ............................................................ 31 2.3.1.2 Disequilibrium of Balance of Payment ............................................ 32 2.3.1.3 Contagion............................................................................................ 33 2.3.2 Impact of the Crisis............................................................................... 33 2.3.2.1 Financial Impact ................................................................................ 33 2.3.2.2 Political Impact .................................................................................. 35 2.3.2.3 Social Impact ...................................................................................... 35 2.4 Thai Sovereign Rating (by Fitch, Moody’s and S&P) .................................. 35 2.4.1 Thailand Sovereign Credit Rating Before the Crisis......................... 35 2.4.2 Thailand Sovereign Credit Rating During the Crisis........................ 36 27 DBA Thesis: Malinee Ronapat 2.4.3 Recent Sovereign Credit Rating .......................................................... 36 2.5 2.6 Thai Capital Market ........................................................................................ 37 2.5.1 History of the Stock Exchange of Thailand........................................ 37 2.5.2 Establishment of the Bangkok Stock Exchange................................. 38 2.5.3 Establishment of the Stock Exchange of Thailand ............................ 38 2.5.4 Roles of the Stock Exchange of Thailand ........................................... 38 2.5.5 Regulatory framework of the Stock Exchange of Thailand.............. 39 2.5.6 The Performance of SET...................................................................... 41 2.5.7 SET Vision 2003 .................................................................................... 48 2.5.8 Listing..................................................................................................... 48 2.5.9 Disclosure Procedure ............................................................................ 51 2.5.10 Material Information............................................................................ 52 2.5.11 Stock Information ................................................................................. 52 2.5.12 Information Dissemination .................................................................. 56 2.5.13 Investor Protection................................................................................ 57 2.5.14 Systems Reliability ................................................................................ 57 Conclusion ........................................................................................................ 58 CHAPTER 3 – LITERATURE REVIEW 3.1 Introduction...................................................................................................... 59 3.2 Background on Financial Investment ............................................................ 61 3.2.1 Definition and Objective of Financial Investment ............................. 61 3.3 Capital Market and Efficient Market Hypothesis ........................................ 63 3.3.1 Capital Market and Efficient Market Hypothesis ............................. 63 3.3.2 Efficient Market Hypothesis (EMH)................................................... 71 3.4 The Importance of Dividends ......................................................................... 76 3.4.1 Dividends and Stock Valuations .......................................................... 76 3.4.2 Components of Dividend Policy and Dividend Payment Procedures............................................................ 78 3.4.3 The Relevance of Dividend Policy ....................................................... 81 3.4.4 Dividend Policy and Share Price Movement ...................................... 85 3.4.5 Other Forms of Dividends.................................................................... 90 3.4.6 Factors Affecting Dividend Policy....................................................... 91 3.4.7 Types of Dividend Policies.................................................................... 93 3.4.8 Firm’s Dividend Payment Decision..................................................... 93 3.5 The Phenomenon of Disappearing Dividends in Capital Markets.............. 95 3.5.1 The Global Phenomenon of Disappearing Dividends........................ 95 3.5.2 The Disappearance of Dividends in Thailand .................................. 103 3.5.3 Thailand in the Light of Research on Capital Market .................... 105 3.6 Research Gaps, Issues and Model Development......................................... 105 3.6.1 Research Gaps and Issues .................................................................. 106 3.6.1.1 Research Gaps.................................................................................. 106 3.6.1.2 Research Issues................................................................................. 107 3.6.2 Development of the Research Model................................................. 112 3.7 Conclusion ...................................................................................................... 113 28 DBA Thesis: Malinee Ronapat CHAPTER 4 – METHODOLOGY 4.1 Introduction.................................................................................................... 114 4.2 Problem Discovery and Definition ............................................................... 116 4.3 Research Design ............................................................................................. 117 4.3.1 Purpose of the Study........................................................................... 117 4.3.2 Selecting the Type of Research Design.............................................. 118 4.3.3 Selecting the Research Technique ..................................................... 119 4.4 Data Classification and Methodology .......................................................... 121 4.4.1 Dependent and Independent Variables............................................. 122 4.4.2 Data Availability and Period of Study .............................................. 123 4.4.3 Proposed Methodology ....................................................................... 127 4.4.3.1 The Incident of Dividend Payment Among Thai Firms ............... 127 4.4.3.2 The Characteristics of Dividend Payers and Non-Payers ............. 130 4.4.3.2.1 The Effects of Profitability, Investment Opportunities and Firm Size ...................................................................... 136 4.4.3.3 Changing Propensity to Pay Dividends ......................................... 139 4.5 Plan for Data Analysis ................................................................................... 141 4.5.1 Data Processing ................................................................................... 142 4.5.2 Validity................................................................................................. 142 4.6 Conclusion ...................................................................................................... 142 CHAPTER 5 – ANALYSIS OF DATA 5.1 Introduction.................................................................................................... 144 5.2 Descriptive Statistics...................................................................................... 146 5.2.1 Descriptive Statistics Definitions ....................................................... 146 5.2.2 A Description of the Sample............................................................... 146 5.3 An Initial Analysis of Descriptive Statistics................................................. 148 5.4 Time Trends in Cash Dividends ................................................................... 148 5.4.1 Payers and Their Dividend Yield ...................................................... 153 5.4.2 Dividend Pattern ................................................................................. 158 5.5 Characteristics of the Dividend Payers........................................................ 166 5.5.1 Firm Characteristics Descriptive Statistics Approach .................... 166 5.5.1.1 Profitability....................................................................................... 166 5.5.1.2 Investment Opportunities ............................................................... 175 5.5.1.3 Size..................................................................................................... 178 5.5.2 ANOVA Test........................................................................................ 182 5.5.3 Review of Descriptive Statistic Results ............................................. 189 5.5.4 Confirmation from Logit Regressions .............................................. 191 5.5.5 Hypotheses Testing Using Logit Regression (Average Coefficient) 196 5.5.6 Hypotheses Testing Using Logit Regression (Group Coefficient) .. 199 5.6 The Propensity to Pay Dividends (Descriptive Statistics Evidence).......... 204 5.7 Changing Characteristics and Propensity to Pay Dividends (Logit Regressions)......................................................................................... 208 5.7.1 Regression Estimates .......................................................................... 208 5.8 Conclusion ...................................................................................................... 213 CHAPTER 6 – CONCLUSIONS AND IMPLICATIONS 6.1 Introduction.................................................................................................... 214 29 DBA Thesis: Malinee Ronapat 6.2 6.3 6.4 6.5 6.6 Conclusions..................................................................................................... 216 6.2.1 Conclusions – Research Hypotheses and Research Questions ....... 216 6.2.1.1 Conclusions Relating to the Chapters ............................................ 216 6.2.1.2 Conclusions Relating to the Testing the Model............................. 219 Implications .................................................................................................... 230 6.3.1 Implication for Theory ....................................................................... 230 6.3.2 Implication for Policy ......................................................................... 231 6.3.3 Implication for Investment Practice.................................................. 232 Contribution of the Study ............................................................................. 233 Limitation of the Study.................................................................................. 234 Suggestions for Further Research ................................................................ 235 REFERENCES.............................................................................................................. 237 APPENDICES ............................................................................................................... 250 30 DBA Thesis: Malinee Ronapat LIST OF TABLES Table 2.1: Table 2.2: Table 2.3: Table 2.4: Table 2.5: Table 2.6: Table 2.7: Table 2.8: Table 2.9: Table 2.10: Table 2.11: Table 2.12: Table 2.13: Table 3.1: Table 3.2: Table 3.3: Table 3.4: Table 3.5: Table 4.1: Table 4.2: Table 5.1: Table 5.2: Table 5.3: Table 5.4: Table 5.5: Table 5.6: Table 5.7: Table 5.8: Table 5.9: Table 5.10: Table 5.11: Table 5.12: Table 5.13: Table 5.14: Table 5.15: Table 5.16: Table 5.17: Table 5.18: The Share of Bank Lending to Property Sector ................................................. 32 Balance of Payment in US Dollar Billion (1985-2000)..................................... 32 Economic Growth Rates of Selected East Asian Countries 1996-1999 ............ 33 Exchange Rate (Per USD) ................................................................................. 34 Stock Market Collapsed (1997 to 1998) ............................................................ 34 Number of Companies Obtaining Securities Business Licenses ....................... 40 Selected Foreign Stock Exchange Indices ......................................................... 46 Percentage of Buying and Selling Value ........................................................... 47 Market Capitalization of Selected Foreign Stock Exchanges............................ 49 Number of Listed Companies of Selected Foreign Stock Exchanges ............... 50 Financial Statement Disclosure Deadlines......................................................... 52 Overview of Stock Exchange of Thailand ......................................................... 54 Turnover Value of Selected Foreign Stock Exchanges ..................................... 56 Key Differences between Debt and Equity Capital ........................................... 64 Characteristics of Preferred Stock and Common Stock..................................... 66 Advantages and Disadvantages of Common Shares.......................................... 67 Advantages and Disadvantages of Preference Shares ....................................... 67 Decision Rules for Investors and Analysts ........................................................ 75 Number of Samples in this Study .................................................................... 124 Comparison of Different Ratios Applied in Different Stages of this Research .................................................................................................... 138 Counts and Percent of SET Firms in Different Dividend Groups ................... 149 Number of Sample Firms in Each Industry ..................................................... 150 Number of Dividend Payers in Each Industry ................................................. 151 Percentage of Payers to Total Number of Listed Firms in Each Industry ................................................................................................... 152 Annual Dividend Yield by Range.................................................................... 155 Percentage of Payers in Two Ranges of Dividend Yield................................. 155 Average Dividend Yield, Average Weighted Dividend Yield, Average Percentage Changes in Security Prices and Percentage of Market Capitalization of Payers of Total Firms ........................................................... 157 What Happens in Year t to SET Firms That Do and Do not Pay Dividends in Year t-1.................................................................... 163 Average Values of Six Characteristic Ratios................................................... 167 Summary of the Characteristics of the Listed Firms in Thailand from Descriptive Statistics Approach............................................... 180 Percentage of Aggregate Values Accounted for by Firms Paying Dividends ......................................................................................................... 181 ANOVA F-statistic Test of Characteristics from 1991 to 2002 ...................... 183 Summary of the Payers’ Characteristics Confirmation ................................... 184 ANOVA Test for Payers, Former Payers and Never Paid Firms .................... 186 Summary of the Characteristics of Payers, Former Payers and Never Paid Firms ............................................................................................. 188 Logit Regression to Explain which Firms Pay Dividends ............................... 192 Summary of Hypothesis Testing Results (Average Coefficient)..................... 199 Logit Regression Results (Controlling Two Investment Variables)................ 200 31 DBA Thesis: Malinee Ronapat Table 5.19: Table 5.20: Table 5.21: Table 5.22: Table 6.1: Table 6.2: Table 6.3: Table 6.4: Table 6.5: Table 6.6: Logit Regression Results (Controlling One Investment Variable) .................. 201 Summary of Hypothesis Testing Summary (Group Coefficient) .................... 203 Hosmer and Lemeshow Test (Goodness of Fit) .............................................. 204 Logit Regression Estimates of the Effect of Changing Firm Characteristics and Decline in the Propensity to Pay Dividends..................... 210 Results of Hypotheses Testing Following Fama and French (2001) Method of Averaging Coefficient (Profitability)............................................. 220 Results of Hypothesis Testing Using Group Coefficient (Profitability).................................................................................................... 221 Results of Hypotheses Testing Following Fama and French (2001) Method of Averaging Coefficient (Investment Opportunities) ....................... 222 Results of Hypotheses Testing Using Group Coefficient (Investment Opportunities) .............................................................................. 222 Results of Hypotheses Testing Following Fama and French (2001) Method of Average Coefficient ....................................................................... 224 Results of Hypotheses Testing Using Group Coefficient (Size) ................................................................................................................ 224 32 DBA Thesis: Malinee Ronapat Figure 1.1: Figure 1.2: Figure 1.3: Figure 1.4: Figure 2.1: Figure 2.2: Figure 2.3: Figure 2.4: Figure 2.5: Figure 2.6: Figure 2.7: Figure 2.8: Figure 2.9: Figure 2.10: Figure 2.11: Figure 2.12: Figure 2.13: Figure 2.14: Figure 3.1: Figure 3.2: Figure 3.3: Figure 3.4: Figure 3.5: Figure 3.6: Figure 3.7: Figure 3.8: Figure 3.9: Figure 3.10: Figure 3.11: Figure 4.1: Figure 4.2: Figure 4.3: Figure 4.4: Figure 4.5: Figure 4.6: Figure 4.7: Figure 4.8: Figure 4.9: Figure 5.1: Figure 5.2: Figure 5.3: Figure 5.4: Figure 5.5: LIST OF FIGURES The Structure of Chapter One ............................................................................. 2 Hypotheses Developed........................................................................................ 9 Categories of Firms by Fama and French (2001) ............................................. 14 Structure of the Study ....................................................................................... 19 The Structure of Chapter Two .......................................................................... 22 Map of Thailand and Geographical Position .................................................... 23 Number of Population from 1997 to 2002........................................................ 24 Minimum Daily Wage in THAI Baht ............................................................... 25 Level of Education 1999................................................................................... 26 Real GDP Growth (%) from 1999 – 2003 ........................................................ 28 Inflation (%) from 1999 to 2003....................................................................... 29 Official Reserve in Billion AUD from 1999 to 2003 ....................................... 29 Prime Minimum Loan Rate (%) from 1999 to 2003......................................... 30 Current Sovereign Credit Rating of Thailand by Three Raters ........................ 37 Regulatory Framework of SET......................................................................... 39 Number of Companies Obtaining Securities Business Licenses ...................... 41 Stock Exchange of Thailand (SET)’s Performance from 1995 to 2003 ........... 45 Net Trading Value Since 1995.......................................................................... 47 The Structure of Chapter Three ........................................................................ 60 Type of Information in the Capital Markets ..................................................... 73 Removal of Mispricings.................................................................................... 75 Dividend-Retention-Financing Trade-Offs....................................................... 79 Sequence of Dividend Payment Dates.............................................................. 81 Two Ways of Raising Cash for Firm ................................................................ 83 Number of Listed Firms Which Paid Dividends............................................. 103 Percentage of Dividend Payers to Total Listed Firms .................................... 104 Research Questions and Hypotheses .............................................................. 108 Research Questions in Relation to Hypotheses .............................................. 111 Model for this Research .................................................................................. 112 The Structure of Chapter Four ........................................................................ 115 Links Between the Research Problem, Research Questions, Research Hypotheses and Methodology......................................................... 116 Alternatives for the Basic Research Method .................................................. 120 Defining Independent and Dependent Variables ............................................ 122 Process of Counts and Percentages of SET Firms in SIMS in Time t............ 127 Payers and Non-payers in Time t-1 ................................................................ 130 Average Firm Profitability, Investment Opportunities and Size .................... 135 How Logit Regressions is Used for Explaining the Dividend Policies of Firms ............................................................................................. 139 Logit Regression Estimates of the Effect of Changing Firm Characteristics and Decline in the Propensity to Pay Dividends.................... 141 The Structure of Chapter Five......................................................................... 145 Annual Average Dividend Yields and the Pattern in the Yield of The Weighted Average Dividends.................................................................. 154 Trends in the Dividend Yield 1990-2002 ....................................................... 156 The Number of SET Firms in Each Group ..................................................... 158 Percent of All SET Firms in Different Dividend Groups ............................... 159 33 DBA Thesis: Malinee Ronapat Figure 5.6: Figure 5.7: Figure 6.1: Figure 6.2: Hypotheses Revisiting .................................................................................... 195 Percentage of Payers Among Firms With (1) Positive Earnings (2) Negative Earnings (3) Earning Above Investment, and (4) Earnings Below Investment ...................................................................... 207 The Structure of Chapter Six .......................................................................... 215 Relationships of Logit Regression, ANOVA and Descriptive Statistics Findings ........................................................................................... 230 34 DBA Thesis: Malinee Ronapat TO MY PARENTS, WHO INSTALLED IN ME THE IMPORTANCE OF EDUCATION AND HARDWORK. 35 DBA Thesis: Malinee Ronapat C HAPTER 1.1 1 INTRODUCTION TO THE STUDY INTRODUCTION This chapter is an introduction to this study. The objective of this chapter is to establish foundations to the following chapters and to provide an overall picture of the study. The chapter is presented in eight sections as indicated by figure 1.1. Section 1.1 contains a general overview and the structure of this study. Section 1.2 discusses the research background. Section 1.3 introduces the research problem and other research issues including research questions and research hypotheses. Section 1.4 is a justification of the study. Section 1.5 briefly discusses the research methodology and introduces the procedure for identifying key variables, collecting data, data sampling, developing the research model and analysing the data. Chapter 4 will discuss this methodology in more detail. Section 1.6 presents the limitations of the study. Section 1.7 outlines the content each chapter of the thesis and finally section 1.8 is a conclusion to this chapter. 36 DBA Thesis: Malinee Ronapat Figure 1.1: The Structure of Chapter One 1.1 1.2 Introduction Background to the Research 1.3 Research Issues 1.3.1 Research Problem 1.3.3 Research Hypotheses 1.3.2 Research Questions 1.3.4 Research Objectives 1.4 Justification for the Study 1.4.1 Disappearing Dividends and Investors 1.4.3 The Research Community 1.5 1.5.1 1.4.2 Regulators Data and Methodology Key Variables Used in the Study 1.5.2 Research Model and Data Analysis Procedure 1.5.3 Testing Periods and Data Sampling 1.6 Structure of the Study 1.7 Conclusion Source: Developed for this research 37 DBA Thesis: Malinee Ronapat 1.2 BACKGROUND TO THE STUDY One of the basic requirements of a developed financial system is the existence of a formal capital market where investors can buy and sell securities (Melicher and Norton 2000). When firms wish to obtain finance, they have two alternatives: borrowing and thereby increasing their debt capital, or issuing stocks (preferred or common) and increasing their equity capital (Brealey and Myers 1991, 2000; Fisher and Jordan 1991; Gitman 2000; Melicher and Norton 2000; Petty et al. 2000; Gitman, Juchau and Flanagan 2002). Firms normally choose to issue stocks because they have no maturity date, and no cash outflow is associated with their redemption (Fisher and Jordan 1991; Gitman 2000; Melicher and Norton 2000; Petty et al. 2000; Brealey and Myers 1991; Gitman, Juchau and Flanagan 2002). Investors who buy a firm’s stocks are called shareholders, or stockholders and they are considered the owners of the firm (Melicher and Norton 2000; Petty et al. 2000). However, investors desire a financial return for the role they have played in the success, or failure of the firm’s operations (Petty et al. 2000). Therefore, firms normally pay dividends for common and preferred stockholders when they have positive earnings (with the recommendation by the Board of Directors). Stockholders view dividend payments as a return for holding the stocks of firms (Brealey and Myers 1991, 2000; Fisher and Jordan 1991; Gitman 2000; Melicher and Norton 2000; Petty et al. 2000; Gitman, Juchau and Flanagan 2002). However, dividends are normally taxed at a higher rate than interest payments (Brealey and Myer 1991; Fisher and Jordan 1991; Gitman 2000; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). Therefore, it is possible that borrowing is cheaper than issuing equities. Despite this fact, some firms still pay dividends and this is a puzzle (Black 1976). Fama and French (1997) and other researchers examined the dividends and debt decisions of some US firms and showed that there is a relationship between the firm’s dividend 38 DBA Thesis: Malinee Ronapat decisions and investment, profitability and size (Fama and French 1995, 1997, 1998, 1999, 2001; La Porta et al. 2000; Wetherilt and Weeken 2002) A research paper by Fama and French (2001) showed that the percentage of firms paying dividends fell from 66.5 percent in 1978 to 20.8 percent in 1999 for three US stock markets (AMEX, NYSE and NASDAQ). Fama and French (2001) attempted to explain whether the fall in the percentage of dividend payers is due to the decreasing propensity to pay dividends, or the changing characteristics of listed firms on the US stock markets (Fama and French 2001). Their methods are simplistic but meaningful. Fama and French (2001) concluded that the changing propensity to pay dividends and the characteristics of the firms are both important factors which explain the decline in the percentage of dividend payers in the US stock markets. In Thailand, the Stock Exchange of Thailand’s (SET) summary statistics indicate that the percentage of firms paying dividends fell from 84.2 percent of listed firms in 1990, to 46.4 percent in 2002 (SET 1990, 2002a). It is not known whether the decline in the percentage of payers is due to the change in the propensity to pay, or other characteristics of firms. A formal study which uses logit regression to analyse dividend patterns has not been conducted in Thailand. The researcher has observed that there is a gap in the literature on disappearing dividends and the percentage of payers in emerging capital markets such as Thailand. An analysis of this topic will provide benefits for investors, regulators, researchers, and contribute to the body of knowledge on disappearing dividends. 39 DBA Thesis: Malinee Ronapat 1.3 RESEARCH ISSUES This section describes the research problem, issues, questions, hypotheses and objectives of this study. 1.3.1 RESEARCH PROBLEM The research problem for this study is defined as: ‘How do the characteristics of publicly traded firms and their propensity to pay dividends influence the payment of dividends and the reasons for their disappearance from the stock market of Thailand?’ In this study, an apparent disappearance of dividends is explained by (1) changes in the characteristics of different dividend groups and (2) the propensity to pay dividends of firms. The study examines whether the disappearance of dividends in Thailand results from changes in firms’ characteristics or/and their propensity to pay dividends over time. Dividends are the only actual cash flow which investors in capital markets receive without selling their shares. Therefore, the disappearance of dividends will affect shareholders, investors, regulators, analysts and firms which make investment decisions. 1.3.2 RESEARCH QUESTIONS Research questions are the questions to which researchers seek answers for in a research project (Black 1999). These questions must be researchable and answerable (Ticehurst and Veal 2000). Five research questions are developed for solving the research problem of this study. These questions are defined below: 40 DBA Thesis: Malinee Ronapat Research question #1 : What is the background and performance of the Stock Exchange of Thailand and its listed firms? Research question #2 : How have researchers investigated dividends and their patterns throughout the world? Research question #3 : What is the appropriate methodology for collecting and analysing data? Research question #4 : Does the data from Thailand support the model? Research question #4.1 : What are the characteristics of dividend payers? Research question #4.2 : Do these characteristics change over time? Research question #4.3 : How has the propensity to pay dividends of firms changed over time? Research question #5 : How could a model be implemented? 1.3.3 RESEARCH HYPOTHESES After formulating a research question, one or a set of research hypotheses which relate to the research question are developed. It is rarely possible to test whether the theory holds when faced with empirical evidence (Clover and Balsley 1984). Instead, one is more likely to find that ‘a hypothesis which relates to a limited fact of the theory, will be deduced from the theory and submitted to a searching enquiry’ (Bryman and Cramer 1999, p. 3). A research hypothesis is defined as ‘a logically conjectured relationship between two or more variables expressed in the form of a testable statement’ (Cavana, Delahaye and Sekaran 2001, p. 98). By testing the hypotheses and conjectured relationships, it is expected that solutions ‘can be found to correct, or answer the research problem 41 DBA Thesis: Malinee Ronapat encountered’ (Cavana, Delahaye and Sekaran 2001, p. 98). This process is useful ‘when a researcher is seeking to prove or disprove the research questions by using statistical analysis’ (Ticehurst and Veal 1997, p. 37). The hypotheses relating to the questions of this study are described below: Step A: Sample listed firms 1991 to 1996 (before the Asian Economic Crisis) Hypothesis # 1A : Does the dividend payment action depend on the profitability of the listed firms? Hypothesis # 2B : Does the dividend payment action depend on the investment opportunities of the listed firms? Hypothesis # 3C : Does the dividend payment action depend on the size of the listed firms? Step B: Sample listed firms 1997 to 2002 (During and after the Asian Economic Crisis) Hypothesis # 1A : Does the dividend payment action depend on the profitability of the listed firms? Hypothesis # 2B : Does the dividend payment action depend on the investment opportunities of the listed firms? Hypothesis # 3C : Does the dividend payment action depend on the size of the listed firms? 42 DBA Thesis: Malinee Ronapat Step C: Sample listed firms 1991 to 2002 (Before, during and after the Asian Economic Crisis) Hypothesis # 1A : Does the dividend payment action depend on the profitability of the listed firms? Hypothesis # 2B : Does the dividend payment action depend on the investment opportunities of the listed firms? Hypothesis # 3C : Does the dividend payment action depend on the size of the listed firms? Figure 1.2 lists the research hypotheses of steps A, B and C. 43 DBA Thesis: Malinee Ronapat Figure 1.2: Hypotheses Developed CHARACTERISTICS OF DIVIDEND PAYERS BEFORE THE ASIAN ECONOMIC CRISIS (Step A) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. CHARACTERISTICS OF THE DIVIDEND PAYERS DURING AND AFTER THE ASIAN ECONOMIC CRISIS (Step B) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. CHARACTERISTICS OF THE DIVIDEND PAYERS FOR THE WHOLE PERIOD (Step C) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends for the whole period of investigation. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends the whole period of investigation. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividends the whole period of investigation. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividends the whole period of investigation. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends the whole period of investigation. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends the whole period of investigation. Source: Developed for this research 44 DBA Thesis: Malinee Ronapat 1.3.4 RESEARCH OBJECTIVES Based on the research background and research issues described earlier, this research has eight objectives: x To classify listed firms in the Thai capital market into different dividend groups according to their dividend actions in time t (dividend payers, former payers and never paid). x To determine the dividend pattern and explain the phenomenon of disappearing dividends in the Thai capital market. x To determine the characteristics and changes in these characteristics of listed firms in different dividend groups (dividend payers, former payers and never paid firms) over time. x To determine the relationship between each characteristic and the dividend payment action of the listed firms. x To determine the propensity to pay dividends of listed firms in Thailand. x To generate a formula to predict the percentage of dividend payers in the market. x To provide an early warning of the likelihood of the disappearance, or the reappearance of dividends. x To draw inferences on the importance of institutional and financial market issues for payers of dividends. 1.4 JUSTIFICATION FOR THE STUDY Research on the phenomenon of disappearing dividends is relatively recent in the US because Fama and French discussed the issue in 2001. Furthermore, Fama and French’s 45 DBA Thesis: Malinee Ronapat (2001) investigation focused on a developed country, the US. There is an absence of academic research in the Thai context. Most of the studies on dividends in the emerging markets have concentrated on dividend policy and the behavior of stock markets (La Porta et al. 2000; Aivazian, Booth and Cleary 2003). This is the first Thai study on the ‘disappearing dividends’ phenomenon. This research will provide new evidence on the phenomenon and will benefit investors and regulators. 1.4.1 DISAPPEARING DIVIDENDS AND INVESTORS Investors would be aware of trends in the payment of dividends in the Thai stock market between 1990 and 2002. They will be able to make investment selections between firms (dividends payers, former-payers and never paid firms) because they will be aware of the different characteristics and propensity to pay of these firms (Fama and French 2001). Institutional investors include Banks, Mutual funds, Pension Funds and Brokers. These institutions have portfolios with values in the millions of Thai baht. This research applies fundamental analysis which involves analysing stocks from published sources, such as financial statements. Shifting from payers to non-payers (former payers or never paid firms) or non-payers to payers, depending on the return, should be done by the institutional investors. They will be able to give accurate analysis and forecasts of expected stock prices and returns in the market, thus, reducing the level of risk in their portfolios. 1.4.2 REGULATORS Regulators need to ensure the investment climate is fair for all market participants. Regulators such as the Securities and Exchange Commission (SEC) and the Stock Exchange of Thailand (SET) should be aware of the publicly available information. SET and SEC need to be aware of changes in the payment of dividends and the regulations imposed on the payment of dividends, such as taxation. They also need to be aware of the changing characteristics and the propensity to pay dividends of the newly listed firms. They should consider announcing more fundamental analysis ratios and summary statistics for investor education and protection. 46 DBA Thesis: Malinee Ronapat 1.4.3 THE RESEARCH COMMUNITY This research aims to contribute to the body of knowledge on ‘disappearing dividends’ in the emerging stock market in Thailand, where the ‘disappearing dividends’ is a new topic with a lack of literature. This will raise the level of understanding on the phenomenon of disappearing dividends, firm characteristics, the measurement of changes in firm characteristics, the propensity to pay dividends in the developed market (the US) and the emerging market of Thailand. 1.5 DATA AND METHODOLOGY OVERVIEW This section is an introductory overview of the methodology used in this study. The research methodology will be discussed in more detail in chapter 4. The methodology was identified after exploring the empirical research papers relating to the disappearing dividend, dividend omissions, Efficient Market Hypotheses (EMH) and several capital market research papers. This section is divided into 3 subsections. Subsection 1.5.1 shows the key variables used in this research. Subsection 1.5.2 briefly summarises the research models in this study and the procedure for data analysis. Subsection 1.5.3 presents the sources of data, testing periods and data sampling. 1.5.1 KEY VARIABLES USED IN THE STUDY This research is an example of causal research as it ‘identifies the cause-and-effect relationships among variables’ (Zikmund, 2000, p. 56). This paper investigates the history of the payment of dividends (secondary data technique) of listed firms in the Thai capital market and applies summary statistics and logit regression to explore the decline in the percentage of payers of dividends. Secondary data will be used as the principal research technique of this study (Fama and MacBeth 1973; Fama and French 1988, 1995, 1997, 1998, 1999, 2000, 2001, 2002; La Porta et al. 2000; DeAngelo, DeAngelo and Skinner 2002; Aivazian, Booth and Cleary 2003). Numerical data on dividends payments are required to complete this study. This data 47 DBA Thesis: Malinee Ronapat is available in secondary form (Fama and MacBeth 1973; Fama and French 1988, 1995, 1997, 1998, 1999, 2001; La Porta et al. 2000; DeAngelo, DeAngelo and Skinner 2002; Aivazian, Booth and Cleary 2003). The research commences with a review of the literature on the phenomenon of disappearing dividends which was discussed by Fama and French (2001). Later, the discussion branches out to include relevant research papers that discuss methodologies and the findings of Fama and French (2001). The researcher’s choice of factors, variables and methodology is motivated by previous empirical findings (Fama and MacBeth 1973; Fama and French 1988, 1995, 1997, 1998, 1999, 2000, 2001; La Porta et al. 2000; DeAngelo, DeAngelo and Skinner 2002; Aivazian, Booth and Cleary 2003). The research methodology has two main components. The first component involves the use of summary statistics, or descriptive statistics such as means, frequencies and standard deviation. These statistics will be used to group listed firms (figure 1.3) and identify the phenomenon of disappearing dividends in the Thai stock market. This technique has been used in developed countries and has been suggested by a number of researchers (Fama and MacBeth 1973; Fama and French 1988, 1995, 1997, 1998, 1999, 2000, 2001; DeAngelo, DeAngelo and Skinner 2002; Aizazian, Booth and Cleary 2003). The second component is the application of logit regression which is used for classifying the characteristics of dividend payers and non-payers (former-payers and never paid firms) as well as identifying changes in the characteristics and the propensity to pay dividends of listed firms in Thailand. Logit regression has been used by many researchers including Fama and MacBeth (1973), Fama and French (1988), (1995), (1997), (1998), (1999), (2001), DeAngelo, DeAngelo and Skinner (2002), and Aizazian, Booth and Cleary (2003). As mentioned earlier, listed firms will be classified into four groups, depending on their dividend action in time t (present time). Payers are firms that pay cash dividends in time t or the present time. Non-payers are firms that do not pay cash dividends in time t, or the present time. Non-payers are divided into (1) former payers which are firms that do not pay at the present time, but paid in a previous year and (2) never payers which are firms that have never paid dividends (Fama and French 2001). 48 DBA Thesis: Malinee Ronapat Figure 1.3: Categories of Firms or Dividend Groups by Fama and French (2001) All Firms Pay Cash Dividends Do not Pay Cash Dividends Formerly Paid Have Never Paid Source: Fama and French (2001) The initial discussion of the characteristics of dividend payers focuses on the findings of the summary statistics. Payers and non-payers differ in terms of profitability, investment opportunities and size. The evidence from the summary statistics will be confirmed by conducting logit regressions (Fama and French 2001). Each characteristic will be measured by financial ratios. The financial ratios include: For Profitability, the indicators are Et/At and Yt/BEt Et/At is the ratio of aggregate earnings before interest to aggregate assets Yt/BEt is the aggregate common stock earnings over aggregate book equity Where: Et : Earnings before interest but after taxes for fiscal year t At : Assets at the end of fiscal year t Yt : After-tax earnings to common stock for fiscal year t BEt : Book common equity at the end of fiscal year t For Investment opportunities, the indicators are dAt/At and Vt/At 49 DBA Thesis: Malinee Ronapat dAt/At is the firm’s rate of growth of assets Vt/At is the ratio of the aggregate market value to the aggregate book value of assets Where: Vt : Total market value at the end of fiscal year t At : Total assets at the end of fiscal year t dAt : The change of total assets from year t-1 For Size, the indicator is At and Lt/At At is the total assets of the firm in time t Lt/At is the ratio of aggregate liability to the aggregate value of a firm’s assets Where: At : Total assets at the end of fiscal year t Lt : Total liabilities at the end of fiscal year t As analysis of data and summary statistics is expected to achieve outcomes which are consistent with the literature: x Profitabilityinvestment opportunities and size are factors in the decision to pay dividends (Fama and French 2001). x Dividend payers tend to be large and profitable firms with lower investment opportunities (Fama and French 2001). x Firms that have never paid dividends tend to be smaller and less profitable than dividend payers, but they have greater investment opportunities (Fama and French 2001). x Former dividend payers have low earnings and few investment opportunities (Fama and French 2001). 1.5.2 RESEARCH MODEL AND DATA ANALYSIS PROCEDURE Fama and French (2001) suggested that logit regression should be used to confirm the characteristics of dividend payers. Three characteristics of the firms will be analysed: 50 DBA Thesis: Malinee Ronapat Profitability, this characteristic will be confirmed with Et/At Where: Et/At is the ratio of aggregate earnings before interest to aggregate assets Investment opportunities this characteristic will be confirmed with Vt/At anddAt/At Where: dAt/At is the firm’s rate of growth of assets Vt/At is the ratio of the aggregate market value to the aggregate book value of assets The Size characteristic will be confirmed with SETt Where: SETt is the percentage of SET (Stock Exchange of Thailand) firms with the same or smaller rates of market capitalisation. This size measure is meant to simplify any effects of the growth in typical firm size through time (Fama and French 2001). The logit regression analysis will determine: x The roles of profitability, investment opportunities and size and in the decision to pay dividends (Fama and French 2001). x Larger firms are more likely to pay dividends (Fama and French 2001). x More profitable firms are more likely to pay dividends (Fama and French 2001). x Firms with more investments are less likely to pay dividends (Fama and French 2001). The analysis of the propensity to pay dividends will apply the quantitative evidence and is expected to indicate: 51 DBA Thesis: Malinee Ronapat x That dividends are less likely to be paid by firms with characteristics such as positive earnings and earnings in excess of investment (Fama and French 2001). x Firms are less likely to pay dividends, whatever their characteristics (Fama and French 2001). The effects of changing firm characteristics and the propensity to pay on the percentage of firms that pay dividends will be measured. The following methodologies will be adopted: x Estimating the probability that firms with given characteristics (profitability, investment opportunities and size) paid dividends during 1991-2002 (Fama and French 2001). x Apply the probabilities (estimated from logit regression) from the base period (1991-1996) for estimating the expected percentage of dividend payers for each year from 1991 to 2002 (Fama and French 2001). x Use the difference between the expected percentage of payers and the actual percentage to measure the change in the propensity to pay dividends. A positive difference between the expected and the actual percentage of payers implies a decline in the propensity to pay (Fama and French 2001). x Use the changes in the expected percentage of payers over time to present the changes in the characteristics of the listed firms (Fama and French 2001). 1.5.3 TESTING PERIODS AND DATA SAMPLING The annual data on dividend payments by the companies listed at SET will be used for conducting a yearly analysis of disappearing dividends between 1990 and 2002. The data is available for the period 1990 to 2002. Information is required for calculating financial ratios such as: annual dividend yields, asset, liabilities, market capitalization, earnings before interest and taxes, earnings before extra ordinary items, taxes and interest expenses between 1990 and 2002. However, the dAt/At ratio, which is the growth rate of asset will require data for time t and t-1. Consequently, the period of analysis will be shortened to 1991 to 2002. 52 DBA Thesis: Malinee Ronapat Sample observations used in this study were collected from all listed firms between 1990 and 2002. This data is available on SIMS (the Stock Exchange of Thailand Information Management System). Sample observations include every sector in the Stock Exchange of Thailand. 1.6 STRUCTURE OF THE THESIS This study is presented in 6 chapters as illustrated in figure 1.4. A brief summary of each chapter is set out below. Figure 1.4: Structure of the Study Chapter 1: Introduction to the Study Chapter 2: Country Review Chapter 3: Review of Literature Chapter 4: Research Methodology Chapter 5: Analysis of Data Chapter 6: Conclusions and Implications Source: Developed for this research Chapter 1 is an introduction to the study. This chapter contains background to the study, defines the research problem and develops a list of questions and hypotheses for answering the research question. Chapter 1 also justifies the research, its methodology, provides definitions of key terms, and discusses its limitations. 53 DBA Thesis: Malinee Ronapat Chapter 2 is a background discussion on Thailand. This chapter provides an overview of the economy of Thailand, the Asian Economic Crisis (in 1997), the Thai capital market, the Securities Exchange of Thailand (SET) and the sovereign credit rating of Thailand. Chapter 3 is a review of research papers relating to the parent disciplines of financial investment, capital markets and dividend policy. This chapter commences with background on finance, investment, risk and returns. The discussion also focuses on the components of Thailand’s capital market and the dividend policies of firms. Emphasis is placed on the issue of the phenomenon of disappearing dividends in the stock markets of developed economies. Chapter 3 concludes with a summary of the gaps in the knowledge and the limitations of the relevant papers on the topic of disappearing dividends. The discussion encourages the investigation of this phenomenon in developing countries, especially in the case of Thailand. Chapter 4 discusses the methodology used for extracting and analysing data, and testing the stated hypotheses. Definitions and measurements of key variables are provided and a model is designed for this study. This chapter concludes with a plan for the analysis of data which will be conducted in Chapter 5.. Chapter 5 discusses the findings of this study. The findings are analysed using summary statistics and the results are confirmed by ANOVA and logit regression. The findings are analysed in a pattern which is consistent with the methodology described in Chapter 4. Finally, Chapter 6 discusses the implications of the findings which were presented in Chapter 5. It also provides answers to the research question and hypotheses which were developed in Chapter 1. The chapter concludes with a discussion on the limitations of the study and identifies several areas for future research. 54 DBA Thesis: Malinee Ronapat 1.7 CONCLUSION This chapter introduced the main features of this research study, including its background, research issues, questions, justification, an overview of the methodology, definitions and its limitations. Chapter 2 will present background on Thailand and its the Stock Exchange (SET) and highlight the performance of the Thai economic and financial environments. 55 DBA Thesis: Malinee Ronapat C HAPTER 2.1 2 THAILAND: COUNTRY REVIEW INTRODUCTION This chapter provides background information on Thailand, with emphasis on its economy, geography populations, capital markets and the Asian Economic Crisis of 1997. The development of Thailand’s capital markets and their performance are also discussed to focus the direction of this study. This chapter is presented in four sections. Section 2.1 is an introduction to the chapter. Section 2.2 is a brief outline of the geography, demography and economy of Thailand. Section 2.3 discusses the causes and impacts of the Asian Economic Crisis on the economy of Thailand and its capital markets. Section 2.4 focuses on the trends in the sovereign credit rating given to Thailand by three international rating companies since 1989. Section 2.5 concentrates on Thailand’s capital markets and the Stock Exchange of Thailand (SET). In particular, this section traces the history, establishment, performance and operation of SET. Finally, section 2.6 is a brief conclusion to the chapter. 56 DBA Thesis: Malinee Ronapat Figure 2.1: The Structure of Chapter 2 2.1 2.2 Introduction Country Profile 2.2.1 Demographic 2.2.3 Economic Highlights 2.2.2 Social Highlights 2.2.4 Trading Environment 2.3 2.4 Asian Economic Crisis Thai Sovereign Rating (by Fitch, Moody’s, S&P) 2.5 Thai Capital Market 2.5.1 History of SET 2.5.8 Listing 2.5.2 Establishment of BSE 2.5.9 Disclosure Procedure 2.5.3 Establishment of SET 2.5.10 Material Information 2.5.4 Roles of SET 2.5.11 Stock Information 2.5.5 Regulatory Framework 2.5.12 Information Dissemination 2.5.6 SET Performance 2.5.13 Investor Protection 2.5.7 SET Vision 2003 2.5.14 Systems Reliability 2.6 Conclusion Source: Developed for this research 57 DBA Thesis: Malinee Ronapat 2.2 COUNTRY PROFILE Figure 2.2: Map of Thailand and Geographical Position Figure removed due to copyright restrictions Source: BOI (2004) The Kingdom of Thailand is situated in the heart of Southeast Asia (see Figure 2.2). Thailand's total area of 514,000 sq. km. includes a land area of 507,171 sq. km. and 2,230 sq. km. of water territory. The country shares borders with Myanmar (Burma), Laos, Cambodia, and Malaysia. (Pringvanich 2002). The name of the country was ‘Siam’ until 1949. On May 11, 1949, an official proclamation declared that the country would henceforth be known as ‘Thailand’ (BOI 2004). Thailand is a warm and humid tropical country. The climate is monsoonal with a pronounced rainy season which lasts from May to September and a relatively dry season from August to April. Temperatures are highest in March and April and lowest in December and January. The average temperature is 28.1o C (BOI 2004). Thailand is divided into four natural regions (1) The North; (2) The Central Plain; (3) The Northeast and (4) The South. The North is a mountainous region comprised of natural forests, ridges and deep narrow alluvial valleys. Central Thailand is a lush, fertile valley. It is the richest and most extensive rice-producing area in the country. Bangkok, the capital of 58 DBA Thesis: Malinee Ronapat Thailand, is located in this region. The Northeast region is arid, characterised by rolling terrain and undulating hills. This region often suffers from climatic extremes such as floods and droughts. The Southern region is hilly to mountainous, with thick virgin forests and rich deposits of minerals and ores. This region is a centre for the production of rubber and cultivation of tropical crops (BOI 2004). 2.2.1 DEMOGRAPHIC The population of Thailand is diverse and includes Thais, Chinese, Malays, Cambodians, Vietnamese, Indians, and others. Approximately 30 percent of the total population is under the age of 15. With a growth rate of 1.2 to 1.4 percent per year the population is projected to exceed 70 million by 2010 (BOT 2000, 2002; BOI 2004). Figure 2.3: Number of Population from 1997 to 2003 Year Population (Million) 1997 60.8 1998 61.5 1999 61.8 2000 62.4 2001 62.4 Population (Million) Million People 2002 62.5 2003 63.8 Population (M illion) 64.5 63.8 64 63.5 63 62.5 62 61.5 61 61.5 62.4 62.4 62.5 2000 2001 2002 61.8 60.8 60.5 60 59.5 59 1997 1998 1999 2003 Y ear Source: BOI (2004); BOT (2000, 2003) 59 DBA Thesis: Malinee Ronapat The size of the workforce exceeds 32.9 million and the majorities are under 30 years of age. About 800,000 people join the workforce annually. The minimum wage in Thailand is currently 162 baht per day (6 AUD at the rate of 27 Baht/1AUD) in Bangkok, and 130 to 140 baht (4.8–5.2 AUD) in the provinces. Although labour is not the cheapest in the region, Thailand's work force is among the most cost-efficient in the world (BOT 2002). Figure 2.4 below presents the minimum daily wage. Figure 2.4: Minimum Daily Wage in THAI baht Provinces Minimum Daily Wage (THB) Phuket Bangkok Chonburi Chaing Mai The rest of the country 168.00 165.00 146.00 143.00 133.00 Minimum Daily Wage (THB) 180 160 140 120 100 80 60 40 20 0 168 165 146 Phuket Minimum Daily Wage (THB) Bangkok 143 133 Chonburi Chaing Mai The rest of the country Source: BOI (2004) In recent years, the Government has made education an important priority. Indeed, education will be important in the 21st century because the development of the nation’s 60 DBA Thesis: Malinee Ronapat human resources is the highest priority of the Eighth National Economic and Social Development Plan (1997-2002) (BOT 2002; BOI 2004). Figure 2.5: Level of Education in 1999 Level of Education Number of graduates Higher than Bachelor Degree 30,719 Bachelor's Degree 238,256 Lower than Bachelor's Degree 11,193 Number of graduates Number of graduates Number of Graduates 300000 238256 250000 200000 150000 100000 50000 30719 11193 0 Higher than Bachelor Degree Bachelor's Degree Lower than Bachelor's Degree Education Level Sources: BOT (2002); BOI (2004) The average rate of the literacy rate in Thailand has risen since the plan was implemented and stood at 96 percent for the total population in 2003. At present, 97.5 percent of males and 94.6 percent of females are literate (BOI 2004). 61 DBA Thesis: Malinee Ronapat 2.2.2 SOCIAL HIGHLIGHTS Buddhism is the national religion and faith of 95 percent of the population. However, Islam, Christianity, Hinduism and other creeds are also present in Thailand (BOT 2002; BOI 2004). The official language is Thai. English is a compulsory subject in public schools and it is widely spoken and understood, in Bangkok and the major cities. Chinese and Malay are also popular (BOT 2002; BOI 2004). Thailand is governed by a constitutional monarchy and an elected parliament. Thailand was an absolute monarchy until 1932. In 1932 a revolution was staged by a small group of disaffected civil servants and military men. Since this revolution, Thailand’s monarchs have ruled under a constitution and their powers are theoretically no greater than those of European monarchs (BOI 2004). The current Prime Minister is Dr Thaksin Shinnawatra (he was elected in January 2001). The country is divided into 76 provinces, each administered by an appointed governor. The provinces are sub-divided into districts, sub-districts, tambons, and villages. The Bangkok Metropolitan area is administrated by an elected governor (BOI 2004). The Thai monarchy is revered and regarded as the central, unifying element of the nation. King Bhumibol was born on December 5, 1927 and this day in December is now the Thai National Day. He ascended the throne on June 9, 1946. 2.2.3 ECONOMIC HIGHLIGHTS More than a decade of strong sustained growth concluded in 1996 when GDP growth slowed to 5.5 percent. The performance of the economics deteriorated until August 1997 when Thailand committed to a US$17.2 billion (or AUD 26.7 billion at THB27/AUD, July 16, 2003) rescue package which was arranged by the International Monetary Fund (IMF). 62 DBA Thesis: Malinee Ronapat In mid-2000 Thailand began to repay its loans from the IMF. Under this program of recovery, Thailand enhanced its revenue streams, substantially reduced spending, implemented a range of finance sector reforms and initiated a program to privatise many of its 60 state-owned enterprises (BOI 2004). Thailand has entered the 21st century on a firm economic footing. In 1999, the economy began to rebound from the crisis of 1997 and GDP expanded by 4.2 percent. This was a marked turnaround from the 10% contraction in 1998. Steady growth continued through 2000 and 4.3 percent grow was recorded in this year (BOT 1999, 2000, 2001, 2002; BOI 2004). Thailand is not immune to fluctuations in the global economy. Weaknesses in the global economy reduced the growth of GDP to 1.8 percent in 2001. However, solid domestic consumption and rising private investment in late 2001 and 2002 helped raise GDP by 4.9 percent in 2002. The National Economic and Social Development Board (NESDB) forecast that this level of growth would be sustained throughout 2003. Indeed, many analysts believe that Thailand could be experiencing a sustained recovery for the first time since the economic crisis of 1997 (BOT 1999, 2000, 2001, 2002; BOI 2004). Figure 2.6: Real GDP Growth (%) from 1999 – 2003 Re al GDP Growth (%) Real GDP Growth (%) 6 Year 1999 4.2 4 2000 4.3 2001 1.8 2002 4.9 2003 4.5 GDP (%) 5 4.2 4.9 4.3 3 Real GDP Growth (%) 4.5 1.8 2 1 0 1999 2000 2001 Year 2002 2003 Source: BOT (1999, 2000, 2001, 2002); BOI (2004) 63 DBA Thesis: Malinee Ronapat Figure 2.7: Inflation (%) from 1999 to 2003 Inflation (%) Inflation (%) 1999 0.3 2000 1.6 2001 1.6 2002 0.7 2003 1.8 2 Inflation (%) Year 1.6 Inflation (%) 1.8 1.6 1.5 1 0.7 0.3 0.5 0 1999 2000 2001 Year 2002 2003 Source: BOT (1999, 2000, 2001, 2002); BOI (2004) Responsible macroeconomic policies have kept inflation under control since 1999. The Bank of Thailand (BOT) estimates that inflation for 2004 will be about 2 percent. Thailand’s official reserves remained strong at US$38 billion at the end of 2003 (see Figure 2.8). The BOT has pledged to keep reserves strong through 2004. Figure 2.8: Official Reserve in Billion AUD from 1999 to 2003 Year Official Reserve (Billion AUD)* 1999 54.0 2000 50.8 2001 51.3 2002 60.5 2003 56.3 O fficial Re se rve (Billion AUD) Billion AUD 62 60 58 56 54 52 50 48 46 44 Official Reserve (Billion AUD) 60.5 56.3 54 1999 50.8 51.3 2000 2001 Ye ar 2002 2003 Source: BOT (1999, 2000, 2001, 2002); BOI (2004) Note: *The spot exchange rate at July 16, 2003 at THB27/AUD 64 DBA Thesis: Malinee Ronapat Figure 2.9: Prime Minimum Loan Rate (%) from 1999 to 2003 Year Prime Minimum Loan Rate (%) * 1999 8.25 - 8.50 2000 7.50 - 8.25 2001 7.00 - 7.50 2002 6.50 - 7.00 2003 6.50 – 6.75 Source: BOT (1999, 2000, 2001, 2002); BOI (2004) Note: * Average prime minimum loan rate for five largest commercial banks The average prime minimum loan rate (2001-2003) has been below 7.5 percent since 2001, thereby ensuring that long-term investment (houses and cars) has become attractive. 2.2.4 TRADING ENVIRONMENT IN THAILAND On the external front, export growth has traditionally been the major driver of the Thai economy and has contributed to the diversification of the country's industrial structure. Manufactured exports have expanded strongly and have accounted for 80 percent of total exports in the past few years. Thailand is now a major exporter of computers and parts, textiles, gems and jewelry, electronics and automotive products, in addition to agricultural products (BOI 2004). 2.3 ASIAN ECONOMIC CRISIS The Asian Economic Crisis (AEC) affected most of South East Asia although it commenced in Thailand. Thailand’s economic problems resulted from its fixed exchange rate, the collapse of the property market, over expansion of the financial sector and a decline in the level of exports. It is generally accepted that the Asian Economic Crisis 65 DBA Thesis: Malinee Ronapat commenced on 2 July 1997 when the Bank of Thailand (BOT) announced that the baht would be floated. The decision, heavily influenced the value of the Philippine peso, Malaysian ringgit and Indonesian rupiah and the region’s financial markets collapsed (BOT 1997, 1998; Radelet and Schs 1998; Chua, Sim and Tham 1999; Rahman, Yui and Foon 2003). 2.3.1 ASIAN ECONOMIC CRISIS REVIEW Goldstein (1998) claimed there were three main causes of the Asian Economic Crisis: (1) weaknesses in the financial sector (2) dis-equilibrium in the balance of payments and (3) contagion from Thailand (Rahman, Yui and Foon 2003). 2.3.1.1 FINANCIAL SECTOR WEAKNESSES A boom in credit and property markets resulted in over excessive lending by the banks to the property sector. Table 2.1 indicates that before the crisis around 30 to 50 percent of Southeast Asian banks lent to the property sector. However, with the onset of the crisis, property prices fell sharply and the level of Non-Performing Loans (NPL) of the region’s banks rose to as high as 35 percent in some cases. Non-Performing Loans (NPL) are defined as ‘loans currently in default for three months or longer regardless of whether they are secured or unsecured’ (TRIS 2001). In addition, borrowing by the banks and corporate short-term and long-term borrowing was in foreign currency. Some of the banks and firms had too little capital to repay their loans and the financials sector collapsed (BOT 1997, 1998; Radelet and Schs 1998; Chua, Sim and Tham 1999; Rahman, Yui and Foon 2003). 66 DBA Thesis: Malinee Ronapat Table 2.1: The Share of Bank Lending to Property Sector Hong Kong Singapore Thailand Malaysia Indonesia South Korea The Philippines End of 1997 40-55 30-40 30-40 30-40 25-30 15-25 15-20 Source: Goldstein (1998); ADB (2000a); Rahman, Yui and Foon (2003, p. 496) 2.3.1.2 DISEQUILIBRIUM OF BALANCE OF PAYMENT Many of the Asian countries facing economic problems had experienced a prolonged period of disequilibrium in their balance of payments because they imported more than they exported. Large deficits in the balance of payments were an important factor leading to the exchange rate crisis (Rahman, Yui and Foon 2003). Table 2.2 shows trends in the balance of payments for seven Asian countries before and after crisis. Table 2.2: Balance of Payment in US Dollar Billion (1985-2000) Country Hong Kong Singapore Thailand Malaysia Indonesia South Korea The Philippines 1990 6.3 3.1 -7.3 -0.9 -3 -2 -2.7 1991 5.7 4.9 -7.6 -4.2 -4.3 -8.3 -1 1992 5.4 5.9 -6.3 2.2 -2.8 -3.9 -1 Pre-Crisis and Crisis 1993 1994 1995 1996 8.2 1.6 -6.1 -2.2 4.2 11.4 14.4 14.5 -6.4 -8.1 -14 -15 -3 -4.5 -8.5 -4.6 -2.1 -2.8 -6.4 -7.7 1 -3.9 -8.5 -23 -3 -3 -2 -4 1997 -6.8 15 -3 -4.8 -4.9 -8.2 -4.4 1998 2.3 17.6 14.2 9.4 4.1 40.6 1.3 Post-Crisis 1999 2000 6.4 5.3 17.2 17.2 11.4 11.6 12.5 8.4 5 4.4 25.1 6.5 5.2 3.2 Source: ADB (2000b); Lim (2000); Rahman, Yui and Foon (2003, p. 497) 67 DBA Thesis: Malinee Ronapat 2.3.1.3 CONTAGION Economists and scholars believed that the value of the currencies of Thailand’s neighbours could not be sustained if there was a large depreciation of the baht. Panic amongst domestic and foreign investors resulted in strong depreciation of the Philippine peso and Indonesia rupiah (Radelet and Schs 1998; Chua, Sim and Tham 1999; Rahman, Yui and Foon 2003) 2.3.2 IMPACT OF THE CRISIS The Asian Economic Crisis has affected most of the region’s countries financially, politically and socially. 2.3.2.1 FINANCIAL IMPACT Every country in East and South East Asia faced recession in 1998 (Lim 2000). Table 2.3 shows the rate of economic growth for six Asian countries from 1996 to 1999. Table 2.3: Economic Growth Rates of Selected East Asian Countries 1996-1999 Hong Kong Singapore Thailand Malaysia Indonesia South Korea 1996 4.5 7.5 5.9 10 7.8 6.8 Period 1997 1998 5 -5.3 8.4 0.5 -1.7 -10.2 7.3 -7.4 4.7 -13 5 -6.7 1999 3.1 5.4 4.2 5.8 0.3 10.7 Source: ADB (2000c), Rahman, Yui and Foon (2003, p. 497) The fixed exchange rate system collapsed resulting in an extreme exchange rate crisis in 68 DBA Thesis: Malinee Ronapat the region. The Thai baht depreciated by 54%, the Malaysian ringgit by 40%, the Korean won by 48% and the Indonesian rupiah declined by more than 83% (see Table 2.4). These declines in the exchange rate are measured against the United State dollar (USD). Table 2.4: Exchange Rate (per USD) Crisis Countries Singapore (S$) Thailand (Baht) Malaysia (Ringgit) Indonesia (Rupiah) South Korea (Won) Jun-97 1.4264 24.318 2.5157 2427.9 887.03 Lowest Rate (Period) 1.7566 (August 1998) 52.551 (January 1998) 4.1941 (August 1998) 13995.9 (July 1998) 1693.65 (January 1998) Depreciation (%) -19% -48% -40% -83% -48% Source: ADB (2000a); Lim (2000); Rahman, Yui and Foon (2003, p. 498) The stock markets of these economies collapsed after the depreciation of their exchange rates. Decline in the net asset value of banks and other financial institutions also resulted from the increase in Non-Performing Loans that was discussed earlier. The withdrawal of funds and investments from the region resulted in a skew of funds from USD 92.8 billion in 1996 to negative USD 12 billion in 1999 (Radelet and Sachs 1998; Rahman, Yui and Foon 2003). Table 2.5 contains the highest (expressed as 100%) and lowest points that were recorded for the stock market indices of six Asian countries during the Asian Economic Crisis. Table 2.5: Stock Market Collapsed (1997 to 1998) Crisis Countries Hong Kong (Hang Seng) Singapore (STI) Thailand (SET) Malaysia (KL Composite) Indonesia (Jarkarta SE Composite) The Philippines (Manila Composite) Highest Pre-Crisis Aug-98 Feb-97 Jan-97 Feb-97 Jul-97 Feb-97 Lowest Point 43% (September 1998) 40% (September 1998) 25% (September 1998) 35% (September 1998) 37% (October 1998) 35% (September 1998) Source: ADB (2000a); Lim (2000); Rahman, Yui and Foon (2003, p. 498) 69 DBA Thesis: Malinee Ronapat 2.3.2.2 POLITICAL IMPACT The level of political instability was relatively high during the Asian Economic Crisis and many policy responses were unsuccessful in restoring economic growth. Political instability was particularly high in Thailand under the government led by Charvalit Yongchaiyudh and in Indonesia under President Suharto and President Habibie. 2.3.2.3 SOCIAL IMPACT Conditions in the region’s labour market were also depressed during the crisis. High levels of unemployment lowered the incomes of households and the standards of living. Social externalities such as high levels of crime were also recorded. (Rahman, Yui and Foon 2003). The next section will discuss the sovereign ratings given to Thailand before, during and after the crisis. Later, more detail on the Thai capital market including its establishment, performance and expansion from 1975 to the present day will be discussed. 2.4 THAI SOVEREIGN RATING (BY FITCH, MOODY’S, S&P) 2.4.1 THAILAND SOVEREIGN CREDIT RATING BEFORE THE CRISIS Thailand was initially rated in 1989 by two well-known credit agencies namely Standard and Poor (S&P) and Moody’s. A credit rating is a measurement of a country’s risk of defaulting on its loans, compared to that of other countries. It also measures the ability of a country to repay its debts over time (BOT 2000, TRIS 2001). Thailand was initially rated, A- by S&P and A2 by Moody’s in the investment grade. These ratings changed little before the Asian crisis in late 1997. The highest rating that has been granted by S&P was A and 70 DBA Thesis: Malinee Ronapat A2 by Moody’s during the economic boom in 1995 and 1996. However, the credit upgrade in those years was not attractive for Thai society because the Government and private sectors rarely borrowed from foreign countries. Also foreign investors came to Thailand for financial and direct investment due to the economic boom in those years (TRIS 2001). 2.4.2 THAILAND SOVEREIGN CREDIT RATING DURING THE CRISIS In 1997, the two raters often downgraded Thailand’s sovereign rating in response to a downgrading by other raters. S&P downgraded Thailand’s rating three times during a five month period from A level to BBB- on 9 January 1998. Moody’s downgraded Thailand’s rating four times in nine months from A2 to Ba1 on 21 December 1997. Ba1 is considered a Speculative, or Non-investment Grade (TRIS 2001). The announcement of the credit downgradings affected business activities directly and indirectly. The Thai money market was especially affected, because investors consider the credit rating of a country when making decisions. As the credit rating was downgraded, this made investors think that Thailand had become more risky and this increased the cost of moving funds among countries (so called capital mobilisation). Consequently, total investments in Thailand decreased considerably in 1998. Total capital outflows from the private sector were 645,096 billion baht (or AUD 24 billion at the rate of 27 Baht/AUD), or AUD 12 billion higher than 1997 (TRIS 2001). The credit downgrading also affected the political climate atmosphere in Thailand. It gave the Government a weaker image and enhanced the level of political instability. However, an upgrading of the credit rating has a positive effect on the Thai economy and this is manifested in the value of the Thai baht and the number of investments at SET (TRIS 2001). 2.4.3 RECENT SOVEREIGN CREDIT RATING Three international credit raters have recently rated Thailand differently. Moody’s rated Thailand at Baa3, S&P at BBB- and Fitch, at BBB-. All these grades are speculative, or 71 DBA Thesis: Malinee Ronapat non-investment types of sovereign credit ratings. The current Sovereign Credit Rating of Thailand (2002) is shown in Figure 2.10 below: Figure 2.10: Current Sovereign Credit Rating of Thailand by Three Raters Raters Sovereign Credit Rating (December 2002) Moody's Baa3 S&P BBB- Fitch BBB- Sources: TRIS (2002) 2.5 THAI CAPITAL MARKET 2.5.1 HISTORY OF THE STOCK EXCHANGE OF THAILAND The creation of Thailand's first officially sanctioned and regulated securities market was initially proposed as part of the Second National Economic and Social Development Plan (1967-1971). The plan stressed that the market's most important role would be to mobilise funds to support Thailand's industrialisation and economic development (SET 2002a). The development of a modern capital market in Thailand can be divided into two phases. The first stage involved ‘The Bangkok Stock Exchange’ which was privately owned while the second stage involved the establishment of ‘The Securities Exchange of Thailand’ (SET 2002b). 2.5.2 ESTABLISHMENT OF THE BANGKOK STOCK EXCHANGE Thailand’s stock market was established in July 1962, when a private group established an organised stock exchange as a limited partnership. The name of this institution was 72 DBA Thesis: Malinee Ronapat changed to the ‘Bangkok Stock Exchange Co., Ltd.’ (BSE) in 1963. BSE was initially rather inactive because annual turnover consisted of only 160 million baht in 1968 and 114 million baht in 1969. Trading volumes fell sharply to 28 million baht in 1971 and BSE ceased operations in the early 1970s. It is generally agreed that BSE failed to succeed because of a lack of official government support and a limited level of investor understanding of the equity market. 2.5.3 ESTABLISHMENT OF THE STOCK EXCHANGE OF THAILAND In 1972 the Government took a step toward building a formal capital market by amending the ‘Announcement of the Executive Council No. 58 on the Control of Commercial Undertakings Affecting Public Safety and Welfare’. By 1975 the basic legislative framework was in place and on April 30 1975, ‘The Securities Exchange of Thailand’ officially started trading. On January 1 1991 its name was formally changed to ‘The Stock Exchange of Thailand’ (SET) (SET 2002c). 2.5.4 ROLES OF THE STOCK EXCHANGE OF THAILAND SET's primary roles are: x To serve as a centre for the trading of listed securities and to provide the essential systems needed to facilitate securities trading (SET 2002c). x To undertake any business relating to the Securities Exchange, such as a clearing house, securities depository center, securities registrar, or similar activities (SET 2002c). x To undertake any other business approved by the Securities Exchange Commission (SEC) (SET 2002c). 2.5.5 REGULATORY FRAMEWORK OF THE STOCK EXCHANGE OF THAILAND 73 DBA Thesis: Malinee Ronapat Figure 2.11: Regulatory Framework of SET Source: SET (2002c) SET’s Board of Governors is comprised of a maximum of eleven people. Its President is appointed by the Board and is an ex-officio member of the Board. The Board is also responsible for formulating SET’s policies and supervising the Exchange's operations. However rules and regulations prescribed by the Board must also be approved by SEC (SET 2002c). Only member companies of SET are authorised to buy, or sell securities on the Exchange. Firms which have obtained a securities licence from the Ministry of Finance (following a recommendation by the SEC) to engage in the securities business, as stock brokers, may apply for membership of SET. Membership status is obtained once approval is granted by SET’s Board of Governors. At present, SET has 28 member firms. Besides brokers, there are sub-brokers which, like brokers, take orders from investors. But sub-brokers are not members of SET and they have to route trading orders to a broker. The broker then transmits these orders to the SET trading system (SET 2002c). 74 DBA Thesis: Malinee Ronapat Table 2.6: Number of Companies Obtaining Securities Business Licenses Number of Companies 1997 1998 1999 2000 2001 2002 % increase (since 1997) Securities Brokerage 41 42 42 43 43 43 5% Securities Dealing 59 59 59 59 60 63 7% Investment Advisory Service 39 39 39 39 39 39 0% Securities Underwriting 58 58 59 60 60 63 9% Mutual Fund Management 14 14 14 14 14 14 0% Private Fund Management 19 18 17 22 24 27 42% Type of Securities Business Source: SEC (2003) Table 2.6 indicates that the number of companies in the securities brokerage business increased from 41 firms in 1997 to 43 firms in 2002, representing growth of 5 percent. The securities brokerage business includes brokerage firms and sub-brokerage firms. Securities dealing expanded by 7 percent, from 59 firms in 1997 to 63 firms in 2002. The securities dealing business includes finance and securities companies and commercial banks. However, the number of firms that offer investment advisory services has been stable since 1997. The number of securities underwriting firms increased by 9 percent from 58 in 1997 to 63 firms in 2002. The securities underwriting business includes securities companies, finance and securities companies, and commercial banks. The mutual fund management business has been stable at 14 licenced firms since 1997. Private funds management has shown the largest expansion at 42 percent. With only 19 firms in 1997, the number of private funds management firms, which includes securities companies, finance and securities companies, fund management companies, commercial banks and insurance companies, increased to 27 licensed firms in 2002. Figure 2.12 shows the number of securities brokerage firms, securities dealing firms, private fund management firms, and securities underwriting firms over the period 1997 to 2002 (SEC 2001, 2002, 2003). 75 DBA Thesis: Malinee Ronapat Figure 2.12: Number of Companies Obtained Securities Business Licenses Number of companies obtained securities business licenses Securities Brokerage Securities Dealing Securities Underwriting Private Fund M anagement 70 63 63 59 58 Number of companies 60 59 58 60 60 59 60 59 59 50 42 41 43 42 43 43 40 27 30 24 22 19 20 18 17 10 0 1997 1998 1999 2000 2001 2002 Year Source: SEC (2003) 76 Chapter 5: Analysis of Data and Findings 2.5.6 THE PERFORMANCE OF SET This section discusses the conditions in the capital market from 1990 to 2002. In 1990, the performance of the capital market was strong in the first six months and the SET index rose from 879.2 to 1,143.8. However, after the crisis in the Persian Gulf, the index declined to 544.3 at the end of 1990. The trading volume was, therefore, low in the last six months of 1990. The main factors contributing to the low level of activity in the capital market were domestic political instability and as stated above, the Persian Gulf crisis, which resulted in strong increases in oil prices and interest rates, as well as the collapse of overseas capital markets (BOT 1990; SET 1990). In 1991, the level of trade at SET recovered and closed at 711.4 in December 1991, which was 16.1 percent higher than in 1990. The volume of trading averaged 3,237 million baht per day, which is almost 30 percent higher than 1990. The volume of trade rose in the first quarter of 1991 and there was some fluctuation later in 1991. Indeed, the volume of trade in 1991 rose 25.6 percent above that recorded in 1990 and the index stood at 893 points. The factor contributing to the unstable index was the political unrest in May (referred to as ‘Black May’) which involved an overthrown of the government by the military. However, the market experienced a rapid increase in the volume of trade after September 1992. This favourable result was due to the implementation of corrective policies by government, the activities of listed companies and the impact of legal action against individuals and firms which were manipulating the market. These measures enhanced the level of confidence and stability both in the short and long term (BOT 1991, 1992; SET 1991, 1992). Trade fluctuated widely in 1993 and closed at 1682.9 (the highest point in the history) in December, up by almost 90 percent over the figure recorded in 1992. The level of fluctuation can be attributed to negative and positive factors. In early 1993, one financial institution faced problems with liquidity and ceased operating and 30 individuals were caught manipulating prices by SEC and SET in April 1993. Therefore, trade at the market suffered slightly but was boosted in May by lower interest rates and higher capital inflow (BOT 1993; SET 1993). 144 Chapter 5: Analysis of Data and Findings In 1994, the SET index fluctuated slightly and fell to 1,360.09 points, a decline of 19.2 percent. Although the performance of listed firms improved, the economy expanded strongly. Listed firms were permitted for the first time to open branches overseas, although this did not overcome negative issues such as domestic political instability and volatility in the international money market (weakening of the U.S. dollar against the Yen and Deutsche Mark in June 1994). The Mexican financial crisis and the associated outflow of capital from foreign investors resulted in a decline of 5.8 percent in the SET index in 1995. In 1996, the capital market closed at 831.57 points, which is 35.1 percent less than the 1995 close. This result can be attributed to the poor performance of some of the listed firms and political uncertainty (BOT 1995, 1996; SET 1995, 1996). In 1997, the capital market fell sharply and closed at 372.69 points, or 55.2 percent below the 1996 close. The daily volume of trade declined by almost 30 percent from the previous year. This decline resulted from the slowdown in economic growth, operational losses of the listed firms, the closure of 52 financial institutions, strong depreciation in the exchange rate and a high level of financial instability due to the Asian Economic Crisis. Many firms were delisted, or received a ‘C’ posting which signals the possibility of a delisting (BOT 1997; SET 1997). The SET index continued to decline in 1998 due to the unfavorable economic outlook, low domestic liquidity and the Asian currency crisis. However, the market improved slightly during the last four months of 1998 due to a decline in domestic interest rates, greater stability of the Thai baht, and the transparency of the Comprehensive Financial Restructuring Scheme which was introduced by SEC on 14 August. Due to the high operational losses which were reported by nearly every listed firm, SET changed the listing criteria for delisting in 1998. The previous criteria stated that every listed firm must have a specific amount of tangible assets totalling at least 50 percent of its total assets and should report net profits for 3 consecutive years. The new delisting rule used the benchmark of non-negative shareholder equity. Some firms were transferred to the 145 Chapter 5: Analysis of Data and Findings restructuring (rehabilitation) sector, rather than delisted as would be the case under old rule (BOT 1998; SET 1998). In 1999, SET closed at 481.92 points, almost 40 percent higher than the previous year. The trading volume improved due to low interest rates, low stock prices, and a return to economic growth. However, problems with non-performing loans were still unresolved and a rise in crude oil prices ensured that economic growth remained relatively slow (BOT 1999; SET 1999). In 2000, the stock market closed at 269.19 points, which was 44.1 percent below the 1999 close. The trading volume also declined by 37.2 percent. The Bank of Thailand identified the positive factors as low domestic interest rates, less inflationary pressure, progress in resolving the problems of financial institutions and a decline in the value of NPLs. The bank claimed that the negative factors still out weighed the positive factors. The negative factors included an increase in public debt, high oil prices in the world market, and the slowdown of the global economy (BOT 2000; SET 2000a, 2000b). In 2001 the index fell to 265.22 in November. It was affected by the terrorist attack in the US, high domestic public debt and higher NPLs. However, the index closed at 303.85 points in December which is slightly higher than the 2001 close (BOT 2001, 2002; SET 2001a, 2001b) In 2002, the index closed at 356.48 points, nearly 20 percent higher than the figure reached at the start of the year. However, the 2002 result was superior than the performance of other markets in the region. This outcome was due to high performance of listed firms, new and more diversified investments, high economic growth and low interest rates. However, the index was affected by war in Iraq (BOT 2001, 2002; SET 2001a, 2002a). Figure 2.13 traces the movement of the SET index from 1995 to 2002. 146 Chapter 5: Analysis of Data and Findings Figure 2.13: Stock Exchange of Thailand (SET)’s Performance from 1995 to 2003 SET index SET index 1400 1200 1281 SET index 1000 832 800 772 600 482 400 373 356 269 200 356 304 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 Year Source: SET (2002c) The SET index stood at 1,281 points in 1995, two years before the Asian economic crisis. As stated above the index started to decline in 1996 and fell to its lowest level in 2000 (269 points). However, the index has risen since 2001 and now stood at 772.15 points in 2003 although it was still well below 1995 levels (SET 2002a; SET 2003). It is important to compare the performance of Thailand’s capital market with the markets of a number of developed and developing countries. Table 2.7 compares SET with the exchange indices from selected countries between 1997 and 2002 (the ending date of the study period) (SET 2002a). 147 Chapter 5: Analysis of Data and Findings Table 2.7: Selected Foreign Stock Exchange Indices Index Exchange Country USA 1997 1998 1999 2000 2001 2002 7,908.25 9,181.43 11,497.13 10,786.85 10,021.50 8,341.63 684.61 688.99 876.97 897.75 847.61 824.38 Industrials-30 Dow Jones AMEX-Composite NASDAQComposite AMEX NASDAQ 1,570.35 2,192.69 4,069.31 2,470.52 1,950.40 1,335.51 NYSE-Composite NYSE 5,405.19 6,299.93 6,876.10 6,945.57 6,236.39 5,000.00 FTSE-100 London UK 5,135.50 5,882.60 6,930.20 6,222.46 5,217.35 3,940.36 Nikkei-225 Tokyo 15,258.74 13,842.17 18,934.34 13,785.69 10,542.62 8,578.95 Hang Seng Hong Kong Japan Hong Kong 10,722.76 10,048.56 16,962.10 15,095.53 11,397.21 9,321.29 Straits Times Singapore Singapore 1,507.65 1,392.73 2,479.58 1,926.83 1,623.60 1,341.03 All Ordinaries Sydney Australia 2,579.50 2,697.00 3,108.80 3,154.70 3,359.90 2,975.50 Weighted Price Taiwan 8,187.27 6,418.43 8,448.84 4,743.94 5,551.24 4,452.45 KLSE-Composite Taipei Kuala Lumpur Malaysia 594.44 586.13 812.33 679.64 696.09 646.32 SET Bangkok Thailand 372.69 355.81 481.92 269.19 303.85 356.48 PSE-Composite Manila Philippines 1,869.23 1,968.78 2,142.97 1,494.50 1,168.08 1,018.41 JSX-Composite Jakarta Indonesia 401.71 398.04 676.92 416.32 392.04 424.95 Source: SEC (2003) SET was at its highest in 1999 (481.92 points) and lowest in 2000 (269.19 points). Table 2.7 indicates that SET rose to 356.48 in 2002. The SET index continues to be small when compared with the US, UK, Japan, Hong Kong and Taiwan. The indices with similar values to SET include the KLSE-Composite (Malaysia) and the JSX-Composite (Indonesia) (SEC 2001, 2002a, 2003). There are different types of investors in the Thai capital market. Each type of investor invests different proportions in the capital market, and accounts for a varying percentage of the total trading value. Table 2.3 shows the percentage of buying and selling which is attributed to three types of investors namely, (1) local institutional investors, (2) local non-institutional investors, and (3) foreign investors. 148 Chapter 5: Analysis of Data and Findings Table 2.8: Percentage of Buying and Selling Value 1997 100 9.94 46.81 43.25 Percentage of Buying and Selling Value (%) Local Institutional Investors Local Non-institutional Investors (Retail) Foreign Investors 1998 100 5.64 59.74 34.62 1999 100 4.90 65.69 29.41 2000 100 5.69 62.12 32.19 2001 100 3.95 77.43 18.62 2002 100 5.34 72.09 22.57 Source: SET (2002b) Note: Percentage of buying and selling value is equal to the sum of buying value and selling value divided by the total market buying and selling value multiplied by 100. The trading value of the three groups of investors is illustrated in Figure 2.14. This figure presents the net trading value (buying and selling) of these investors between 1995 and 2001. Figure 2.14: Net Trading Value Since 1995 Local institute retail foeign Net Trading Value 80000 55437 60000 47302 Million Baht 40000 34016 30227 20000 -756 13377 3680 6963 -538 6006 0 -3239 -20000 -17056 -22453 -32984 -40000 -2872-3134 -948 -26987 -6426 -33068 -46546 -60000 1995 1996 1997 1998 1999 2000 2001 Source: SET (2002b), SEC (2003) Note: ( - ) = negative value/selling more than buying After viewing Figure 2.14 and Table 2.8 it can be seen that the net trading value of foreign investors and retail investors have mirrored each other since 1995. This may be due to the different information obtained and the investment objectives of these investors. 149 Chapter 5: Analysis of Data and Findings Furthermore, the most important investors in the Thai capital market since 1995 have been retailers and foreign investors. These two groups accounted for more than 90% of buying and selling in the market. 2.5.7 SET VISION 2003 The SET vision for year 2003 is: ‘To be one of the most attractive capital markets in Asia by providing quality products, representing the Thai economy truthfully, having effective risk management tools, and complying with international standards of enforcement and corporate governance (SET 2002b)’. Achieving the goals of SET’s vision will bring greater benefits to participants in the market. Investors will be able to reduce their investment risks and receive better protection of their rights. Listed companies will see an overall improvement in quality and will be able to mobilise funds more easily at lower costs. In addition, members will be more fairly regulated and able to generate higher business volumes (SET 2002b). 2.5.8 LISTING As a company expands its business, additional funding is usually required to support this expansion and further investment. Offering securities to the public is an alternative method of finance, with lower funding costs than borrowing from financial institutions. A public offering must be approved by SEC before the securities can be released on the primary market. To add flexibility to the capital market and create a channel to enhance the liquidity of public companies, these securities may be traded in the secondary market at SET. SET plays a vital role in helping companies to raise capital and this is beneficial not only for the company but also the country's long-term economic development. Table 2.9 shows the market capitalisation of selected foreign stock exchanges in million of US dollars from 1997 to 2002 (SET 2002b). 150 Chapter 5: Analysis of Data and Findings Table 2.9: Market Capitalization of Selected Foreign Stock Exchanges Country Millions of US Dollars Exchange 1997 1998 152,512 126,307 n.a. 86,775 60,292 44,769 NASDAQ 1,737,510 2,524,373 5,204,620 3,578,593 2,896,856 1,994,494 NYSE 8,879,631 10,271,900 11,440,767 11,442,380 11,026,518 9,015,167 UK London 2,068,246 2,297,651 2,954,816 2,576,990 2,149,501 1,785,199 Japan Tokyo 2,085,370 2,439,549 4,455,348 3,157,220 2,264,528 2,069,299 Hong Kong Hong Kong 413,323 343,567 609,090 623,400 506,073 463,055 Singapore Singapore 104,438 94,987 192,983 152,830 115,688 99,807 Australia Sydney 295,766 328,929 427,835 372,790 375,131 380,087 Taiwan 287,813 260,015 375,991 247,600 292,621 261,211 Malaysia Taipei Kuala Lumpur 93,230 95,562 139,908 113,160 118,981 125,778 Thailand Bangkok 22,886 34,252 57,177 29,220 35,943 45,504 Philippines Manila 31,261 35,297 48,080 25,980 21,245 18,507 Indonesia Jakarta 29,050 22,078 64,045 26,810 22,998 30,067 USA AMEX 1999 2000 2001 2002 Source: SEC (2002a) Capitalisation of the Thai market was at its lowest in 1997 (22,886 million USD), during the Asian economic crisis. The highest level of market capitalisation for Thailand occurred in 1999 (57,177 million USD). As stated earlier, the value of market capitalisation fell to 29,220 million USD in 2001, but recovered to 45,504 USD in 2002. It appears that a period of sustained expansion of Thailand’s capital market is finally occurring after the Asian economic crisis. However, the rate of capitalisation in the Thai market remains small when compared with the US, UK and Japan. The markets of the Philippines and Indonesia possess a similar rate of capitalisation as Thailand (SET 2002b). SET has placed emphasis on the quality of its listed companies by improving the standards of its listing rules and regulations. A listed company must comply with SET's listing requirements prior to obtaining listing status (SET 2002b). Listed companies must publicly disclose all materials that are important for investors' decision making. SET has implemented a full disclosure policy, which permits investors to obtain accurate, 151 Chapter 5: Analysis of Data and Findings adequate and timely information. This helps to ensure market transparency and integrity. Disclosure occurs by facsimile and on-line through the SET Information System (SET 2002b). Table 2.10 presents the number of listed companies on selected foreign stock exchanges from 1997 to 2002. The number of listed companies in Thailand reached its highest in 1997 and fell to its lowest in 2000. The most recent number of companies listed on Thailand’s exchange is 398 firms and this is still below the number attained in 1997 (SET 2002b). Table 2.10: Number of Listed Companies of Selected Foreign Stock Exchanges Exchange Number of Companies Country 1997 1998 1999 2000 2001 2002 771 711 n.a. 693 605 572 NASDAQ 5,487 5,010 4,829 4,726 4,128 3,649 NYSE 2,626 2,669 2,592 2,468 2,400 2,366 AMEX USA London UK 2,991 2,920 2,791 2,929 2,891 2,824 Tokyo Japan 1,865 1,890 1,932 2,096 2,141 2,153 Hong Kong Hong Kong 658 680 701 790 867 978 Singapore Singapore 294 295 317 377 386 385 Sydney Australia 1,219 1,222 1,287 1,409 1,410 1,421 Taipei Taiwan 404 437 462 531 586 640 Kuala Lumpur Malaysia 703 731 752 790 807 860 Bangkok Thailand 431 418 392 381 382 398 Manila Philippines 221 221 226 230 232 235 Jakarta Indonesia 282 287 276 286 315 330 Source: SEC (2002a) Table 2.10 indicates that the number of companies listed on the Thai stock market since 1997 has fallen. This phenomenon is also evident in the US and UK. However, stock markets, such as Australia, Hong Kong, and Japan, have continued to increase since 1997 (see Table 2.10). The fall in the number of listed companies in Thailand, the US and UK may be due to the bursting of the Asian investment bubble and the hi-tech boom (SEC 2002a). 152 Chapter 5: Analysis of Data and Findings 2.5.9 DISCLOSURE PROCEDURE To ensure equal access in the SET, any disclosure of material information must be made by the listed company at least one hour prior to the commencement of each trading session or after the close of the day's trading. There are two trading sessions, the morning trading session 10:00 a.m.-12:30 p.m. and the afternoon trading session 2:30-5:00 p.m.. Therefore, listed companies must submit their information: 1. before 9.00 a.m. for the morning trading session (SET 2002a) 2. between 12:30 and 1:30 p.m. for the afternoon trading session or (SET 2002a) 3. anytime from after 5:00 p.m. and before 9:00 a.m. on the morning of the next day of trading (SET 2002a). To facilitate instantaneous disclosure, SET has established guidelines for determining the of information which requires prompt disclosure as follows: 1. Information which is likely to have a significant effect on the market price of the firm's securities (SET 2002a). 2. Information which is likely to be important for investment decisions (SET 2002a). 3. Information which is likely to affect the interests of shareholders (SET 2002a). 2.5.10 MATERIAL INFORMATION Vital information, which may affect the stock prices, investment decisions, or investor interests, must be disclosed within a specific period. This information includes the company’s report and financial statements (audited and unaudited). Vital information is important for both technical and fundamental analyses (this will be explained later) especially ratio and stock valuation analysis (SET 2002b). This information should be presented on time to SET and SEC. 153 Chapter 5: Analysis of Data and Findings Table 2.11: Financial Statement Disclosure Deadlines 1. Financial Statements Unreviewed and unaudited financial statements Reviewed quarterly financial statements Audited annual or semiannual financial statements 2. Annual Report 3. Disclosure report of additional information 4. Information on any operational and financial results which is price-sensitive and/or affects shareholder interests Within 30 days of the end of the accounting period. Within 45 days of the end of each quarter. Within 3 months of the end of the accounting period. Within 110 days of the end of the accounting period. Within 3 months of the end of the accounting period. At least one hour prior to the trading session. Sources: SET (2002b) 2.5.11 STOCK INFORMATION The Department of Investors Relations and the Stock Exchange of Thailand (2003) advise investors to analyse stock information before making any investment decision in the stock market. x Price: The stock price is an important factor for investors, as buying and selling affects the level of supply and demand of a particular stock and its price. Changes in stock price reflect the amount of money involved and the investor’s decision on whether to buy, sell, or hold a particularly stock. When valuing a stock the price has to be considered, together with other performance variables such as earnings per share and dividends (SET 2002b). x Price-Earnings Ratio (P/E Ratio): This ratio is a measure of the stock's fundamental value. P/E Ratio is calculated by dividing the Close Price (P) with Earnings per share (E). P/E = Price / Earning per share A stock with a low P/E ratio is preferable to one with a high P/E. In brief, a low P/E stock has higher earnings potential, or is cheaper than a high P/E stock, when its earnings potential is considered (SET 2002b). 154 Chapter 5: Analysis of Data and Findings x Dividends: Dividends as described by SEC, refer to the proportion of a firm’s profits which are distributed to shareholders (common, preferred or investment unit) and depend on the right of the holders. A preferred share’s dividend is normally fixed as a percentage of par, while a common share’s dividend (or investment unit) depends on the firm’s success over a fiscal year. The Board of Directors of a firm announce the dividend payment to common shareholders, yearly, semi-annually or quarterly. The dividend to common shareholders can be a cash dividend, or stock dividend (both concepts will be discussed later). However, SEC stated that the stock dividend in Thailand is not as popular as the cash dividend and firms do not want to pay stock dividends because of the tax burden. SEC is now considering legislative changes in relation to stock dividends (SEC 2002a). x Dividend Yield: The rate of dividend return is shown as a percentage. A stock with a high dividend yield is more attractive because a higher rate of return is received in the form of dividends. Dividend Yield = (Dividend * 100) / Earning per share x Trading Volume: The trading volume, or liquidity of a stock is important because a high trading volume indicate that investors can buy/sell a stock easily as there are many buyers and sellers (SET 2002b). x Financial Analysis: This refers to an analysis of a company's growth potential, stability, financial and management strengths and profit potential for its investors. Financial analysis is a complex exercise which requires strong investment knowledge (SET 2002b). In addition to market conditions and factors specific to individual stocks, a myriad of variables influence the stock market and the movement of prices. Investors can follow these issues in various media reports. Table 2.12 presents an overview of the important indicators in the Stock Exchange of Thailand (SET) from 1997 to 2002 (SET 2002b). 155 Chapter 5: Analysis of Data and Findings Table 2.12: Overview of Stock Exchange of Thailand 1997 372.69 1998 355.81 1999 481.92 2000 269.19 2001 303.85 2002 356.48 929,598 3,764 1,133,344 855,169 3,505 1,268,199 1,609,787 6,571 2,193,067 923,697 3,740 1,279,224 1,577,758 6,440 1,607,310 2,047,442 8,357 1,986,236 Number of Listed Companies 431 418 392 381 382 389 Number of Listed Securities P/E Ratio (times) P/BV Ratio (times) Dividend Yield (%) 529 6.59 0.89 6.04 494 10.04 1.05 1.34 450 14.70 1.72 0.61 438 5.52 1.11 1.78 449 4.92 1.29 2.06 471 6.98 1.36 2.72 SET Index Total Turnover (B mil.) Daily Average Turnover (B mil.) Market Capitalization (B mil.) Source: SET (2002b) As stated earlier, the performance of the Stock Exchange of Thailand (SET) has fluctuated since 1997. Total turnover and daily average turnover have increased since 1997, while the number of listed companies and securities has declined since 1997 (SET 2002b). The Table 2.12 presents trends in a range of SET’s performance indicators since 1997, including the P/E and P/BV ratios and the Dividend Yield. As stated earlier, P/E is the price to earning ratio, P/BV is the price to book value ratio and the dividend yield is the percentage of dividends per share to earnings per share. The price-book value ratio sets the current share quote in relation to its book value per unit as stated by the balance sheet. The higher this Figure, the better are a stock’s investment potential according to market participants (SET 2002b). The P/E ratio was at its highest in 1995. Consequently, it appears that two years before the Asian economic crisis investors believed that Thai stocks were over-valued (SET 2002b). The P/E ratio was 6.59 in 1997 and rose to 14.7 in 1999. However, the P/E ratio fell to 5.52 in 2002 and to 4.92 in 2001. A stock with a high Dividend Yield is more attractive to some investors because they receive a higher rate of return in the form of dividends. Dividend yields were at their 156 Chapter 5: Analysis of Data and Findings highest in 1995. Consequently, listed firms paid higher dividends before the Asian economic crisis. The dividend yield reached its lowest level (0.89) in 1997. Table 2.13 presents the turnover value of selected foreign stock exchanges in millions of US dollars from 1997 to 2002 (SET 2002b). Table 2.13: Turnover Value of Selected Foreign Stock Exchanges Country Millions of US Dollars Exchange 1997 1998 1999 2000 2001 2002 143,230 287,929 n.a. 945,391 817,042 642,181 NASDAQ 4,481,682 5,518,946 10,466,613 19,798,799 11,000,223 7,533,875 NYSE 5,777,602 7,317,949 8,945,205 11,060,041 10,489,324 10,311,156 UK London 1,989,489 2,887,990 3,399,349 4,558,661 4,550,503 4,006,732 Japan Tokyo 896,055 750,831 1,675,641 2,315,502 1,544,662 1,565,824 Hong Kong Hong Kong 453,657 206,153 230,032 376,664 241,011 194,425 Singapore Singapore 74,138 58,510 107,406 95,152 71,771 62,770 Australia Sydney 171,004 163,054 198,195 226,485 244,099 295,649 Taiwan 1,308,634 895,986 913,610 986,269 544,597 632,666 Malaysia Taipei Kuala Lumpur 145,688 26,840 42,431 52,868 21,323 32,038 Thailand Bangkok 28,796 20,976 37,246 21,119 31,034 41,292 Philippines Manila 20,349 10,148 19,716 8,526 3,148 3,107 Indonesia Jakarta 42,605 10,637 20,033 15,109 9,527 13,050 USA AMEX Source: SEC (2003) As stated earlier, the turnover value of the Thai capital market fluctuated over the 6 years between 1997 and 2002. It was at its lowest in 1998, (20,976 million USD). The highest market capitalisation turnover for Thailand occurred in 2002 (41,292 million USD). However, the turnover value of the Thai market is small compared to that of the US, UK, Japan, Hong Kong and Taiwan. However, the level of turnover in the capital markets of the Philippines, Indonesia and Thailand were similar between 1997 and 2002 (SET 2002b). 157 Chapter 5: Analysis of Data and Findings 2.5.12 INFORMATION DISSEMINATION Listed company and trading information is distributed through various channels and they are very important sources of information for investors and aid in enhancing the level of accessibility for investors. The three most important sources of information are as follows: SETinfo: All data and information are updated in the SET database (SIMS), and disseminated via an electronic system known as, SETINFO. A comprehensive range of information alternatives is provided on SETinfo (SET 2001b). Intranet: Investors can access the SET website: http://www.set.or.th to obtain comprehensive information (SET 2001b). Mass media and publications: Information on SET’s members and listed companies, while its educational programs on the stock market are also available on the media and in its publications (SET 2001b). 2.5.13 INVESTOR PROTECTION The Stock Exchange of Thailand recognises the crucial role investor protection plays in enhancing investor confidence and contributing to market growth. Such protection is best understood as a combination of different, but closely integrated measures including, but not limited to, market regulation and enforcement, trading and settlement system reliability, information disclosure and equal accessibility. 2.5.14 SYSTEMS RELIABILITY 158 Chapter 5: Analysis of Data and Findings The SET trading system has been fully-computerised since 1991, thereby expanding the capacity for trade and improving efficiency and transparency. The trading system has the capacity to handle 600,000 orders daily (SET 2001b). 2.6 CONCLUSION This chapter has provided a brief background on Thailand, its economy, sovereign credit ratings and capital markets. It has, therefore, provided a political, economic and institutional setting for this study. In next Chapter reviews the literature on dividends and the dividend disappearance phenomenon. 159 Chapter 5: Analysis of Data and Findings C HAPTER 3.1 3 REVIEW OF LITERATURE INTRODUCTION Chapter 2 discussed the development and role of Thailand’s capital market. This chapter reviews the literature on the issue of disappearing dividends, firm’s characteristics, ratio analysis, efficient market hypothesis (EMH), and value and growth stocks. It also identifies gaps in the body of knowledge and aids in developing an effective model for explaining the phenomenon of disappearing dividends. The chapter is presented in seven sections as shown by Figure 3.1. Section 3.2 introduces the topic of financial investment, its definitions and the concepts of finance and investment. Section 3.3 introduces the capital market and efficient market hypothesis (EMH). Section 3.4 discusses the importance of dividends, the dividend payment procedure, the components of dividends, share price movements, the diverse types of dividends, factors affecting dividend policy, types of dividend policies, the firm’s dividend payment decision and company analysis. Section 3.5 integrates the topics of capital markets and dividends by reviewing the global experiences on the phenomenon of disappearing dividends. The discussion also focuses on recent experiences in Thailand. Section 3.6 identifies gaps in earlier research and develops a model for this study. Finally, section 3.7 is a brief conclusion to this chapter. 160 Chapter 5: Analysis of Data and Findings Figure 3.1: The Structure of Chapter Three 3.1 Introduction 3.2 Background to Financial Investment 3.3 Capital Market and Efficient Market Hypothesis 3.3.1 Capital Market, Risk and Return 3.3.2 Efficient Market Hypothesis 3.4 The Importance of Dividends 3.4.1 Dividend Payment Procedure 3.4.2 Components of Dividend Policy 3.4.3 The Relevance of Dividend Policy 3.4.4 Dividend Policy and Share Price Movement 3.4.5 Other Forms of Dividends 3.4.7 Types of Dividend Policies 3.4.6 Factors Affecting Payment Decisions 3.4.9 Company Analysis 3.4.8 Firms Dividend Payment Decisions 3.5 Disappearing Dividends in Capital Market 3.5.1 Disappearing Dividends 3.5.2 Disappearing Dividends in Thailand 3.5.3 Asian Stock Markets 3.5.4 Thai Capital Market Research 3.6 Research Gaps and Model Development 3.6.1 Research Gaps and Issues 3.6.2 Research Model Development 3.7 Conclusion Source: Developed for this research. 161 Chapter 5: Analysis of Data and Findings 3.2 BACKGROUND ON FINANCIAL INVESTMENT This study will investigate the phenomenon of disappearing dividends and relate this issue to the characteristics of listed firms and their propensity to pay dividends. Changes in the characteristics of the listed firms and their propensity to pay dividends will be discussed in relation to Thailand. This section commences with a discussion of definitions and the concepts of finance, assets and markets. The final subsection links the capital market with the efficient market hypothesis. 3.2.1 DEFINITION AND OBJECTIVE OF FINANCIAL INVESTMENT Finance is ‘the study of how individuals, institutions, and businesses acquire, spend, and manage their financial resources’ (Melicher and Norton 2000, p. 3). Finance has its origins in economics and accounting. Economists apply the supply-and-demand framework, while accountants use a record-keeping mechanism to record revenues, expenses and profitability (Case and Fair 1996; Melicher and Norton 2000). The interaction, or mixture of intermediaries, markets, instruments, policy makers, and regulations which aid the flow of funds from savings to investments is called the financial system (Melicher and Norton 2000). There are three basic requirements for a modern financial system (Melicher and Norton 2000). 1. An efficient monetary system for creating and transferring money. The financial system is an efficient medium for exchanging and paying for goods and services and measuring prices (Melicher and Norton 2000). 2. Capital formation is the channelling of savings into investment. It takes place whenever resources are used to produce buildings and equipment for the production of goods for consumers, or producers (Melicher and Norton 2000). 3. Financial markets are where people buy and sell, or transfer measures of (or claims to) wealth, such as real estate, or financial assets (stocks, bonds). They encourage investment by providing the means for savers to quickly and easily 162 Chapter 5: Analysis of Data and Findings convert their claims into cash when needed (Reilly and Norton 1999; Melicher and Norton 2000). This study will focus on financial markets because they are important for modern business. Financial markets help corporations raise money by selling different types of financial assets with different maturities (Brealey and Myers 1991, 2000). There are two types of assets, namely: (1) Real assets which include the ownership of land, buildings, machinery, inventory and even precious metals; and (2) Financial assets, which are claims against the income and assets (wealth) of those who issued them. Examples of financial assets include securities, bonds and options (Brealey and Myers 1991, 2000; Reilly and Norton 1999; Melicher and Norton 2000). Liquidity is one of many properties that distinguish real assets from financial assets. Liquidity refers to the ease of converting an asset into money, with minimal costs of exchange (Fisher and Jordan 1991). Real assets are generally less liquid than financial assets because they are adapted to a specific use and yield benefits only in limited circumstances (Fisher and Jordan 1991). Furthermore, returns on real assets are difficult to measure and the markets for these assets are not broad, ready and active (Fisher and Jordan 1991; Haugen 1995; Melicher and Norton 2000). This study will concentrate on investment in financial assets. There are two key financial markets (Brealey and Myers 1991, 2000; Gitman, Juchau and Flanagan 2002) which are classified by maturities (either short-term or long-term). x Money markets are markets where financial assets, such as debt instruments of 1 year, or less, (hereafter called short-term securities) are traded. (Brealey and Myers 1991, 2000; Madura 2000; Gitman, Juchau and Flanagan 2002). x Capital markets are markets for financial assets such as debt securities (notes and bonds) and debt instruments (mortgages and stocks) which mature after a period of time which is greater than one year (hereafter called long-term securities) (Brealey and Myers 1991, 2000; Gitman 2000; Gitman, Juchau and Flanagan 2002). 163 Chapter 5: Analysis of Data and Findings Many financial assets do not require the existence of a formal financial market. Only securities and debt instruments are marketed and traded in financial markets (Melicher and Norton 2000). This study will focus on capital markets because they help firms raise long-term funds by issuing securities to the market (so called flotation or public offering). According to Fisher and Jordan (1991, p. 3), ‘capital markets help firms grow, jobs are provided, families prosper, and the economy grows’. 3.3 CAPITAL MARKET AND EFFICIENT MARKET HYPOTHESIS This section commences with a definition of the term ‘capital’ (debt and equity capital). Capital markets, the place where different types of assets are bought and sold will then be outlined. A variety of types of investments will be identified and the concepts of risk and return will be defined. In the latter stage of this section, the efficient market hypothesis (EMH) is discussed to identify the variety in the types of efficient markets, as defined by Fama (1970). 3.3.1 CAPITAL MARKET, INVESTMENT, RISK AND RETURN The term capital refers to the ‘long-term funds of the firm’ (Gitman, Juchau and Flanagan 2002, p. 37). Therefore, when analysing a firm’s balance sheet, all the liabilities, or financing items, excluding short-term liabilities, are classified as longterm capital (Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). Long-term capital consists of debt capital and equity capital. Debt capital ‘includes all long-term borrowing incurred by the firm, including bonds’ (Gitman, Juchau and Flanagan 2002, p. 37). Equity capital ‘consists of long-term funds provided by the 164 Chapter 5: Analysis of Data and Findings firm’s owners, the shareholders’ (Gitman, Juchau and Flanagan 2002, p. 37). There are two firms of equity capital, namely, internal and external raised capital. Capital is raised internally by spending a firm’s retained earnings, while capital is raised externally by selling ordinary (common), or preferred shares to the public (Gitman, Juchau and Flanagan 2002). The key differences between debt and equity capital are summarised in Table 3.1. Table 3.1: Key Differences Between Debt and Equity Capital Type of capital Debt No Voice of Management Senior to equity Claims on Income and Assets Stated Maturity Interest deduction Tax Treatment Source: Gitman, Juchau and Flanagan (2002), p. 37 Equity Yes Subordinate to debt None No deduction Equity capital is necessary for a firm’s growth because all firms must initially be financed with equity capital. There are four key differences between debt and equity capital (see Table 3.1), which ensure that equity capital is the preferred source of finance of a firm’s managers and more interesting as a topic of study (Reilly and Brown 2000; Gitman, Juchau and Flanagan 2002). The key differences are as follows: Voice of Management (or voting rights): Holders of equity capital (shareholders) are owners of the firm. They have voting rights which determine the make up of a firm’s board of directors and other special issues, while holders of debt capital (lenders) do not have these rights (Brealey and Myers 1991, 2000; Melicher and Norton 2000; Gitman 2000; Gitman, Juchau and Flanagan 2002). Claims on Income and Assets: Shareholders have claims on the income and assets of a firm after interest has been paid to creditors or lenders. The remaining funds, or residual amount can be distributed as dividends and/or retained by the firm. The shareholders’ claims on a firm’s assets are repaid if the firm fails and its assets are sold on the market. However, the claims of shareholders are of lesser value than those of a 165 Chapter 5: Analysis of Data and Findings firm’s creditors. This is because shareholders expect greater returns from dividends and/or growth in share (stock) prices (Brealey and Myers 1991, 2000; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). Maturity: equity capital is a perpetual form of financing because it does not mature, and no repayment is required. Debt capital, however, has a maturity date (Brealey and Myers 1991, 2000; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). Tax Treatment: interest payments from debt capital is tax-deductible, while dividend payments from equity capital are taxable (Brealey and Myers 1991, 2000; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). In summary, equity capital does not mature or involve repayments and its holders have voting rights. However, dividend payments for equity capital are not tax-deductible. The level of dividends also depend on a firm’s annual profits and debt capital is repaid before equity capital if a firm fails (Brealey and Myers 1991, 2000; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). Firms can raise equity capital externally by issuing ordinary or preferred shares (Brealey and Myers 1991, 2000; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). (1) Ordinary Shares Holders of ordinary shares are owners of a corporation which stock share in the company’s success and problems (Reilly and Brown 2000). Common shareholders can vote to select a firm’s board of directors, and help decide the outcome of major issues facing the firm, such as changes to its corporate charter and potential mergers with other companies (Melicher and Norton 2000; Christensen and Jordan 2001; Gitman, Juchau and Flanagan 2002). Common shareholders have a claim on a firm’s remaining profits after the holders of all other classes of debt and equity securities have received their interest payments or returns (Melicher and Norton 2000; Christensen and Jordan 2001; Gitman, Juchau and Flanagan 2002). In the event of liquidation creditors, 166 Chapter 5: Analysis of Data and Findings bondholders, and preferred stock holders must be paid in full before common stockholders receive any proceeds from the firm (Melicher and Norton 2000; Christensen and Jordan 2001; Gitman, Juchau and Flanagan 2002). However, owners of a firm’s common stock are entitled to receive dividend payments. Common stock can be divided into two groups, Class A which does not have voting rights and Class B which has voting rights. Par Value is used mainly for accounting purposes and some legal needs. It has little relationship to the current price of the shares (Melicher and Norton 2000). (2) Preferred Shares A preferred share is an equity security with a senior claim on a firm’s earnings and assets (Melicher and Norton 2000). In the event of liquidation, the claims of preferred shareholders are satisfied before common shareholders receive any proceeds from the firm (Melicher and Norton 2000; Christensen and Jordan 2001; Gitman, Juchau and Flanagan 2002). Preferred shareholders receive fixed dividends, which are paid before common shareholders’ receive their dividends (Melicher and Norton 2000). This dividend is specified as a percentage of par value, or a fixed amount per year (Melicher and Norton 2000). Unlike common stock, the par value of preferred stock is important, because dividends are expressed as a percentage of par and the par value represents the holder’s claim on corporate assets in the event of liquidation (Gitman, Juchau and Flanagan 2002; Christensen and Jordan 2001; Melicher and Norton 2000). Table 3.2 presents the variety of claims and characteristics of preferred and common stocks. Table 3.2: Characteristics of Preferred Stock and Common Stock Voting Priority for Priority claim Higher Higher right dividend in bankruptcy risk Expected return D D D Common stock D Preferred stock D Source: Developed from this research 167 Chapter 5: Analysis of Data and Findings Table 3.3 indicates that there are a number of advantages and disadvantages of raising new funds by issuing common shares, debt and preference shares (Petty et al. 2000). Table 3.3: Advantages and Disadvantages of Common Shares Advantages to the firm Disadvantages to the firm (1) The firm is not legally obligated to pay dividends. Thus, during financial distress, there is no need to have cash outflow from dividends, while there must be with debt and preference shares. (2) Common shares have no maturity date, thus, no cash outflow associated with their redemption. (3) By issuing common shares, the firm is increasing its financial base and its future borrowing capacity. By issuing debt, the firm is cutting its ability to borrow. (1) Issuing debt is tax-deductible while dividends on common shares are not. (2) The issue costs on new equity tend to be larger than those on new debt (3) Common shares are issued with voting rights. Therefore, issuing new common shares may result in a change in the ownership and control of the firm. Source: Petty et al. (2000) Table 3.4 indicates that there are advantages and disadvantages of issuing preference shares over debt and common shares (Petty et al 2000). Table 3.4: Advantages and Disadvantages of Preference Shares Advantages to the firm Disadvantages to the firm (1) Preferred shares do not have default risk to the issuer arising from the non-payment of dividends. (1) Debt interest is tax deductible while preferred shares dividend is not, thus, increasing cost of borrowing for the firm. (2) Too much issuance of preferred shares can lead to the omission of dividend by the firm. (2) Preferred shares carry dividends limited to a fixed amount. They do not participate in excess earnings like common shares. (3) No dilution of the ownership control. (4) Preferred shares get higher dividend compared to debtholders but lower than common shareholders. (5) Preferred shares increase the firm's equity base and enable the firm to borrow more in the future. Source: Petty et al. (2000) 168 Chapter 5: Analysis of Data and Findings Capital equities are traded in capital markets, or securities markets (Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). Securities markets are classified by the types of securities bought and sold through them (Fisher and Jordon 1991). The broadest classification depends on whether the securities are new issues, or are already owned by someone (Fisher and Jordon 1991; Melicher and Norton 2000). Securities markets can be classified into (1) a primary market and (2) a secondary market according to the types of securities which are traded (Fisher and Jordan 1991; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). (1) Primary Market The Primary market is the market for newly issued securities. When firms wish to raise capital the initial sale of newly issued debt, or equity securities is called a Flotation (Melicher and Norton 2000). The initial sale of equities to the public is called an Initial Public Offering (IPO) (Melicher and Norton 2000). To raise money, firms often use the services of investment bankers. The main activity of investment bankers is to market securities and to act as intermediaries between corporations and the general public when corporations wish to raise capital (Fisher and Jordan 1991; Melicher and Norton 2000; Petty et al. 2000). (2) Secondary Market The Secondary market is the market for securities already bought, sold, or owned by investors (Fisher and Jordon 1991). Selling securities to investors would be difficult if there was no easy way to profit from their holdings, or no way to sell them for cash (Melicher and Norton 2000). The secondary market provides liquidity and marketability for investors. According to Melicher and Norton (2000), the benefits of the secondary market are as follows: 169 Chapter 5: Analysis of Data and Findings x The secondary market provides liquidity. x The secondary market allows investors to shift their assets into different securities and markets. x The secondary market provides information pricing. x The secondary market provides a means for evaluating a firm’s management and for managers to determine how investors are interpreting their actions. The securities market is beneficial for issuers and investors (people who invest their money in capital markets) (Fisher and Jordan 1991). The Government and firms are issuers of securities. Securities markets help issuers to raise funds. They also ensure savings are converted to capital and directed to productive investments. Governments also borrow to fund their activities. In short, the market helps to efficiently transfer funds from surplus to deficit sectors (Fisher and Jordon 1991; Reilly and Norton 1999). Investors also benefit from the activities of securities markets. As mentioned earlier, securities markets provide investors with liquidity to convert assets to cash. Investors could not easily resell, or buy securities without a liquid market and they would be hesitant to acquire them. This would reduce the quantity of funds which are available to the finance industry and the government (Fisher and Jordon 1991; Reilly and Norton 1999). ‘Those who own securities must be assured of a fast, fair, orderly and open system of purchase and sale at known prices’ (Fisher and Jordon 1991, p. 17). Investors provide the primary source of capital for business (Brigham, Gapenski and Ehrhardt 1999). Fisher and Jordan (1991, p. 2) defined investment as ‘a commitment of funds with the expectation of a positive rate of return.’ Investors include stockholders, bondholders and lenders such as banks (Brigham, Gapenski and Ehrhardt 1999). Brigham, Gapenski and Ehrhardt (1999, p. 39) suggested ‘if the investment is properly undertaken, the return will be commensurate with the risk the investor assumes’. 170 Chapter 5: Analysis of Data and Findings Further, investors ‘are interested in a good rate of return, earned on a consistent basis for a relatively long period of time’ (Fisher and Jordan 1991, p. 2). The Fundamentals of Risk and Return Investors are concerned with two key factors (1) the level of risk and (2) the rate of return. Risk is ‘the chance of financial loss, or the variability of returns associated with a given asset’ (Gitman, Juchau and Flanagan 2002, p. 196). Return is ‘the total gain or loss experienced on an investment over a given period of time. The level of return is calculated by dividing the asset’s change in value plus any cash distributions during the period by its beginning-of-period investment value’ (Gitman, Juchau and Flanagan 2002, p. 196). Investors have different feelings about risk. There are ‘three basic risk preference behaviours: risk averse, risk indifferent and risk seeking’ (Gitman, Juchau and Flanagan 2002, p. 197). Risk-Indifferent Investors feel that ‘no change in return would be required for an increase in risk’ (Gitman, Juchau and Flanagan 2002, p. 197). Risk-Averse Investors feel that ‘an increase in return would be required for an increase in risk’ (Gitman, Juchau and Flanagan 2002, p. 197). Risk-Seeking Investors feel that ‘a decrease in return would be required for an increase in risk’ (Gitman, Juchau and Flanagan 2002, p. 197). Most investors are risk-averse. For a given increase in the level of risk, an increase in the rate of return is desired. Higher returns will compensate investors for holding high risk assets. Therefore, different classes of investments are associated with different levels of risk and rates of return (Reilly and Norton 1999; Gitman, Juchau and Flanagan 2002). 171 Chapter 5: Analysis of Data and Findings Investment Categories Investment is viewed as ‘a single asset held in isolation, or a portfolio which is a collection, or group, of assets’ (Gitman, Juchau and Flanagan 2002, p. 196). When making investment decisions, investors make selections from various investment categories, including: Debt Instruments, Institutional Deposits and Contracts, Government Debt Securities, Private Issues, International Bond Investing, Equities Real Estate, Options and Futures (Brealey and Myers 1991, 2000; Gitman, Juchau and Flanagan 2002). 3.3.2 EFFICIENT MARKET HYPOTHESIS (EMH) As stated earlier, markets which exchange securities are liquid so that firms can obtain finance (Gitman, Juchau and Flanagan 2002). ‘A market is efficient if current market prices fully reflect available information’ (Fama 1970, p. 383). Information should be publically and privately available to investors in the market place. When all available information is freely available, there are no transaction costs, market participants are price takers and a perfect capital market is said to exist (Fama and Miller 1972, Fama 1976, Fama 1998). Fama and Miller (1972) added that: ‘Such a market has a very desirable feature, in particular, at any point in time market prices of securities provide accurate signals for resource allocation; that is, firms can make production-investment decisions, and consumers (investors) can choose among the securities that represent ownership of firms’ activities under the presumption that security prices at any time ‘fully reflect’ all available information. A market in which prices fully reflect available information is termed efficient’ (Fama and Miller 1972, p. 335). In an efficient market, it is impossible to receive above-average returns and there are no reasons to believe that current prices are too high or low (Christensen and Jordan 2001). Academics and investors have attempted to measure the performance of the market’s price mechanism and to apply descriptions to the market. The accepted description is efficient-markets theory (EMT) or, the efficient market hypotheses (EMH) of Fama 172 Chapter 5: Analysis of Data and Findings (1970). EMH has been accepted as a description of the world’s major stock markets such as the New York Stock Exchange (Firth 1977) and the London Stock Exchange (Firth 1977; Fisher and Jordan 1991; Haugen 1995; Petty et al. 2000; Christensen and Jordan 2001; Gitman, Juchau and Flanagan 2002). EMH states that a market is efficient with respect to a given information set, if it is impossible to make profits by trading on the basis of this information set (Jensen 1978). Fama (1970) divided EMH into three categories of market efficiency depending on the type of information which is available: (1) Weak, (2) Semi-strong, and (3) Strong (Fama 1970). Weak market efficiency is present if past prices cannot be used for forecasting future prices (Fama 1970; Jensen 1978; Sutcliffe 1997). This implies that technical analysis is ineffective (Fama 1970; Jensen 1978; Sutcliffe 1997). Technical Analysis is the study of an investor’s psychology by observing his or her behaviour (Fama 1970; Jensen 1978; Sutcliffe 1997). This technique is widely used by stock brokers, speculators and investment advisors (Sutcliffe 1993, 1997). See figure 3.2. In a semi-strong market, public information is ineffective for forecasting futures prices (Fama 1970; Jensen 1978; Sutcliffe 1997). This implies that fundamental and technical analysis is inappropriate (Fama 1970; Jensen 1978; Sutcliffe 1997). ‘Fundamental analysis is applied when a security analyst uses public information (e.g. published accounts, interviews with the company management and public statements by the company) to value, predict or forecast the prices of stock (Fisher and Jordan 1991, p. 612). Therefore, trading techniques which are past pricing and published information are not helpful for investors that wish to increase their returns in a semi-strong market (Fama 1976 1998). See figure 3.2. When the level of efficiency is strong, all information (both public and private) is ineffective for forecasting future prices (Fama 1970; Jensen 1978; Sutcliffe 1997). This also implies that the use of a firm’s internal information (insider trading) is not profitable (Fisher and Jordan 1991). See figure 3.2. 173 Chapter 5: Analysis of Data and Findings Fundamental research says that ‘prices reflect all relevant information and that changes in prices are unpredictable’ (Brealey and Myers 2000, p. 358). However technical research says that ‘current prices reflect past information and prices, and that future price changes cannot be predicted from past prices’ (Brealey and Myers 2000, p. 358). If the capital market is not efficient, technical and fundamental analysis could be profitable. However, if the capital market is at least weakly efficient, technical analysis will be ineffective, but fundamental analysis can still be useful. If the efficiency of a capital market is semi-strong, fundamental analysis will also be ineffective (Sutcliffe 1993). Fundamental analysis can therefore, be applied in markets with less than semistrong efficiency, while technical analysis can be applied only when the level of efficiency is less than weak (Fama 1976). This supports the view that fundamental analysis is effective in a wider range of markets than technical analysis (Fama 1976, 1998). If public information is regularly updated and readily available to investors, fundamental analysis can be very useful in investment strategies (Fama 1976, 1998). Figure 3.2: Type of Information in the Capital Markets Past Prices Public Information Private Information All Information Source: Jensen (1978); Sutcliffe (1997) 174 Chapter 5: Analysis of Data and Findings After assessing the information which is available to the public, fundamental investors and analysts normally estimate the intrinsic value of a stock. This estimate is equivalent to its equilibrium share price after trading in the current market (Fisher and Jordan 1991). ‘Intrinsic value, or fair value, is the present value of the investment’s expected future cash flows, discounted at the investor’s required rate of return’ (Petty et al. 2000, p. 263). Future cash flows can be estimated using publicly available information. Therefore, the intrinsic value of a security is equivalent to its investor value (Petty et al. 2000). After the investor has determined the intrinsic value of a security, this can be compared with the security’s current market value (market price) (Petty et al. 2000). If the intrinsic value is not equal to the current market value (price), the security is mispriced (Firth 1977). A mispriced security can be categorised in three ways. If the intrinsic value is greater than the market value, the security is underpriced (Petty et al. 1991). If the intrinsic value is lower than the market value, the security is overpriced (Petty et al. 1991). If the intrinsic value is equal to the market value, the market is efficient (Petty et al. 2000) because price reflects all the available information. Over time, the forces of supply and demand tend to adjust the price of a mispriced share towards its intrinsic value (Fisher and Jordan 1991). Investors and analysts sell overpriced stocks, thereby driving their prices towards their intrinsic value. The same scenario applies to underpriced stocks (Fisher and Jordan 1991). Figure 3.3 indicates that prices of overpriced and underpriced stocks move towards their intrinsic value as times passes. 175 Chapter 5: Analysis of Data and Findings Figure 3.3: Removal of Mispricings Overpriced Share price Intrinsic Value Intrinsic Value Underpriced Time Source: Sutcliffe (1997) The decision rules of investors and analysts are summarised below: Investors will buy a stock if the intrinsic value is greater (>) than the current price, or will sell a stock if the intrinsic value is less (<) than the current price. Table 3.5 illustrates the decision rules for investors and analysts. Table 3.5: Decision Rules for Investors and Analysts Decision rules Intrinsic Value > Market Value Buy Sell D Intrinsic Value < Market Value D Source: Developed for this research In practice, no stock market is perfectly efficient in the terms described above (Firth 1977). Complete knowledge about a firm is not publicly available, and if it was, it is unlikely that it would be interpreted accurately enough to reflect the value of the share price (Firth 1977). Although there are shortcomings with EMH, researchers have considerable scope for measuring the performance of the stock market and the relevance of efficient-markets theory (Fama 1970; Firth 1977). 176 Chapter 5: Analysis of Data and Findings 3.4 THE IMPORTANCE OF DIVIDENDS As discussed earlier, firms issue equity capital to obtain funds from investors in equity markets. However, investors require a reasonable return on the funds they lend to firms. This leads to a discussion of dividends and dividend policy. 3.4.1 DIVIDENDS AND STOCK VALUATIONS The profits of successful firms can be reinvested, used to acquire securities, retire debt or distributed to shareholders as dividends (Bringham, Gapenski and Ehrhardt 1999). Dividends are a cash flow to shareholders of firms (Gitman 2000; Gitman, Juchau and Flanagan 2002). They are considered a return for holding stocks and provide a method for determining the value of a share (Gitman 2000; Gitman, Juchau and Flanagan 2002). The Dividend Discount Model (DDM) is used for determining the value of shares in equity markets (Reilly and Brown 2000). Dividends are the only direct cash flows from firms to investors (Brealey and Myers 2000; Gitman, Juchau, Flanagan 2002). The concept of discounted dividend is expressed as: PV (stock) = PV (expected future dividends) The reduced form of the Dividend Discount Model (DDM) is known as the Gordon model. PV (stock) = Expected Dividend / (Discount rate - Growth) This model is useful for discussing the valuation of stable or mature firms, where an assumption of relative constant growth and a stable discount rate for the long term is appropriate (Reilly and Brown 2000). Therefore, investigating the dividend phenomenon will contribute to a knowledge of the stock’s return and its valuation. 177 Chapter 5: Analysis of Data and Findings Investors may buy stocks and expect to receive capital gains in addition to dividends (Brealey and Myers 2000). Consequently, the cash flow from investing in stocks is in two forms (1) cash dividends and (2) capital gains or losses. The current price of a stock is equal to the sum of future dividends and the expected future price of the stock discounted to the present (Brealey and Myers 2000). Therefore; PV (stock) = PV (Future dividend + Expected Future stock price) Although firms prefer issuing common stocks rather than preferred stocks, investors are aware that common shares are more risky than preference shares (Brealey and Myers 2000). Common stocks are more difficult to value than other securities for three reasons: Firstly, the cash flows of common stocks are undertaken because the value of dividends and expected future stock prices, are not known in advance (Christensen and Jordan 2001). Secondly, common stocks do not mature, therefore they have an indefinite life span (Christensen and Jordan 2001). Thirdly, the required market rate of return is not easy to observe (Christensen and Jordan 2001). Two main classes of techniques have been used for valuing stocks: (1) Discounted Cash-Flow Valuation Techniques: the value of stocks depends on the present value of several cash flow items, such as dividends, the operating cash flow, the free cash flow, the Dividend Discount Model, the present value of operating cash flows and the present value of free cash flows (Reilly and Brown 2000). (2) The Relative Valuation Techniques: the value of stocks depends on their current prices relative to several significant variables. These variables include the Price/Earnings Ratio (P/E), Price/Cash Flow Ratio (P/CF), Price/Book Value Ratio (P/BV) and Price/Sales Ratio (P/S) (Reilly and Brown 2000). 178 Chapter 5: Analysis of Data and Findings This study will investigate the phenomenon of disappearing dividends, which states that dividends are no longer being paid by listed firms (Fama and French 2001). If this phenomenon exists, it will raise the degree of complexity for valuing stocks when using the discounting dividend technique and analysing ratios which are associated with dividends (such as dividend yields and dividends per share). If a firm does not wish to distribute its earnings as dividends, it holds them as retained earnings which can be re-invested for upgrading a firm’s operations (Gitman, Juchau and Flanagan 2002). If the firm requires finance, it can use the retained earnings, (internal financing) or borrow from elsewhere (external financing) (Gitman, Juchau and Flanagan 2002). Therefore, if a firm pays high dividends it is likely that it will need to borrow more funds from external sources (Gitman 2000; Gitman, Juchau and Flanagan 2002). As stated earlier, dividends are an observable cash flow (Wetherilt and Weeken 2002) for investors. It is also possible to measure and analyse these flows. 3.4.2 COMPONENTS OF DIVIDEND POLICY AND DIVIDEND PAYMENT PRODECURES A firm’s dividend policy has two basic components: (1) A Dividend Payout Ratio which indicates the value of dividends paid relative to the company’s earnings (such as 50% of the earnings or 60% of the earnings) (Petty et al. 2000). (2) The Stability of Dividends is important because profits and investment opportunities vary over time (Brigham, Gapenski and Ehrhardt 1999). Investors desire certainty with future payments of dividends, but it is impossible to accurately estimate the value of the dividends they will receive in any specific period in the future. However, several indicators can help investors estimate the level and stability of a firm’s dividends. These indicators help enhance the level of confidence of investors and aid in their future decision making (Petty et al. 2000). 179 Chapter 5: Analysis of Data and Findings x Listed firms forecast the level of their earnings and dividends five to ten years in advance, however these estimates are not available to the public. However, analysts prepare similar forecasts which are available to the public (Brigham, Gapenski and Ehrhardt 1999). x A Stable Dividend Policy commits a firm to pay the same value of dividends each year (Brigham, Gapenski and Ehrhardt 1999). Firms with unstable earnings may be unwilling to make such a commitment and investors would not be aware of this decision by management, because it would not be detailed in the firm’s annual report. When determining a policy for dividends, a firm’s management faces a trade-off between the satisfaction of shareholders and the required level of external financing (Petty et al. 2000; Brigham, Gapenski and Ehrhardt 1999). If a firm makes large dividend payments there will be less retained earnings (or profit retention) and possibly a greater need to borrow from external sources (Petty et al. 2000). The reverse is true when a firm pays a lower level of dividends. Figure 3.4 illustrates a firm’s trade-off between large and small payments of dividends. Figure 3.4: Dividend-Retention-Financing Trade-Offs Given Investment Decisions Debt-Equity mix Choices between Large Dividends Lower Profit Retention Heavy External Equity Financing Small Dividends or High Profit Retention Negligible External Equity Financing Source: Petty et al. (2000), p. 466 180 Chapter 5: Analysis of Data and Findings The value of dividends which will be paid to shareholders is decided by the firm’s board of directors. The firm’s past performance and history of dividend payments are considered before the announcement of the next dividend payment to shareholders. The dividend payment procedures include of the amount to be paid, the record date and the date of payment (Gitman 2000; Gitman, Juchau and Flanagan 2002). These procedures are discussed below: Record Date Firms usually pay dividends twice a year. Dividend policies vary between countries, but generally firm’s will announce their dividend policies in the financial news media (Gitman, Juchau and Flanagan 2002). Gitman, Juchau and Flanagan (2002, p. 450) mentioned that the ‘date of dividend records is when all shareholders are notified that they will receive their dividends on a specific date. Ex-Dividend Date According to Gitman, Juchau and Flanagan (2002) the ex-dividend date is a period which commences five business days prior to the date of record. All shareholders who sell their stocks during this period will not receive the current dividend. To be eligible for a dividend payment, investors must buy shares before the ex-dividend date (Fisher and Jordan 1991; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). Payment Date The payment date is the date when dividends are distributed to shareholders (Gitman 2000; Petty et al. 2000; Gitman, Juchau and Flanagan 2002). Figure 3.5 illustrates the sequence of payment dates from the announcement date to payment date. 181 Chapter 5: Analysis of Data and Findings Figure 3.5: Sequence of Dividend Payment Dates Announcement Date Record Date Declaration Date Payment Date Ex-dividend Date Source: Petty et al. 2000, p. 493. 3.4.3 THE RELEVANCE OF DIVIDEND POLICY The Residual Theory of Dividends suggests that the dividends paid by a firm should be viewed as a residual amount remaining after all investment opportunities have been met (Gitman, Juchau and Flanagan 2002). When a firm makes long-term investment decisions, the cost of capital is an important concept (Gitman, Juchau and Flanagan 2002). ‘The cost of capital is the rate of return a firm must earn on its project investment to maintain the market value of its shares’ (Gitman, Juchau and Flanagan 2002; p. 392). The cost of capital is a benchmark for justifying whether an investment should be undertaken. Projects with rates of return greater than the cost of capital will increase the value of the firm and should be accepted (Gitman 2000; Petty et al. 2000; Gitman, Juchau and Flanagan 2002). Firms desire and attempt to maintain an optimal mix of debt and equity financing (Gitman, Juchau and Flanagan 2002). The firm’s capital structure is comprised of three sources of finance, namely, (1) long-term debt (2) preference share capital and (3) common share capital (Gitman 2000; Gitman, Juchau and Flanagan 2002). The overall cost of capital can be defined as: ka = ( w i x k i ) + ( w p x k p ) + ( w s x k r or n ) 182 Chapter 5: Analysis of Data and Findings where ka = the cost of capital wI = the proportion of long-term debt in the capital structure wp = the proportion of preference share capital in the capital structure ws = the proportion of common share capital in the capital structure w s + w p + w I = 1 or 100% kI = after-tax cost of long-term debt in the capital structure kp = after-tax cost of preference share capital in the capital structure k r or n = after-tax cost of common share capital in the capital structure If the project’s return is greater than, or equal to the weighted marginal cost of new finance, the firm should accept the investment project’ (Gitman, Juchau and Flanagan 2002, p. 412). However, if the returns from the investment are lower than the cost of capital, the firm should distribute the earnings (as a dividend) to stockholders (Gitman 2000). Dividend Relevance Arguments The dividend relevance argument has been supported by Gordon (1962) and Lintner (1962). These two scholars suggested that there is a direct, or positive correlation between a firm’s dividend policy and the market value of this firm. Lintner (1962) and Gordon (1962) suggested the bird-in-the-hand argument which says that investors are generally risk-averse (investors try to minimize risk, or if the level of risk is high they will expect a higher rate of return as compensation for consuming high levels of risk). Therefore, the payment of dividends should reduce the level of uncertainty (risk) of holding stocks. If the level of dividends is reduced, however, the level of uncertainty will increase and this will place downward pressure on stock values (Gordon 1962; Lintner 1962; Brigham, Gapenski and Ehrhardt 1999; Gitman, Juchau and Flanagan 2002). 183 Chapter 5: Analysis of Data and Findings Dividend Irrelevance Argument Academics and experienced practitioners perceive the relationship between dividends and share prices in a variety of ways (Petty et al. 2000). Practitioners believe changes in share prices result from announcements of dividends (Petty et al. 2000). However, the dividend irrelevance argument has been suggested by Miller and Modigliani (1961). These researchers claim, in a perfect world, the value of a firm is unaffected by the distribution of dividends and is determined by the earnings ability of the firm and the risk of holding assets (Miller and Modigliani 1961; Gitman, Juchau and Flanagan 2002). This argument suggests that firms can transfer value, and preserve the market value of shares with a variety of combinations of assets, earnings and investment opportunities (Brealey and Myers 2000). If the firm pays dividends, each share will be worth less because new shares must be issued to finance the cash outflow. Investors can obtain cash by selling their shares in the market (if the market is liquid or efficient) (Miller and Modigliani 1961; Brealey and Myers 2000). Therefore, firms need not worry about their dividend policy (Brealey and Myers 2000). Figure 3.6 shows the relationship between new stockholder, existing stockholders and the firm’s trade-off between cash and shares. Figure 3.6: Two Ways of Raising Cash for Firm Dividends Financed by Stock Issue Shares New Stockholders Cash Firm No Dividend, No Stock Issue New Stockholders Cash Shares Cash Old Stockholders Old Stockholders Source: Brealey and Myers (2000), p. 445. 184 Chapter 5: Analysis of Data and Findings Dividend irrelevance theory has two preconditions or assumptions. Firstly, Miller and Modigliani assume that investment and borrowing decisions have been undertaken and will not affect the value of the dividend payment (Miller and Modigliani 1961; Petty et al. 2000). Secondly, the scenario described by these researchers assumes that (1) investors can buy and sell stocks without incurring transaction, or brokerage costs and, (2) firms can issue shares without issuing costs and, (3) firms and investors are not taxed and, (4) all information about the firm is available to the public and, (5) there is no conflict of interest between management and stockholders and, (6) borrowing and lending are unlimited and, (7) there is no problem with liquidity (Petty et al. 2000). Given these assumptions, Miller and Modigliani suggested that there is no relationship between a firm’s dividend policy and the value of its shares (Miller and Modigliani 1961; Petty et al. 2000). An absence of taxes and brokerage costs (Brigham, Gapenski and Ehrhardt 1999) and the assumptions made by Miller and Modigliani indicate that the dividend irrelevance argument theory can be contested (Brigham, Gapenski and Ehrhardt 1999). Studies have indicated that changes to dividends affect share prices, or are positively correlated with share prices (James, Dodd and Kimpton 1985; Rappaport 1986; March and Merton 1987; Penman and Sougiannis 1998; Hurley and Johnson 1998; Yao 1997; Fama and French 1999, 2000, 2001; James, Fama and French 2000, Wetherilt and Weeken 2002). A higher payment of dividends is a positive signal for investors to buy the shares, while a lower payment is a negative signal and suggests that the investor should sell the shares (Gitman, Juchau and Flanagan 2002). As discussed earlier, an increase in dividend payments is often followed by an increase in the price of a firm’s stocks and this may indicate that investors prefer dividends to capital gains (Brigham, Gapenski and Ehrhardt 1999). However, Miller and Modigliani (1961) argued that an increase, or decrease in the payment of dividends is merely a signal that a firm’s managers have forecast a positive, or negative outlook for profits. It is suggested that the evidence termed ‘signaling of dividends’ is unclear (Brigham, Gapenski and Ehrhardt 1999). 185 Chapter 5: Analysis of Data and Findings The evidence on whether dividends are relevant, or irrelevant is mixed. It appears, however, that investors possess a variety of views on the importance of dividends. Some investors prefer high dividends, while others are satisfied with low dividends (Brigham, Gapenski and Ehrhardt 1999). In addition, some investors prefer capital gains rather than dividends. Overall it appears that investors prefer firms with a stable and predictable dividend policy (Brigham, Gapenski and Ehrhardt 1999). 3.4.4 DIVIDEND POLICY AND SHARE PRICE MOVEMENT As mentioned earlier, a firm’s dividend policy is likely to affect the value of its shares. There are three views of the relationship between dividend policy and share prices: (1) dividend policy is irrelevant; (2) high dividends increase share values and; (3) low dividends increase share values (Brigham, Gapenski and Ehrhardt 1999; Gitman 2000). (1) Dividend policy is irrelevant: As discussed earlier, several analysts believe that a firm’s dividend policy has no effect on its share price (Miller and Modigliani 1961; Petty et al. 2000). (2) High dividends increase share values: Dividends are more predictable than capital gains and managers can control dividends, although they cannot directly influence the price of shares. Shareholders believe the risks associated with capital gains are higher than those with dividends (Brigham, Gapenski and Ehrhardt 1999; Brealey and Myer 2000; Gitman 2000). The value of dividends are therefore more predictable than the size of capital gains. The ‘bird-in-the-hand’ theory is therefore accepted (Linter 1956). Investors value a unit of expected dividends more highly than a unit of expected capital gains (Brigham, Gapenski and Ehrhardt 1999; Gitman 2000). Consequently, high dividends increase the value of shares (Brigham, Gapenski and Ehrhardt 1999; Gitman 2000; Gitman, Juchau and Flanagan 2002). (3) Low dividends increase share values: Income in the forms of dividends and capital gains is taxable. However, tax on dividend income is paid when the dividends are received, while tax on capital gains can be deferred until the stock is sold (this is 186 Chapter 5: Analysis of Data and Findings termed deferring the payment of taxes). Investors attempt to defer tax payments for as long as possible (Petty et al. 2000). Therefore, if taxation is considered, investors prefer capital gains rather than the payment of dividends (Brigham, Gapenski and Ehrhardt 1999; Gitman 2000; Gitman, Juchau and Flanagan 2002). Corporate dividend policies have been the subject of intensive theoretical modelling and empirical examination. There are many views on dividend payments. Large payments can be harmful to investors and firms, although, some firms are willing to pay high dividends (Petty et al. 2000). This phenomenon is termed the ‘dividend puzzle’ and many scholars and practitioners has provided evidence to explain this phenomenon (Petty et al. 2000). However, the conclusions of theoretical models are often conflicting and lack empirical support when attempting to explain the dividend policies of firms. Policy mainly considers the likely effect of dividend payments on share prices. As discussed earlier, there are two schools of thought and their conclusions are in conflict. The relevance theories of Gordon (1962, 1963) and Lintner (1962) suggest that dividends have positive and negative influence on stock prices. As stated earlier, irrelevance theory states that a firm’s dividend policy does not affect the value of its stocks. Recent models focus on corporate dividend policy and taxation and their impact on the behaviour of investors. Some of these models have been formulated using full information, information asymmetries and behavioral principles. Several recent models are discussed below. A. Models of Full Information and Tax Factor Investors expect higher returns in the form of dividends due to increasing tax liabilities (Frankfurter and Gong 1992, 1993). The shareholder’s pre-tax returns should therefore be increased. Miller and Modigliani (1961) separated investors into a variety of dividend tax clienteles. More recent research by these scholars (Miller 1977; Modigliani 1982) found that the ‘clientele effect’ has a minor influence on the composition of an investor’s portfolio. Masulis and Trueman (1988) found that 187 Chapter 5: Analysis of Data and Findings investors have a variety of views on a firm’s investment and dividend policies. Assuming that investors aim to maximize their after-tax income, Farrar and Selwyn (1967) developed a model which suggests that dividends need not be paid, although share repurchases should be used for distributing corporate earnings. Farrar and Selwyn (1967) assumed that investors can decide whether they want to receive their returns in the form of cash dividends or capital gains. These findings were also supported by Brennan (1970) and Auerbach (1979a). It has been suggested that investors will buy stocks which pay dividends and invest in tax free securities if they are likely to be highly taxed (Miller and Scholes 1978; Auerbach 1979b; Miller 1986). The returns from tax-free securities are preferred to dividends which are highly taxed. However, this theory was rejected by DeAngelo and Masulis (1980) and Fung and Theobald (1984). These researchers concluded that the model is insufficient, incomplete and does not consider interest and liquidity. B. Models of Information Asymmetries B1. Signaling Models A firm’s dividend policy is often used as a less costly way of convincing shareholders that the firm is being well managed. Many researchers support this finding, including, Laub (1976), Bar-Yosef and Kolodny (1976, 1978), Bhattacharya (1979, 1980), Talmor (1981), Hakansson (1982), John and Williams (1985), Asquith and Mullins (1983), Miller and Rock (1985), Bar-Yosef and Huffman (1986), Makhija and Thompson (1986), Ambarish, John and Williams (1987), Ofer and Thakor (1987), Kumar (1988), Kale and Noe (1990) and Rodriguez (1992). B2. Agency Cost The problems of agency costs are not new. Adam Smith (1937) and (Kindleberger 1984) found that it was not possible for shareholders to monitor a firm’s managers because of the inefficiencies and costs involved. Agency problems result if a firm’s 188 Chapter 5: Analysis of Data and Findings managers engage in high-risk-high-return projects and overlook projects with a positive net present value. Easterbrook (1984) suggested that a large dividend payout could reduce agency costs by forcing managers to seek funds from capital markets. B3. The Free Cash Flow Hypothesis The problem of agency costs was also considered by Berle and Means (1932). These researchers concluded that the funds which remain after a firm has invested in profitable areas were used inefficiently and raised the level of conflict between managers and shareholders. Jensen (1986) developed a free cash flow hypothesis which states that dividends and debt interest payments decrease the volume of free cash flows which are available to managers. In short, dividends and interest repayments help prevent managers from engaging in non-productive investments. C. Behavioral Models Several recent models have focused on the behavior of investors because it may influence the dividend policies of firms. Shiller (1984) noted that the behaviour of investors is influenced by societal norms and attitudes. The behavioral models initially failed to gain acceptance but could be used to explain some issues in the area of dividend policy. Shiller (1984) also mentioned that social pressures can lead to poor judgment by investors or shareholders. C1. Managerial Surveys Lintner (1956) interviewed a number of corporate chief executive officers and chief financial officers and found they are reluctant to reduce dividends, or adopt a policy of stable dividends because this may provoke a negative response from investors. Darling (1957), Turnovsky (1967), Fama and Babiak (1968), Baker, Farrelly and Edelman (1985), Baker and Farrelly (1988), DeAngelo, DeAngelo and Skinner (1992) supported Linter’s findings. DeAngelo, DeAngelo and Skinner (1992) also suggested that factors such as regulatory constraints, investment magnitude, debt and firm size also affect a 189 Chapter 5: Analysis of Data and Findings firm’s dividend policy. Managers consider current and expected earnings, a firm’s dividend payment history, dividend stability, cash flows and investment opportunities and the view of shareholders when determining the level of dividends (Harkins and Walsh 1971). They also consider issues such as sustainability, current profitability, expectations of future cash flows, and industry norms when determining their payout (Frankfurter and Lane 1992). C2. Theoretical Behavioral Models Shefrin and Statman (1984) explained dividend preferences with the theory of selfcontrol and the descriptive theory of choice under uncertainty (Kahneman and Tversky, 1982). The theory of finance says that dividends and capital gains are perfect substitutes. Thaler and Shefrin (1981) argued that dividends and capital gains are not always perfect substitutes. The costs and payoffs of high risk investments are evaluated separately because losses are more significant than gains (Thaler and Shefrin 1981). Individuals have saving and investment goals, therefore, they will not sell parts of their portfolio if they expect to receive dividends from the shares (Frankfurter and Wood 2001). Variation in the results obtained by the above models is explained by differences in modeling, the method of analysis, the variables and definitions used, the type of data and the sample period (Watts 1973, 1976a, 1976b; Roll 1977, 1983; Miller and Scholes 1982; Frankfurter and Gong 1992, 1993; Watts 1973; Morgan 1980, 1982; Baker and Farrelly 1988). Black (1976) and Feldstein and Green (1983) concluded that dividend policy remains a puzzle because the existing models lack empirical support. However, some general conclusions can be drawn about dividend policies. Managers decrease the level of dividend payments only when necessary (Myers, 1984; and DeAngelo, DeAngelo, and Skinner, 1992). Shareholders must pay taxes on dividend income and research supports the hypothesis that returns on dividend-paying stocks are increased to offset the tax liability (Frankfurter and Gong 1992, 1993). 190 Chapter 5: Analysis of Data and Findings Managers are reluctant to reduce dividend payments if a firm is suffering from financial distress. Likewise, dividends are only increased if managers are confident that high levels of profit can be sustained. Managers are also aware that shareholders expect dividends to be paid and feel they deserve this payment (Myers 1984). 3.4.5 OTHER FORMS OF DIVIDENDS Share dividends or bonus shares, share splits, share buybacks and share repurchases are alternative forms of dividends which are paid to shareholders. Firms pay stock dividends, or buy back shares as a replacement for, or a supplement to cash dividends (Gitman, Juchau and Flanagan 2002). Share Dividends Share dividends, or a bonus share issue is a method of paying dividends by giving shares to the existing shareholders. Share dividends are normally issued as a proportion of current holdings, such as one bonus share for ten shares owned (Gitman 2000; Petty et al 2000; Gitman, Juchau and Flanagan 2002). Share Dividends: the Shareholder’s Viewpoint A share dividend signals potential growth and higher confidence in a firm’s current operations. This is because a firm needs to be confident that it can generate enough earnings to pay dividends on the expanded, or additional shares (Gitman 2000; Petty et al. 2000; Gitman, Juchau and Flanagan 2002). Share Dividends: the Company’s Viewpoint Share dividends are more costly than cash dividends because the firm needs to pay dividends on the additional shares (Gitman, Juchau and Flanagan 2002). However, share dividends are a good option if the firm is growing fast and needs internal financing (Gitman, Juchau and Flanagan 2002). Share dividends are also a way of giving shareholders a bonus without paying cash (Gitman, Juchau and Flanagan 2002). Share dividends are appreciated by the shareholders as long as 191 Chapter 5: Analysis of Data and Findings the internal funds are used to enhance the firm’s growth and not for paying overdue bills (Gitman 2000; Gitman, Juchau and Flanagan 2002). Share Buybacks or Share Repurchase Share buybacks are also a form of dividend payment because excess cash, or retained earnings are used by the firm to purchase its own shares. Share buybacks increase shareholders value in three ways: (1) by reducing the number of shares outstanding, and thereby raising Earnings Per Share (EPS); (2) giving a positive signal that management believes the shares are undervalued and; (3) by preventing an unfriendly takeover (Gitman 2000; Gitman, Juchau and Flanagan 2002). This study will focus on cash dividends because they constitute actual cash which is paid to shareholders. Cash dividends can significantly affect a firm’s financing because they affect the volume of retained earnings (internal funds available) and the funds which are available for investment (James, Dodd and Kimpton 1985; Melicher and Norton 2000; Gitman, Juchau and Flanagan 2002). 3.4.6 FACTORS AFFECTING DIVIDEND POLICY Legal Constraints Firms only pay dividends when there are current, or accumulated past profits which are available for distribution to shareholders. Firms are prevented from paying cash dividends from its legal, or registered capital. The capital is required to register a firm and must not be invested (James, Dodd and Kimpton 1985; Gitman, Juchau and Flanagan 2002). Internal Constraints Profit is not cash, therefore firms with a high level of profitability may not possess a high ability to pay dividends. Dividend payments are a drain on a firm’s liquid assets (cash). The level of dividend payments, therefore depends on the internal constraints of 192 Chapter 5: Analysis of Data and Findings a firm such as the liquidity of its assets (James, Dodd and Kimpton 1985; Grinblatt and Titman 1998; Gitman, Juchau and Flanagan 2002). Contractual Constraints Firms with restrictive provisions in loan agreements may have difficulty paying dividends. Generally, loan agreements prohibit the payment of cash dividends until a certain level of earnings has been achieved. These provisions protect a firm’s creditors from any losses in the case of insolvency (James, Dodd and Kimpton 1985; Grinblatt and Titman 1998; Gitman, Juchau and Flanagan 2002). Growth Prospects Dividend policies may vary according to the stages of a firm’s growth. Growing firms may need cash for additional investment projects. In addition, low growth firms may also need cash for the replacement of existing assets. Therefore, a firm’s potential growth can affect its dividend policies (James, Dodd and Kimpton 1985; Grinblatt and Titman 1998; Gitman, Juchau and Flanagan 2002). Owner Considerations The goal of the firm is to maximize the shareholders’ wealth. Consequently, the views of shareholders also influence a firm’s dividend policy. Issues which will affect the views of shareholders include tax status, investment opportunities and the dilution of ownership (James, Dodd and Kimpton 1985; Grinblatt and Titman 1998; Gitman, Juchau and Flanagan 2002). Market Considerations Lintner (1956) suggested that the market’s response to a firm’s dividend policy is also important for its success and is helpful in developing future dividend policies. As stated earlier, shareholders often view the firm’s dividend payments as an indicator of its future success (Grinblatt and Titman 1998; Gitman, Juchau and Flanagan 2002). 193 Chapter 5: Analysis of Data and Findings 3.4.7 TYPES OF DIVIDEND POLICIES Gitman, Juchau and Flanagan (2002, p. 452) defined dividend policy as ‘the firm’s plan of action to be followed whenever a decision concerning dividends must be made’. Three dividend policies which are commonly used are described below. Constant-Payout-Ratio Dividend Policy The constant payout ratio is the most commonly used dividend policy and is calculated by dividing the firm’s cash dividend per share by its earnings per share (EPS). The constant payout ratio is used to ensure a specific percentage of earnings is paid to owners in each dividend period (Grinblatt and Titman 1998; Gitman, Juchau and Flanagan 2002). Stable Dollar Dividend Per Share This policy aims to maintain the value of dividend payments over time. A firm’s management must be convinced that a specific dividend level can be maintained before adopting such a policy (Brigham, Gapenski and Ehrhardt 1999). Low-Regular-and-Extra Dividend Policy Some firms establish a low-regular-and-extra dividend policy, which involves distributing a low dividend payment on a regular basis and supplementing this with a bonus when earnings are high (Gitman, Juchau and Flanagan 2002). 3.4.8 FIRM’S DIVIDEND PAYMENT DECISION Lintner’s Model Lintner (1956) conducted a series of interviews on the dividend policies of firms with managers of a variety of firms. He found that dividends are determined by four stylised facts: x ‘Firms have long-run target dividend payout ratios. Mature firms have a high payout ratio while growth firms have a low payout ratio’ 194 Chapter 5: Analysis of Data and Findings x ‘Managers focus more on dividend changes than on the absolute level’ x ‘Dividend changes follow shifts in long-run, sustainable earnings. Managers try to smooth dividends’ x ‘Managers are reluctant to make dividend changes’ (Brealey and Myers 2000, p. 443). The findings of Lintner (1956) were published in DeAngelo, DeAngelo and Skinner (1990). These findings were as follows: In times of financial difficulties, managers are more likely to reduce dividend payments than cease paying them. In addition, managers of firms with long dividend histories tend to avoid cuts, or omissions, to preserve the favourable views of the firms which are held by investors (DeAngelo and DeAngelo 1990). From these findings, Lintner developed a model for explaining the payment of dividends: DIV 1 = Target Dividends = Target Ratio x EPS 1 = Target Ratio x EPS 1–DIV 0 The dividend change would equal DIV 1 – DIV 0 = Target Change When a firm decides to alter its target payout ratio (according to Lintner’s survey they are reluctant to do this) the changes are formulated are follows: DIV 1 – DIV 0 = Adjustment Rate x Target Change = Adjustment Rate x (Target Ratio x EPS1 – DIV0) Fama and Babiak (1968) suggested the findings of Lintner (1956), however, this study found that the model could not explain all of the reasons that firms pay dividends. Consequently, there is no universal approach that explains dividend policy. Managers 195 Chapter 5: Analysis of Data and Findings may wish to maintain a long history of continuous dividend payments, strategic considerations, and binding debt covenants (DeAngelo, DeAngelo and Skinner 1990). 3.5 THE PHENOMENON OF DISAPPEARING DIVIDENDS IN CAPITAL MARKETS 3.5.1 THE GLOBAL PHENOMENON OF DISAPPEARING DIVIDENDS In the US, dividends are generally taxed at a higher rate than capital gains. Therefore, firms presume that dividends are less valuable than capital gains (Fama and French 2001). Consequently, firms that pay dividends are disadvantaged because they experience higher costs of equity than firms that do not pay dividends, or finance their activities with equity (Fama and French 2001). Even though dividends are a high cost method of obtaining finance, some firms continue to pay dividends. Some firms claim that they need to compensate and be honest to their loyal stockholders (Petty et al. 2000). Fama and French (2001) investigated the issue of disappearing dividends in the US. This study focused on the period from 1978 to 1999 and found that the percentage of firms paying dividends in the US declined sharply after 1978. In 1973, 52.8 percent of listed firms paid dividends and this figure rose to 66.5 percent in 1978. However, by 1999 only 20.8 percent of US listed firms paid dividends. Therefore, the percentage of firms paying dividends decreased by more than 50 percent between 1973 and 1999. Fama and French (2001) provided two explanations for this outcome. Firstly, the characteristics of newly listed firms changed alter 1978. At present, many listed firms experience characteristics such as low profitability, strong growth and have never made dividend payments (Fama and French 2001). Secondly, Fama and French (2001) studied the characteristics of listed firms over time and found that less firms appear to pay dividends than in the past. Fama and French (2001) entitled this trend; ‘the declining propensity to pay dividends’. Fama and French (2001) used summary statistics (or descriptive statistics) such as, frequencies and means, to categorise the firms listed (excluding financials and utilities 196 Chapter 5: Analysis of Data and Findings firms) on three US stock markets (AMEX, NASDAQ and NYSE). Three characteristics were found to affect the level of dividend payments, namely profitability, investment opportunities and a firm’s size. Logit regression analysis was used to confirm the significance of these factors. Earlier researchers such as Tobin (1958), Gordon (1962), (1963), Higgin (1972) Fazzari, Hubbard, and Petersen (1988), Fama and French (1997) and Aivazian, Booth and Cleary (2003) focused on the relationship between internally generated funds, profitability and investment. Ratios were used to assess the level of profitability, investment, growth, size or the liquidity of firms to standardise the measurement of these variables. Ratios which are often used by researchers include the debt ratio (total liabilities divided by total assets), the M/B ratio (market to book value), ROA (net income divided by total assets), ROE (net income divided by shareholders’ equity), total assets, and the growth in assets (Fazzari, Hubbard, and Petersen 1988; DeAngelo and DeAngelo 1990; Fama and French 1995, 1997; James, Fama and French 2000, Aivazian, Booth and Cleary 2003). In relation to a firm’s profitability, Fama and French (2001) used Et/At (the ratio of aggregate earnings before interest to aggregate assets) and Yt/BEt (aggregate common stock earnings over aggregate book equity). After calculating the values and the average for the two ratios, the average values of ratios for payers of dividends, nonpayers and former payers were calculated using statistical software (Fama and French 2001). Fama and French (2001) found that the profitability of the firms (as measured by Et/At and Yt/BEt ratios) varies with the type of firm (dividends payers, former payers and non-payers). Firms that distribute dividends were found to have higher profitability than non-payers (Fama and French 2001). As mentioned earlier, the level of investment opportunities was seen as a factor which influences the dividend policies of firms which are listed in the US. Fama and French (2001) claimed that investment opportunities vary across dividend groups (payers, nonpayers: former-payers and never paid firms). 197 Chapter 5: Analysis of Data and Findings In this way, the investment opportunities of firms listed on the Thai stock market will be measured by using Vt/At (assets’ growth rates which is the ratio of the aggregate market value to the aggregate book value of assets), dAt/At (the rate of change of total assets, or the firms’ assets growth rate) and RDt/At (research and development or R&D) expenditures following Fama and French’s research paper on ‘disappearing dividends’. Summary statistics (descriptive statistics) and logit regression will be used to analyse the significance of the investment opportunities of firms which are listed in Thailand. Fama and French (2001) concluded that firms which have not paid dividends also have better growth opportunities than payers and former payers. Firms that have never paid dividends also grow faster than those which have not paid dividends. Former payers of dividends have poor investment opportunities, because their internal source of finance has been depleted by the payment of dividends (Fama and French 2001). The average assets size (At) and debt ratios (total liabilities to total assets) of firms listed on the US stock markets has influenced firm size (Fama and French 2001). Summary statistics were used to calculate the values of the average assets size ratio and logit regression was applied to confirm this characteristic. Fama and French (2001) found that dividends payers are much larger than former payers and non-payers, in the case of US listed firms. The propensity to pay dividends is measured by applying the average coefficient to the value of each firm’s characteristic to form a formula on profitability, investment opportunities and size (Fama and French 2001). A base period (or years) was set and the ratios a variety of groups of firms were compared in relation to their propensity to pay dividends in different time periods. This process of analysis and comparison was undertaken with the aid of logit regression (Fama and French 2001). Fama and French (2001) stated that listed firms have a lower propensity to pay dividends and this results in lower payments and decline in the number of payers. 198 Chapter 5: Analysis of Data and Findings To measure change in the characteristics of listed firms; the sample was divided into four groups: (1) dividends payers, (2) non-payers (3) former-payers, and (4) never paid (Fama and French 2001). The three important characteristics of listed firms, namely: (1) profitability, (2) investment opportunities, and (3) size, were measured and a logit regression analysis was used to test for any change in firm characteristics, or divergence between the expected number of payers and the actual number of payers (Fama and French 2001). Fama and French (2001) found that listed firms have changed their characteristics over time and this has resulted in lower dividend payment in the US stock markets. Fama and French (2001) concluded that firms which have never paid dividends are more profitable and experience stronger growth opportunities than former payers. Consequently, it can be argued that low earnings, strong investment and growth opportunities, and small size are characteristics of firms which have never paid dividends (Fama and French 2001). In addition, former payers tend to be distressed firms (Fama and French 2001). Finally, payers of dividends are large firms with high earnings and low investment opportunities (Fama and French 2001). Fama and French (2001) concluded that the characteristics and their propensity to pay dividends of listed firms have changed. The change has resulted in a disappearance of dividends and a reduction in the number of payers (Fama and French 2001). Types of Listed Firms Analysts have categorized listed firms into four groups according to their earnings. These groups include: (1) growth, (2) defensive, (3) cyclical, and (4) speculative firms (Reilly and Brown 2000). Growth Firms Growth firms generally experience ‘above-average growth in their sales and earnings’ (Reilly and Brown 2000). However, academics define this category as firms which yield rates of return which are greater than the firm’s required rate of return or weighted average cost of capital (WACC) (Reilly and Brown 2000). The sales and earnings of 199 Chapter 5: Analysis of Data and Findings growth firms expand faster than firms with different trends in the value of their earnings, although the risk associated with investing in these firms are similar. (Reilly and Brown 2000). Defensive Firms The earnings of defensive firms are likely to be maintained during a downturn (Reilly and Brown 2000). Defensive firms are also considered low risk investments. Public utilities are an example of a defensive firm (Reilly and Brown 2000). Cyclical Firms The sales and earnings of cyclical firms are strongly affected by fluctuations in the business cycle or life-cycle (Reilly and Brown 2000). This type of firm is regarded as a high risk investment. Speculative Firms Speculative firms have high risk assets (Reilly and Brown 2000). However, this type of firm may yield high returns for investors and compensate them for holding these assets (Reilly and Brown 2000). Analysts and investors have classified the listed stocks of firms into two groups (1) value and (2) growth stocks. This classification is based on how the stocks generate income, or earnings for investors (Reilly and Brown 2000). Value Stocks appear to be undervalued for reasons other than their earnings, or growth potential (Reilly and Brown 2000). Value stocks normally pay dividends and have low P/E ratios (Petty et al. 2000; Reilly and Brown 2000). Growth Stocks often experience rapid growth in their sales and earnings (examples include Microsoft and Intel) (Reilly and Brown 2000). Investors in this type of stock expect the price to appreciate (Petty et al. 2000). Firms with a high book-to-market equity ratio (B/M), earnings to price ratio (E/P), or cash flow to price ratio (C/P) are 200 Chapter 5: Analysis of Data and Findings classified as growth stocks (Lakonishok, Shleifer, and Vishny 1994; Fama and French 1995, 1998). A paper by Fama and French (1998) suggested that value stocks tend to outperform growth stocks in the capital markets around the world. This analysis also found that the markets undervalue the stocks of distressed firms and overvalue successful firms (growth stocks) (Fama and French 1998). These findings were consistent with the trends in 13 developed capital markets including the US, UK, France, Germany, Italy, the Netherlands, Belgium, Switzerland, Sweden, Australia, Hong Kong and Singapore. Fama and French (2001) concluded that the characteristics of firms which pay dividends tend to be similar to the characteristics of value stocks. However, if these characteristics were held constant, the study found that the characteristics of listed firms tended to change over time. Indeed, firm characteristics initially resembled those of value stocks but later resembled growth stocks. Some firms also ceased paying dividends. It was therefore concluded that the payment of dividends is becoming less common in the US stock markets because the characteristics of listed firms are tending to resemble those of growth stocks. The relative performance of value stocks and growth stocks appears mixed throughout the world’s capital markets. Sometimes, value stocks outperform growth stocks, while sometimes, growth stocks tend to outperform value stocks (Reilly and Brown 2000). There is no universal view on whether growth stocks perform better, or yield higher profits than value stocks. Fama and French (2001)’s study change in the characteristics of listed firms is a useful guide for investors around the world. Investors should be aware that the characteristics of firms are changing and this will aid in improving the quality of their investment decisions when choosing between value and growth stocks. A study by Benito and Young (2001) supported the findings of Fama and French (2001) in relation to the decline in the payment of dividends. This study used secondary data from the UK stock market (FTSE) which was collected during the period 1974 to 1999. It was found that the proportion of firms which do not pay dividends rose from 201 Chapter 5: Analysis of Data and Findings 14.3 percent in 1974 to 25.2 percent in 1999 (Benito and Young 2001). The study also found that the proportion of firms that were not paying dividends tended to increase during recessions. Benito and Young (2001) suggested that the increase in the proportion of firms omitting dividends was largely due to an increase in the proportion of firms that have never paid dividends. In addition, firms which paid high dividends more than doubled their volume of payments to shareholders between 1977 and 1999 (Benito and Young 2001). The study also found that low cash flows, high leverage and greater investment opportunities are associated with an increase in the propensity to omit dividends in the UK (Benito and Young 2001). This finding is consistent with the conclusions of Fama and French (2001). Benito and Young (2001) suggested that non-payers tend to be small firms with strong investment opportunities. This indicates that firms which do not pay dividends have positive prospects for growth and are not distressed. The decisions to avoid the payment of dividends was regarded as evidence of financial distress. In short, Benito and Young (2001) concluded that the characteristics of firms and the propensity to pay dividends have a large impact on the payment of dividends in the UK (Benito and Young 2001). The two previous studies concentrated on the payment of dividends by firms listed on stock markets of developed countries. Aivazian, Booth and Cleary (2003) investigated eight developing countries including South Korea, India, Malaysia, Thailand, Zimbabwe, Jordan, Pakistan and Turkey. This study used secondary data collected between 1981 and 1990, and used descriptive statistics and t-statistics to compare these emerging markets with the US. The findings indicate that firm profitability, growth and investment opportunities affect the payment of dividends of emerging markets in a similar fashion to the US (Aivazian, Booth and Cleary 2003). They added that a high ROE (which represent a firm’s profitability) results in the payment of high dividends, while high debt ratios are associated with a low payment of dividends. However the ‘sensitivity to these factors varies across countries and emerging market firms tend to 202 Chapter 5: Analysis of Data and Findings pay higher dividends than the US firms and are more sensitive to some of the variables’ (Aivazian, Booth and Cleary 2003, p. 387). These researchers, therefore, concluded that firms in emerging markets respond to the same variables as those in the US (Aivazian, Booth and Cleary 2003). However, the study did not shed light on why the payment of dividends is in decline (Aivazian, Booth and Cleary 2003). Many researchers have attempted to explain the phenomenon of disappearing dividends. Yardeni (2003) stated that the payment of dividends was important in the US prior to 1982, but has become less important over the last 20 years (Yardeni 2003). The researcher claimed that ‘if the US Congress eliminates the double taxation of dividends, then dividends should matter more again’ (Yardeni 2003 p. 14). Mr. Allan Greenspan (present Chairman of the Federal Reserve Bank, United States) observed that shareholders’ obsession with earnings is a relatively new phenomenon: ‘Prior to the past several decades, earnings forecasts were not nearly so important a factor in assessing the value of corporations. In fact, I do not recall price-to-earnings ratios as a prominent statistic in the 1950s. Instead, investors tended to value stocks on the basis of their dividend yields’ (Yardeni 2003, p. 15). The phenomenon of disappearing dividends remains a puzzle (Black 1976; Fama and French 2001, Benito and Young 2001; Aivazian, Booth and Cleary 2003). Fama and French’s (2001) findings have added to the knowledge of dividends although more research is needed to explain the pattern of dividends, payers’ characteristics and the propensity to pay dividends in developing capital markets such as Thailand. This leads to an investigation of the relationship between dividends and a given set of characteristics of listed firms and other important market factors which may explain why firms pay, or avoid paying dividends. As stated earlier, the analysis of Fama and French (2001) focused on firms which are listed in the US and decline in the number of payers and the propensity to pay dividends was evident. However, more analysis is needed on whether this is occurring in foreign markets, (Ferris, Sen and Yui 2002) particularly when the characteristics of firms and investors differ from those of the US. The issue of whether decline in the payment of dividends is an ‘international 203 Chapter 5: Analysis of Data and Findings phenomenon’ needs more research (Ferris, Sen and Yui 2002). This study will respond to this need to this topic by focusing on the emerging capital markets such as Thailand. 3.5.2 THE DISAPPEARANCE OF DIVIDENDS IN THAILAND The following trends were found after a preliminary investigation of dividends in Thailand. Figure 3.7: Number of Listed Firms Which Paid Dividends Number of Payers and Non-Payers from 1990-2002 All firms Payers 450 Number of samples 400 350 300 250 200 150 100 50 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Source: SET (2002a, 2002b) 204 Chapter 5: Analysis of Data and Findings Figure 3.8: Percentage of Dividend Payers to Total Listed Firms Dividend trend in Thai capital market Payers Percent Non-Payer 100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Source: SET (2002a, 2002b) Figures 3.7 and 3.8 indicate that the phenomenon of ‘disappearing dividends is also evident in Thailand although this market has not been studied in detail. Both Thai and international investors in the capital market of Thailand lack confidence in (1) the type of stocks to invest in and (2) the trading strategies and valuation techniques which should be applied. This investigation of the phenomenon will enhance the level of confidence in stock investments in Thailand by providing information for investors on the type of stocks to invest, appropriate trading strategies and the techniques which should be applied. A common technique for valuing stocks is the Dividend Discount Model (DDM or Gordon Model). Stock prices are calculated by discounting the expected cash dividend payments back to the present time. Therefore, the most important component of DDM is the level of the cash dividends during each period. However, if it is shown that the payment of cash dividends is disappearing, DDM may not be as effective as it has been in the past. Investors may need to apply, or develop more complex models to value stocks. Conversely, if it is shown that the payment of cash dividends is not 205 Chapter 5: Analysis of Data and Findings disappearing, DDM will continue to be supported as an accurate technique for valuing dividends by investors and analysts. 3.5.3 THAILAND IN THE LIGHT OF RESEARCH ON CAPITAL MARKET The Stock Exchange of Thailand (2003) intends to devote more funds to projects for developing quality analysis and for research into the capital markets. This funding will increase the quantity and quality of research in the securities field, enhance the level of analysis and aid in establishing standards in the industry (SET 2003). SET (2003) is encouraging Thai investors to focus more on improving their knowledge of the following: the structure of capital markets, securities trading regulations, roles of securities analysis in the capital market of Thailand, accessing information, industry and company analysis, financial ratio analysis, stock evaluation, how to provide and gather information, how to customise report presentations according to the users’ needs, critical research examinations, investment instruments, analysis ethics, and finally the Thai capital market development plan. Financial support will be granted via the ‘Securities Research Support Project’ to securities companies securities research on the industrial sectors. There are currently 12 companies which prepare research reports on 70 of the existing 600 securities which are listed in Thailand (SET 2003). This represents only 11% of the existing listed securities. Currently, most securities companies provide research reports on highly capitalized and liquid shares. Therefore, research on trends in share prices, the number of dividend payers, characteristics of payers, listed firms and the propensity to pay dividends will benefit investors, firms, the Stock Exchange of Thailand (SET) and the Securities Exchange Commission (SEC). 3.6 RESEARCH GAPS, ISSUES AND MODEL DEVELOPMENT This section focuses on the research gaps which were identified in this review of literature and will develop the research problems and hypotheses for this study. The 206 Chapter 5: Analysis of Data and Findings research issues, which are composed of research problems and hypotheses, are derived from the research gaps which were identified. 3.6.1 RESEARCH GAPS AND ISSUES 3.6.1.1 RESEARCH GAPS This review has discussed how the characteristics of firms, the propensity to pay dividends and financial ratios of listed firms affect the payment of dividends and has shed light on the phenomenon of disappearing dividends. Although several researchers of capital markets have commented on the phenomenon and the level of omission by listed firms in a variety of markets and countries, there are some limitations in these studies which are listed below: x The phenomenon of disappearing dividends was investigated by Fama and French (2001). Research on this topic is limited and requires additional analysis. x Most of the literature on this topic has focused on the capital markets of developed countries such as the US and the UK. There is a lack of empirical evidence from emerging capital markets, such as Thailand. x The relevant studies have used data from non-financial and non-utility (industrial) firms. Financial institutions were excluded because due to variation in accounting styles and the large level of assets and liabilities of these firms. This study will assess the financial sector. The findings of earlier studies may not apply to all listed companies. As stated above, the samples which were analysed were limited. The gaps in the literature which will be examined include: x The need for additional research on disappearing dividends and the omission of dividend payments in developing countries. x This will be the first analysis of the phenomenon of disappearing dividends or in Thailand. 207 Chapter 5: Analysis of Data and Findings x This study will be the first analysis of all sectors in the capital market of Thailand. 3.6.1.2 RESEARCH ISSUES The research issues depend on the research gaps which have been identified by the review of literature and research problem. After identifying the research issues, a list of research questions and hypotheses will be developed for analysing these issues. The Research Problem The research problem is restated as follows: ‘How do the characteristics of publicly traded firms and their propensity to pay dividends influence the payment of dividends and the reasons for their disappearance from the stock market of Thailand?’ This study focuses on the characteristics of listed firms and their propensity to pay dividends, therefore, the information outlined on firms is summarised by a set of financial variables which are termed firm characteristics. The characteristics which were suggested by Fama and French (2001) have been adopted for this study. These characteristics include: (1) profitability (2) investment opportunities and (3) size. The methodology to measure change in these characteristics and their impact and the propensity to pay dividends will be discussed in more detail in chapter 4. Research Questions and Hypotheses Several research questions and hypotheses have been developed to answer the research problem. The research questions and hypotheses are presented in figure 3.9. 208 Chapter 5: Analysis of Data and Findings Figure 3.9: Research Questions and Hypotheses RESEARCH PROBLEM ‘How do the characteristics of publicly traded firms and RESEARCH HYPOTHESES RESEARCH QUESTIONS Q1: What is the background and performance of the Stock Exchange of Thailand and its listed firms? their propensity Q2: How have researchers investigated to pay dividends dividends and their patterns throughout influence the payment of dividends and the reasons for their the world? Q3: 9 Research Hypotheses What is methodology for the appropriate collecting and analysing data? disappearance Q4: Does the data from Thailand support from the stock the model? market of Thailand?’ Q5: How could the model be implemented? Source: Developed for this research The four research questions are answered in this study. In addition, research question 4 (Q4) contains four sub-questions (Q4.1, Q4.2 and Q4.3) are stated as follows: Research Questions #4.1: What are the characteristics of the dividend payers? As stated earlier, Fama and French (2001) suggested that the characteristics of listed firms have changed over time and the number of firms paying dividends in the US stock markets has fallen. This study will consider the phenomenon of disappearing dividends in Thailand by responding to hypotheses 1 to 3 and the impact of changes in the characteristics of listed firms. 209 Chapter 5: Analysis of Data and Findings Fama and French (2001) used summary statistics to analyse the characteristics of firms listed on the US stock markets. These firms were divided into four groups: (1) Dividend payers, (2) Non-payers, (3) Former payers and (4) Never-paid firms. Three characteristics were identified to explain the pattern of dividends and actions, namely: (1) profitability, (2) size and (3) investment opportunities. Logit regression analysis was employed to test the significance of these characteristics. This research will investigate whether the characteristics of firms listed in the developing economy of Thailand are similar to those listed on the US capital markets. These hypotheses are stated as follows: Research Hypothesis # 1: Ho : There is no significant relationship between the profitability of listed firms and the probability that these firms will pay dividends. Ha : There is a significant relationship between the profitability of listed firms and the probability that these listed firms will pay dividends. Research Hypothesis # 2: Ho : There is no significant relationship between the size of listed firms and the probability that these firms will pay dividends. Ha : There is a significant relationship between the size of listed firms and the probability that these firms will pay dividends. Research Hypothesis # 3: Ho : There is no significant relationship between the investment opportunities of listed firms and the probability that these firms will pay dividends. Ha : There is a significant relationship between the investment opportunities of listed firms and the probability that these firms will pay dividends. SET (2002) found that almost 70 percent of listed firms made negative earnings and only 20 percent of listed firms paid dividends during the Asian economic crisis. There could be a relationship between the Asian economic crisis, the characteristics of listed firms and the number of dividend payers in Thailand’s capital market. Hypotheses # 1 210 Chapter 5: Analysis of Data and Findings to 3 will be tested during three time periods, namely Step A, all firms from 1990 to 1996 (pre-crisis), Step B, all firms from 1997 to 2002 (post-crisis), Step C, all firms from 1990 to 2002 (total outlook). Figure 3.10 shows the linkage between each of the 9 research hypotheses and the 5 research questions. Research Questions #4.2: Do these characteristics change over time? As stated earlier, Fama and French (2001) suggested that the characteristics of listed firms have changed over time and the number of dividend payers in the US stock markets has fallen. This study will consider the phenomenon of disappearing dividends in Thailand and the impact of changes in the characteristics of listed firms by responding to hypotheses 1 to 3. Research Question #4.3: How has the propensity to pay dividends of firms changed over time? As stated earlier, Fama and French (2001) also found that many firms are not paying dividends although they have sufficient earnings. This outcome is known as a lower propensity to pay dividends. The propensity to pay dividends will be addressed by responding to hypotheses 1 to 3. This will lead to an answer to Research Question 4.3. 211 Chapter 5: Analysis of Data and Findings Figure 3.10: Research Questions in Relation to Hypotheses CHARACTERISTICS OF DIVIDEND PAYERS BEFORE THE ASIAN ECONOMIC CRISIS (Step A) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends before the Asian Economic Crisis. CHARACTERISTICS OF THE DIVIDEND PAYERS DURING AND AFTER THE ASIAN ECONOMIC CRISIS (Step B) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends after the Asian Economic Crisis. CHARACTERISTICS OF THE DIVIDEND PAYERS FOR THE WHOLE PERIOD (Step C) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends for the whole period of investigation. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends the whole period of investigation. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividends the whole period of investigation. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividends the whole period of investigation. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends the whole period of investigation. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends the whole period of investigation. Source: Developed for this research 212 Chapter 5: Analysis of Data and Findings 3.6.2 DEVELOPMENT OF THE RESEARCH MODEL Figure 3.11 summarises the concepts which will be investigated and their linkage to the payment of dividends in the Thailand’s capital market and the proposed methodology for this study. Figure 3.11: Model for this Research Summary Statistics Logit Regression Number of Payers, Non-Payers and Former Payers Cash Dividends Characteristics (Profitability, Investment Opportunity and Size) Characteristics (Profitability, Investment Opportunity and Size) (Confirmed) Changes in Characteristics Changes in Characteristics (Confirmed) Changes in Propensity to Pay Dividends Changes in Propensity to Pay Dividends (Confirmed) Source: Developed for this research As illustrated by figure 3.11, the phenomenon of disappearing dividends in Thailand will be explained by summary statistics and logistic regression. The characteristics of listed firms, their propensity to pay dividends and the concentration of dividends will firstly be identified by the summary statistics and confirmed by logit regression analysis. The significance of each factor will be analysed with logit regression analysis and this will answer the research hypotheses, questions and problems which were developed for this study. 213 Chapter 5: Analysis of Data and Findings The characteristics of firms, the propensity to pay dividends and the concentration of dividends and market factors are discussed in chapter 4. All terms and definitions will also be defined for this research in chapter 4. A discussion of summary statistics and logistic regression analysis and why they represent the best methodology for this study will also be presented in Chapter 4. 3.7 CONCLUSION This chapter reviewed the key theoretical issues associated with finance, capital market, investment, dividend policies and the phenomenon of disappearing dividends. This review of literature has shed light on the issues associated with the phenomenon, efficient market hypothesis, financial ratio analysis, stock valuation and value and growth stocks. The discussion also focused on the existence of the phenomenon in the capital markets of the developed economies and indicated that the issue could be present in Thailand. Finally, a model was developed to illustrate how the research problem will be answered. The methodology of this study will be discussed in detail in Chapter 4. 214 Chapter 5: Analysis of Data and Findings C HAPTER 4.1 4 RESEARCH METHODOLOGY INTRODUCTION Chapter 3 reviewed the literature on the phenomenon of disappearing dividends, the characteristics of listed firms, the propensity to pay dividends, dividend policy and capital market theory. This chapter discusses the methodology and data used for investigating whether the phenomenon of disappearing dividends is present in Thailand. This methodology will investigate the characteristics of listed firms, changes in these characteristics over time and the propensity to pay dividends of firms. Figure 4.1 illustrates the methodology of this study and the structure of this chapter. Section 4.2 discusses the research problem and how it was defined. The research design of this study is explained in section 4.3. The techniques used for categorizing and confirming data are presented in section 4.4. This section also presents the model which has been developed for this study. Section 4.5 presents a plan for analysing the data. Finally, section 4.6 is a conclusion to this chapter. 215 Chapter 5: Analysis of Data and Findings Figure 4.1: The Structure of Chapter Four 4.1 Introduction 4.2 Problem Discovery and Definition 4.3 Research Design 4.3.1 Purpose of the Study 4.3.2 Selecting Design the 4.3.3 Selecting the Research Technique Research 4.4 Data Classification and Methodology 4.4.1 Data Availability and Period of Study 4.4.2 Summary Statistics 4.4.3 Logit Regression 4.5 Plan for Data Analysis 4.6 Conclusion Source: Developed for this research 216 Chapter 5: Analysis of Data and Findings 4.2 PROBLEM DISCOVERY AND DEFINITION The first step in a business research project is to define the problem. ‘Problem definition is a specific area that will be clarified by answering a number of research questions or hypotheses’ (Zikmund 2000, p. 88). The nature of the research problem will therefore influence important parts of the research, including the research design, selection of samples, data gathering, data analysis and evaluation (Black 1999; Zikmund 2000). Figure 4.2 presents the research problem and issues of this study, the research hypotheses and methodology developed for responding to the research problems and questions. Figure 4.2: Links between the Research Problem, Research Questions, Research Hypotheses and Methodology RESEARCH PROBLEM ‘How do the characteristics of publicly traded firms and their propensity to pay dividends influence the payment of dividends and the reasons for their disappearance from the stock market of Thailand?’ RESEARCH HYPOTHESES RESEARCH QUESTIONS RESEARCH METHODOLOGY RESEARCH DESIGN Q1: What is the background and performance of the Stock Exchange SAMPLING DESIGN of Thailand and its listed firms? Q2: How have researchers investigated dividends and their patterns throughout the world? Q3: What is the 9 Research Hypotheses DATA COLLECTION METHODS appropriate methodology for collecting and analysing data? DATA ANALYSIS TECHNIQUES Q4: Does the data from Thailand support the model? Q5: How could a model be CONCLUSIONS AND REPORTS implemented? Source: Developed for this research 217 Chapter 5: Analysis of Data and Findings 4.3 RESEARCH DESIGN The research design of a study is a plan, road map, or framework specifying the methods and procedures for collecting and analysing the information which is required for answering a research problem (Emory 1995; Zikmund 2000). The determination of appropriate sources of information, the design technique and sampling methodology of a research project are carried out in the research design process (Zikmund 2000). 4.3.1 PURPOSE OF THE STUDY There are two main forms of business research, namely, basic research and applied research (Zikmund 2000; Cavana, Delahaye and Sekaran 2001). Basic Research Basic research is conducted for the purpose of theory development. It involves building a body of knowledge which is refined by attempting to explain, or understand the problems occurring in organisations. It does not directly involve solving a particular problem (Zikmund 2000; Cavana, Delahaye and Sekaran 2001). Applied Research Applied research is undertaken for applying or testing the existence of a theory and for evaluating how well this theory describes, or solves current business problems (Zikmund 2000; Cavana, Delahaye and Sekaran 2001). It is possible that organisations will apply knowledge from basic research to solve specific problems. Therefore, basic and applied research may be adopted for solving a business research problem (Cavana, Delahaye and Sekaran 2001). Basic and applied research need to be ‘undertaken in a scientific manner’ (Cavana, Delahaye and Sekaran 2001, p. 13). However, applied research is more popular than basic research because it often involves a shorter period of investigation and lower costs than basic research (Cavana, Delahaye and Sekaran 2001). 218 Chapter 5: Analysis of Data and Findings This study is an example of applied research which will answer questions to specific problems or situations. It will examine whether dividends are disappearing from the Thai capital market and considers changes in the characteristics of listed firms and their propensity to pay dividends. 4.3.2 SELECTING THE TYPE OF RESEARCH DESIGN Business research can be classified on the basis of technique or function (Zikmund 2000). The nature of the research problem will determine the research design. According to Zikmund (2000), there are three types of research designs, namely: (1) exploratory research, (2) descriptive research, and (3) causal research. Exploratory research is the initial research which is conducted to clarify and define the nature of a problem (Zikmund 2000). This research is conducted when the researchers lacks a clear idea of the problems that will be met during the course of the study (Emory and Cooper 1991). Exploratory research aids researchers or organisations to develop the concepts more clearly, establish priorities and improve the research design especially when the area of investigation is new, or vague (Emory and Cooper 1991). Exploratory research also helps to save time and money. Researchers and managers are often urged to obtain prompt answers to research questions (Emory and Cooper 1991). Descriptive research seeks to identify and describe, but not to explain a phenomenon (Ticehurst and Veal 2000). It helps in determining answers to who, what, when, where and how questions (Zikmund 2000). Descriptive studies are based on a previous understanding or knowledge of the nature of the research problem or phenomenon (Zikmund 2000). The goal of Causal research is to identify, explain, understand and predict the relationship between the cause and effect of variables (dependent and independent variables) (Zikmund 2000). 219 Chapter 5: Analysis of Data and Findings Emory and Cooper (1991) discussed John Stuart Mill’s method of agreement, which states ‘when two or more cases of a given phenomenon have one and only one condition in common, that condition may be regarded as the cause (or effect) of the phenomenon’ (Emory and Cooper 1991, p. 151). In short, causal research attempts to prove when one action is undertaken, another action will follow. After considering the research problem which was stated in section 4.2, the research questions and hypotheses will investigate the relationship between dividends and (1) the characteristics of listed firms and (2) the propensity to pay dividends of firms. The relationships which are identified will aid in determining whether dividends are disappearing from Thailand’s capital market. These relationships will be of a cause-andeffect type. They will indicate whether changes in firm characteristics and the propensity to pay dividends of listed firms have led to a disappearance of dividends from Thailand’s capital market. An analysis of these relationships is an example of ‘causal research’ and this study is therefore classified as casual research. 4.3.3 SELECTING THE RESEARCH TECHNIQUE A research design (exploratory, descriptive and causal research) is conducted with the aid of research methods and techniques. These techniques include surveys, experiments, observation and the analysis of secondary data (Zikmund 2000). 220 Chapter 5: Analysis of Data and Findings Figure 4.3: Alternatives for the Basic Research Method Problem Discovery and Definition Selection of Basic Research Method Survey: (1) Interview (2) Questionnaire Experiment Observation Secondary Data Source : Emory and Cooper (1991); Zikmund (2000), p. 55 A Survey is commonly used for collecting primary data. There are two main techniques for conducting a survey, namely, interviews and questionnaires. The most important part of developing a survey is preparing a questionnaire, or questions which will be asked at an interview. The process includes preparing questions, determining the list of questions, and designing the format and structure of the questionnaire interview. The wording of the survey questions (avoiding the use of jargon) is very important because it will ensure the response is accurate, unbiased and useful (Clover and Balsley 1984; Ticehurst and Veal 2000; Zikmund 2000). An experiment is an investigation of a cause-and-effect relationship. Experiments deal with variables, which are controlled, whilst others are not controlled (Zikmund 2000). An observation is an unobtrusive technique (Emory 1995) which records, or gathers information on the objects, or behaviour which is being studied, without intervening with these objects, or people or letting them know that they are being observed or recorded (Ticehurst and Veal 2000, Zikmund 2000). 221 Chapter 5: Analysis of Data and Findings The study of secondary, or historical data, involves the use of data which was collected by earlier researchers for a reason which may be related to the current research topic. The secondary data technique also includes the use of historical data which is generally available to the public. Secondary data techniques can be applied in all types of research (exploratory, descriptive and causal research) (Clover and Balsley 1984; Emory and Cooper 1991; Zikmund 2000). The data which is needed for investigating the phenomenon of disappearing dividends in Thailand’s capital market exits is secondary form. Useful secondary data is contained in the financial statements of listed firms, SET’s annual report, BOT’s annual report and the financial press. This information includes cash dividend payments, earnings, assets, liabilities, interest expenses, earnings before interest and taxes, interest rates, market capitalisation rates, and finally, data on the volume of trade. An analysis of recording data is an appropriate technique for this study because it focuses on the cause-effect relationship between dividends, their disappearance, the characteristics of firms and the propensity to pay dividends. All information required for the analysis is readily available in secondary form and at minimal cost. Therefore, the secondary data approach was considered appropriate research method for this study. 4.4 DATA CLASSIFICATION AND METHODOLOGY This section discusses the availability of data, the period of study and the techniques used for explaining the phenomenon of disappearing dividends. Subsection 4.4.1 identifies the dependent and independent variables, listed firms are categorized into several groups and the three characteristics of these groups are also discussed. Subsection 4.4.2 and 4.4.3 discuss the summary statistics and logit regression in detail. 222 Chapter 5: Analysis of Data and Findings 4.4.1 DEPENDENT AND INDEPENDENT VARIABLES The data used in this study measures cash dividend payments, the characteristics of firms and their propensity to pay dividends. Zikmund (2000, p. 91) defines a variable as ‘anything that may make differences in numerical values’. The term dependent and independent variables are frequently used in statistical analysis to describe cause-effect relationships (Ticehurst and Veal 2000). This section identifies and defines the variables that are used in this study. Emory and Cooper’s (1991) illustration of how to define dependent and independent variables is summarised in figure 4.4. Figure 4.4: Defining Independent and Dependent Variables Independent Variable Dependent Variable Presumed Cause….. Presumed Effect….. Stimulus….. Response….. Predicted from….. Predicted to….. Antecedent….. Consequence….. Manipulated….. Measured Outcome….. Source: Emory and Cooper (1991) The independent variable is the presumed cause, predicted from, manipulated, antecedent and it stimulates the dependent variable. The dependent variable is the presumed effect, predicted to, measured outcome, consequence, and it responds to the independent variable (Emory and Cooper 1991; Page and Meyer 2000). 223 Chapter 5: Analysis of Data and Findings Dependent Variable A dependent variable ‘is a criterion, or a variable that is to be predicted or explained’ (Zikmund 2000, p. 91). Therefore, in this study the dependent variable is the percentage of dividend payers in the Thai capital market. Independent Variable(s) An independent variable ‘is a variable that is expected to influence the dependent variable’ (Zikmund 2000, p. 92). In this study, the independent variables are the characteristics of listed firms (profitability, investment opportunities and sizes) and their propensity to pay dividends. 4.4.2 DATA AVAILABILITY AND PERIOD OF STUDY The data used in this study is obtained from the 1998, 1999, 2000, 2001 and 2002 Listed Company Info (CD-ROMs) which is prepared by the Department of Investor Relations and Communication (The Stock Exchange of Thailand), SIMS (The Stock Exchange of Thailand Information Management System), and SET’s annual reports from 1990 to 2002. SIMS provides trading and company information of listed firms in Thailand in a Microsoft Excel compatible format. SET’s records of trading activities and company data is available on CD-ROM from 1990 till the present. The year 1990 is the starting period for this study because SET’s computerised data base was only established in 1990. Consequently, the data which was collected before 1990 is incomplete and incompatible due to a different Securities Exchange Act applying to the disclosure of information for listed firms. After the repealing of the SEC Act, B.E. 2517 (1974) and the SEC Act II, B.E. 2527 (1984), the SEC Act B.E. 2535 (1992) is the most recent securities and exchange act. Although adjustments have been made to this act (the SEC Act B.E. II 2542 (1999), III 2546 (2003)), it still applies to all listed firms from 1992 onward. In addition, the data for this study was collected in mid 2003. Consequently, the most recent information which was available was found on the annual financial statements for 2002. 224 Chapter 5: Analysis of Data and Findings As a result, 2002 was chosen as the final year for the focus period of this study. Therefore the sample period for this study will be 1990 – 2002. As discussed earlier, Thailand has been affected by the Asian Economic Crisis of 1997, and this study will examine the effect of this crisis. The time frame of this study is divided into three periods, or steps; Step A: pre-crisis; Step B: post-crisis and; Step C: the whole period. The ‘pre-crisis’ period covers data from 1990 to 1996 (seven years). The ‘post-crisis’ period lasts from 1997 to 2002 (six years). The ‘whole period’ covers data from 1990 to 2002 (thirteen years). Differences in the relevant variables will be analysed over these periods. The data used for this study will also be analysed according to the time frames which were discussed above. As a result: Step A, represents firms from 1990 to 1996 (pre-crisis) Step B, represents firms from 1997 to 2002 (post-crisis) Step C, represents firms from 1990 to 2002 (all times, total outlook) The analysis of descriptive statistics and logistic regression will also use this classification. Given the study’s time frame, the number of firms in the sample is indicated by table 4.1. Table 4.1: Number of Samples in this Study Counts of Sample SET Firms All Firms 1990 209 1991 228 1992 254 1993 288 1994 328 1995 344 1996 378 1997 393 1998 388 1999 385 2000 368 2001 376 2002 386 Source: SET (2002b) 225 Chapter 5: Analysis of Data and Findings Firms that are included in the sample must provide full information on their financial statements (income statement and balance sheet) and trading information. The sample contains financial data of firm (1990 to 2002) by calendar year on the following: Total assets, average stock prices, shares outstanding, income before extraordinary items, interest expenses, dividend yields, dividend per share, preferred dividends, preferred stock value par value at the end of year t. All firms in the sample have shareholder’s equity, total liabilities and common equity. However, the figures for total assets in year t and t-1 permit the calculation of changes to total assets as a ratio. This ratio represents the investment opportunity of a firm. Other items used for time t. A number of firms have provided information on balance sheet deferred taxes, investment tax credits and income statement deferred taxes. However, this information is not required in this study. The firms included in the sample are found on SET’s database (SIMS) at the end of each fiscal year. This ensures that the stocks of these firms are publicly traded in the market. A firm must also have market equity data for December of year t, which includes average prices, shares outstanding and the market capitalization rate. The sample firms are categorized into 31 sectors, according to the nature of business; (1) Agribusiness (2) Banking (3) Commerce (4) Communication (5) Construction, building materials and furniture (6) Electronic components (7) Electrical and computer parts (8) Energy (9) Entertainment (10) Finance (11) Food and Beverage (12) Health Care Services (13) Hotels (14) Insurance (15) Jewelry (16) Mining (17) Packaging (18) Pharmaceutical (19) Printing (20) Professional Services (21) Property (22) Pulp and Paper (23) Textiles (24) Transportation (25) Vehicles (26) Warehouse (27) Household 226 Chapter 5: Analysis of Data and Findings (28) Machinery (29) Chemicals (30) Others (31) Rehabilitation. The listed firms include Thai and foreign firms. Classification of Dividend Groups As stated earlier, the dependent variable is the percentage of dividend payers from 1990 to 2002. Therefore, it is necessary to define the term ‘dividend payer’ or ‘payer’ so that each firm can be classified in an appropriate group. A firm which is listed at SET is classified as: Payer, in a calendar year t, if the firm had positive dividends per share (Fama and French 2001). Non-payer, if the firm has not paid dividends in time t (Fama and French 2001). An all firms listing includes payers and non-payers, where payers pay dividends in year t and non-payers do not (Fama and French 2001). Former payer is a firm that does not pay in year t but paid dividends in year t-1 (previous year) or in a previous year (Fama and French 2001). Never paid is a firm that has not paid dividends in time t-1 and t or in a previous year (Fama and French 2001). New list, if the firm is added SET’s database between January and December of t (Fama and French 2001). 227 Chapter 5: Analysis of Data and Findings Dividends are observed at the end of each year (31st December). This is an identical time period as the factors which determine firm characteristics and the propensity to pay dividends (profitability, investment opportunities and size). The study period commences after the full financial statements of the listed firms have been published and recorded on SIMS (The Stock Exchange of Thailand’s Information Management System). As stated earlier, the number of firms in each category will be summed and calculated as a percentage of total N, with a table of summary statistics, while t represents each calendar year from January 1990 to December 2002. Figure 4.5 illustrates the process for calculating the percentage of payers, non-payers, never paid firms and former payers. Figure 4.5: Process of Counts and Percentages of SET Firms in SIMS in Time t Listed Firms in Time t All firms (N) (Seasoned and New lists) Payers (n, % of N) Non Payers (n, % of N) Never Paid (n, % of N) Former Payers (n, % of N) Source: Developed for this research 4.4.3 PROPOSED METHODOLOGY 4.4.3.1 THE INCIDENCE OF DIVIDEND PAYMENT BY THAI FIRMS This study will identify whether the percentage of listed firms which pay dividends is decreasing in Thailand. 228 Chapter 5: Analysis of Data and Findings As stated earlier, the pattern of dividend behaviour in Thailand will be observed by grouping the listed firms into the following: (1) dividend payers, (2) non-payers, (3) former payers, or (4) never paid firms. New lists which commenced trading in time t will be identified as newly listed firms and will be grouped as, newly listed firms which pay dividends or newly listed firms which do not pay dividends (Fama and French 2001). The summary statistics, or descriptive statistical techniques, are subgroup statistics for the variables (Tabachnick and Fidell 2001). The descriptive statistics include the sum, number of cases, mean, median, grouped median, standard error of the mean, minimum value, maximum value, the range, variable value of the last category of the grouped variable, standard deviation, variance, kurtosis, standard error of kurtosis, skewness, standard error of skewness, percentage of total sum, percentage of total N, percentage of sum in, percentage of N in, geometric mean and the harmonic mean (Saengkaew 1996; Bryman and Cramer 1999, 2001; Tabachnick and Fidell 2001). Frequencies and means are the descriptive statistics which are used most frequently. Frequencies include the sum and percentage of each variable, while the mean is the average value of a total population, or samples of variables (Ticehurst and Veal 2000). The dividend policies of firms will also be considered by determining whether firms continue, or cease paying dividends. The study will investigate whether the payers in time t continue to pay dividends in subsequent years, or whether they cease paying and become non-payers. In addition, the study will observe whether non-payers of period t commence distributing dividends in the next period, or continue being non-payers (Fama and French 2001). In short, the study uses descriptive statistics to summarise the behaviour of firms which are categorised as payers and non-payers in time t-1 and the current period (t). Both payers and non-payers in time t-1 are classified according to whether they pay, or do not pay dividends in time t. These firms will continue to pay, stop paying, continue to be a non-payer, start paying or will be delisted from SET. 229 Chapter 5: Analysis of Data and Findings The dividend policy of payers (at time t-1) and non-payers (at time t-1) in time t are expressed as follows: Continue to pay: Firms that continue paying dividends in time t and t-1 (Fama and French 2001). Start paying: Firms that commence paying dividends in time t but were non-payers in t-1 (Fama and French 2001). Stop paying: Firms that were payers in time t-1 but cease paying dividends in time t (Fama and French 2001). Continue to be non-payer: Firms that were non-payers in time t-1 and continue to be non-payers of dividends in the current period (t) (Fama and French 2001). Delisted firms: Firms that were payers, or non-payers in time t-1 and were delisted in time t (Fama and French 2001). Figure 4.6 classifies the behaviour of listed firms by their dividend policy at time (t-1) and time t. 230 Chapter 5: Analysis of Data and Findings Figure 4.6: Payers and Non-payers in Time t-1 All firms (N) Payers (t-1) Non Payers (t-1) Continue to Pay (t) Start to Pay (t) Stop Paying (t) Do Not Pay (t) Merge (t) Merge (t) Delist (t) Delist (t) Source: Developed for this research The percentage of firms in each category will be calculated for each year between 1990 and 2002 using summary statistical methods (descriptive statistics approach). The method for presenting the number of listed firms in each group will be similar to the analysis of Fama and French (2001) (see table 4.1). 4.4.3.2 THE CHARACTERISTICS OF DIVIDEND PAYERS AND NON-PAYERS A firm’s dividend policy is related to its profitability, investment opportunities and size (Lintner 1963; James, Dodd and Kimpton 1985; Rappaport 1986; Fama and French 1988, 1995, 1997, 1999, 2000, 2001; Fisher and Jordan 1991; Penman and Sougiannis 1998; Haugen 1997; Hurley and Johnson 1997; Yao 1997; Bringham, Gapenski and Ehrhardt 1999; Brealey and Myers 2000; Reilly and Brown 2000; Melicher and Norton 2000; Petty et al. 2000; Gitman, Juchau and Flanagan 2002; Wetherilt and Weeken 2002). This study categorises the characteristics of the sample firms with the aid of several financial 231 Chapter 5: Analysis of Data and Findings variables (Fama and French 2001). As stated earlier, this analysis will observe whether dividend payers, non-payers, former payers, and never paid firms differ according to (1) profitability, (2) investment opportunities, and (3) size. The results will be compared over the thirteen year period (1990-2002). The characteristics are measured in the following ways: Profitability: Two ratios are adopted as indicators of a firm’s profitability namely: the ratio of aggregate earnings before interest to aggregate assets (Et/At) and the ratio of aggregate common stock earnings over aggregate book equity (Yt/BEt). The components of these ratios are discussed below: For Profitability, the indicator is Et/At Et/At is the ratio of aggregate earnings before interest to aggregate assets, or the amount of earnings before interest generated by a single dollar of assets (aggregate asset) (Fama and French 2001). Where: Et : Earnings before interest for fiscal year t Earnings Before Interest (Et) = Earnings Before Extraordinary Items + Interest Expense + Income Statement Deferred Taxes (if available) At : Assets at the end of fiscal year t Total Assets (A t) = Short-term and Long-term assets For Profitability, the additional indicator is Yt/BEt 232 Chapter 5: Analysis of Data and Findings Yt/BEt is the ratio of aggregate common stock earnings divided by the aggregate book equity, or the value of earnings from common stock which is generated by a single dollar of the firm’s book equity (Fama and French 2001). Where: Yt : After-tax earnings to common stock for fiscal year t Earnings Available for Common Stock (Yt) = Earnings Before Extraordinary Items – Preferred Dividends (If available) + Income Statement Deferred Taxes (If available) BEt : Book common equity at the end of fiscal year t Book Equity (BEt) = Stockholder’s Equity (or common equity + preferred stock par value or asset – liabilities) – Preferred Stock + Balance Sheet Deferred Taxes and Investment Tax Credit (if available) – Post Retirement Asset (if available) Investment Opportunities: The two ratios which are used as indicators of investment opportunities include the following: the ratio of aggregate market value to the aggregate book value of assets (Vt/At) and the ratio of the change in last year’s assets in baht to this year’s total assets (dAt/At). The components of these ratios are discussed below: For Investment opportunities, the indicator is Vt/At Vt/At is the ratio of aggregate market value to the aggregate book value of assets (Fama and French 2001). Where: Vt : Total market value at the end of fiscal year t Market Value of Firm (Vt ) = Assets – Book Equity + Market Equity At : Assets at the end of fiscal year t Total Assets (At) = Short-term and Long-term assets 233 Chapter 5: Analysis of Data and Findings For Investment opportunities, the additional indicator is dAt/At dAt/At is the ratio of change in the asset value (in baht) of this year, compared to last year divided by the total asset value (in baht) of this year. It is also the rate of growth of assets of the firm (Fama and French 2001). Where: At : Assets at the end of fiscal year t Total Assets (At) = Short-term and Long-term assets dAt : Change in value of assets (in baht) Change in Assets (dAt) = Time t assets in baht – Time t-1 assets in baht Size is measured by a firm’s aggregate assets (At) and the ratio of aggregate liabilities to aggregate assets (Lt/At). The components of this ratio are discussed below: For Size, the indicator is At At : Assets at the end of fiscal year t (Fama and French 2001). Where: Total Assets (At) = Current assets and Long-term assets For Size, the indicator is Lt/At Lt/At is the ratio of aggregate liabilities to aggregate assets, or the value of liabilities created from a single baht of assets (Fama and French 2001). 234 Chapter 5: Analysis of Data and Findings Where: Lt : Liabilities at the end of fiscal year t Total Liabilities (Lt) = Short-term and Long-term liabilities At : Assets at the end of fiscal year t Total Assets (At) = Short-term and Long-term assets The information required for classifying the characteristics of listed firms suggested by Fama and French (2001) include: Earnings Before Interest: are earnings on a firm’s assets in Thai baht at the end of every fiscal year from 1990 to 2002. The earnings before interest is found on the income statements of firms listed on SET’s database (SIMS). Total Assets at year end for every listed firm is found on the published balance sheets which are available on SIMS. After-Tax Earnings to Common Stock is a ratio of available earnings to common stocks and it is an important factor which influences the dividend policy of firms. Book Value of Common Equity, data for every fiscal year is available on SIMS. Total Market Value of listed firms is also available on SIMS from 1990 to 2002. The Et, dAt, At, Yt, BEt, At,Vt, MEt and Lt items are also presented for each year and are used for calculating the percentage average aggregate values of firms that pay dividends. The discussion on descriptive statistics concludes with a summary report which identifies the percentage of each dividend group (payers, non-payers, former payer, non-payer, new listed) in Thailand’s capital market (1990 to 2002) and the average (mean) of each ratio which has been used as an indicator of a firm’s characteristics (profitability, investment 235 Chapter 5: Analysis of Data and Findings opportunities and size). This will include a three-stage discussion of the relative changes in the size of the dividend groups before and after the Asian economic crisis (1997), and during the total period of investigation (1991 to 2002). The figure 4.7 indicates how the results for the three characteristics will be presented. Figure 4.7: Average Firm Profitability, Investment Opportunities and Size Listed Firms Characteristics Profitability Et/At, Yt/BEt Investment Opportunities dAt/At Vt/At All Firms Payers Size At Lt/At All New Listed Non-Payers Former Payers Payers Non-Payers Never Paid Source: Developed for this research After calculating the mean values of each ratio and comparing the results for each group according to the period scenarios, an ANOVA will be calculated for each year between 236 Chapter 5: Analysis of Data and Findings 1991 and 2002. ANOVA is the analysis of variance which is used for confirming that the means of each group are significantly different (Mason, Lind and Marchal 1999; Ticehurst and Veal 2000; Tabachnick and Fidell 2001). ANOVA is a statistical tool which will confirm that the mean of the three characteristics for each group is statistically different. An F-test is used to express the level of significance of the variation in the means. The variation will be tested at the significance level of 0.05 (Mason, Lind and Marchal 1999; Tabachnick and Fidell 2001). Two ANOVA tests will be conducted for this study. The first ANOVA test will be for payers and non-payers, and will confirm whether characteristics of payers differ from non-payers. The second test will be conducted for payers and the two sub-groups of nonpayers (former payers and never paid firms). This test will confirm whether the characteristics of payers differ from those of former payers and never paid firms. The Stock Exchange of Thailand (SET) and the Bank of Thailand (BOT) will provide information on fluctuations in the SET index, economic data, trends and forecasts, benchmarks and other market indicators for this research. 4.4.3.2.1 THE EFFECTS OF PROFITABILITY, INVESTMENT OPPORTUNITIES AND FIRM SIZE To assess the effects of firm characteristics (profitability, investment opportunities, and size) on the payment of dividends, Fama and French (2001) suggested that an estimation of logistic regression is appropriate. Logistic regression (hereafter logit regression) ‘allows prediction of group membership, when predictors are continuous, discrete, or a combination of the two’ (Tabachnick and Fidell 2001, p. 24). Thus, ‘it is an alternative to both discriminant function analysis and logit analysis’ (Tabachnick and Fidell 2001, p. 24). Logit regression ‘allows one to evaluate the odds (or probability) of membership in one of the groups based on the combination of values of the predictor variables’ (Tabachnick and Fidell 2001, p. 24). 237 Chapter 5: Analysis of Data and Findings Therefore ‘Logit regression documents more formally the marginal effects of profitability, investment opportunities and size on the likelihood that a firm pays dividends’ (Fama and French 2001, p. 12). The results of logit regression are expected to confirm the inferences on the role of profitability, investment opportunities and firm size in influencing a firm’s dividend policies (Fama and French 2001). The hypotheses developed in chapter 3 shall be tested and answered using the t-statistics provided with the logit regression. Fama and MacBeth (1973, p. 621) stated that ‘using t-statistics in the usual way does not lead to serious errors’. This research applies t-statistics to test the significance of the predetermined hypotheses. As stated earlier the independent variables are profitability, investment opportunities and size. Fama and French (2001) defined each characteristic as follows: Probability is measured by the Et/At (Fama and French 2001) or earnings before interest but after taxes standardized by total assets. Investment opportunities are measured by dAt/At and Vt/At (Fama and French 2001) or the growth rate in assets and the market-to-book ratio. Firm size is measured by the percentage of SET firms that have the same or smaller market capitalization as the specific firms (Fama and French 2001). Following Fama and French (2001), the logit regression is divided and presented in two sets. The first set is the logit regression which controls for two investment variables, 238 Chapter 5: Analysis of Data and Findings namely, dAt/At and Vt/At. The second set is the logit regression control for the investment variable dAt/At. In both sets, the remaining characteristics are unchanged. Table 4.2: Comparison of Different Ratios Applied in Different Stages of this Research Characteristics Approach (1) Descriptive Statistic Profitability Investment Opportunities Size Et/At Yt/BEt dAt/At Vt/At At Lt/At Approach (2) Logit Regression Logit Regression (controlling 2 (controlling 1 investment variables) investment variable) Et/At Et/At dAt/At Vt/At At dAt/At At Source: Developed for this research Later, logit regression is also used to estimate the probability that firms will pay dividends. It will therefore allow the researcher to analyse and confirm the effects of a firm’s characteristics on its dividend policy. In addition, logit regression tests whether the findings of Fama and French (2001), Benito and Young (2001) and Aivazian, Booth and Cleary (2003) could be applied for a developing capital market such as Thailand. Figure 4.8 indicates how the characteristics of firms are confirmed by logit regression analysis. 239 Chapter 5: Analysis of Data and Findings Figure 4.8: How Logit Regressions is Used for Explaining the Dividend Policies of Firms Average Coefficient SETt Size Vt/At t-statistic dAt/At Investment Et/At Profitability Source: Developed for this research Using t-statistics, each characteristic is confirmed as significant (<0.05) or nonsignificant (>0.05) (Tabachnick and Fidell 2001) variable in predicting dividend behaviour of the listed firms. T-statistic values indicate the strength of relationship between each characteristic and the likelihood that a firm will pay dividends (Tabachnick and Fidell 2001). 4.4.3.3 CHANGING PROPENSITY TO PAY DIVIDENDS Fama and French (2001) found that the characteristics of firms have changed and this has affected the payment of dividends. As stated earlier, they also suggest that firms are less likely to pay dividends if these characteristics are controlled over time. This phenomenon was entitled ‘the lower propensity to pay dividends of the listed firms’. The methodology adopted to test the propensity to pay dividends has several steps as suggested by Fama and French (2001). The first step is to apply descriptive statistics by categorizing listed firms into four groups, namely, (1) firms with positive earnings (Yt > 0), (2) firms with negative earnings (Yt < 0), (3) firms with earnings above investment opportunities (Et > dAt) and (4) firms with earnings below investment opportunities (Et < 240 Chapter 5: Analysis of Data and Findings dAt) (Fama and French 2001). This is to investigate the trend in the propensity to pay dividends of listed firms from 1990 to 2002. The next step is to estimate the probability that firms with constant characteristics (profitability, investment opportunities and size) paid dividends between 1990 and 2002. Later, the probabilities which were calculated for the base years 1991 to 1996 or before the Asian Economic Crisis will be applied to the sample firm’s characteristics. This will forecast the expected percentage of dividend payers from 1991 to 2002. The probabilities of the characteristics are fixed at their base year (1991 to 1996). Therefore, any variation in the number of expected payers will result from changes in the characteristics of the sample firms (Fama and French 2001). The actual percentage of payers must be calculated for each year from 1991 to 2002. The difference between the expected percentage of payers and the actual percentage of payers will be an estimate of the change in the propensity to pay dividends by listed firms between 1991 and 2002. A decline in the propensity is indicated by the difference between the expected percentage of payers and actual percentage of payers. Figure 4.9 below indicates how the number of total firms and payers is counted and how the proportion of payers in the sample is expressed as a percentage. 241 Chapter 5: Analysis of Data and Findings Figure 4.9: Logit Regression Estimates of the Effect of Changing Firm Characteristics and Decline in the Propensity to Pay Dividends Firms Actual Percent Payers Expected Percent Expected Actual Logit Regression forecasted N n n/N x 100 Characteristics changes Propensity to pay Source: Developed for this research Note: N is total number of firms, n is number of payers in time t The Total number of firms (N) and payers (n) are counted. Then the actual percentage of payers to total firms is found and compared to the expected percentage of payers forecasted by logit regression. The results show the changes in characteristics over time. Lastly, the differences between expected and actual percentage of payers in the market determine the changes in propensity to pay dividends of listed firms. Therefore the propensity to pay dividends would be discussed using descriptive statistics and logit regression and the changes in the characteristics will also be formally identified in this stage. 4.5 PLAN FOR DATA ANALYSIS This section explains how the data will be analysed after the descriptive statistics and logit regression have been conducted. In particular, this section discusses the data processing, screening and analysis techniques which are applied in this research. 242 Chapter 5: Analysis of Data and Findings 4.5.1 DATA PROCESSING Data processing has two main functions, namely, editing and coding of the raw data (Zikmund 2000). In this study, the raw data is the published financial statements of firms listed on the Thailand stock exchange. This data must be converted into an organised form which can be used for answering the research problem. If the database is large, a computerized data processing program is needed to arrange, analyse and interpret the meaning of the data. In this study, the raw data is presented on the financial statements as dividend payments and firm characteristics. The information which is selected will be entered onto spreadsheets with the aid of Microsoft EXCEL and SPSS (The Statistical Package for Social Sciences) software. This will ensure the data can be analysed effectively. 4.5.2 VALIDITY Validity refers to the extent that information collected in a research study, actually reflects the phenomenon in the real world (Ticehurst and Veal 2000, Zikmund 2000). There are two types of validity (1) internal validity which addresses whether the independent variables are related to the dependent variables and (2) external validity which addresses whether the scope of this research is applicable to the real world (Ticehurst and Veal 2000; Zikmund 2000). The ratios which estimate the importance of each firm characteristic have readily been used by professionals and academics and are regarded as sound measures of profitability, investment opportunities and firm size. The methodology is also widely accepted in the field of finance throughout the world (Lintner 1963; Fama and Babiak 1968; Fama and MacBeth 1973; James, Dodd and Kimpton 1985; Rappaport 1986; Fama and French 1988, 1995, 1997, 1999, 2001; Fisher and Jordan 1991; Hurley and Johnson 1994; Haugen 1997; Yao 1997; Bringham, Gapenski and Ehrhardt 1999; Brealey and Myers 243 Chapter 5: Analysis of Data and Findings 2000; Melicher and Norton 2000; Petty et al. 2000; Reilly and Brown 2000; Gitman, Juchau and Flanagan 2002; Wetherilt and Weeken 2002). The terms, categories, ratios and methodologies are commonly used by researchers in the field of finance and investment. 4.6 CONCLUSION This chapter discussed the research design and methodology which will be used in this study. A range of statistical techniques for investigating the disappearance of dividends have been discussed. Chapter 5 provides the results of the data analysis which is used for explaining the significance of the phenomenon of disappearing dividends in Thailand. 244 Chapter 5: Analysis of Data and Findings C HAPTER 5.1 5 ANALYSIS OF DATA AND FINDINGS INTRODUCTION Chapter 4 discussed the methodology and the plan for investigating the disappearing dividends phenomenon in the capital market of Thailand. This chapter applies the methodology and data together to answer the research problem, questions and hypotheses developed in chapter 3. Chapter 5 presents the results of the descriptive statistics and logistic regression. In particular, significant variables are identified which demonstrate the characteristics of listed firms which pay dividends, changes in these characteristics over time and the propensity to pay dividends of these firms. This chapter also discusses the findings of the hypothesis tests which were developed in chapter 3, and the implications of this study for Thailand’s capital market. The chapter is presented in eight sections as indicated by Figure 5.1. Section 5.2 defines the descriptive statistics and provides the sample with these descriptions. The initial findings of the descriptive statistics and a discussion of the phenomenon of disappearing dividends in Thailand are provided in section 5.3. Section 5.4 outlines the trends identified in the payment of cash dividends. Section 5.5 analyses the characteristics of the dividend groups with the aid of ANOVA and logic regression. Section 5.6 analyses the propensity to pay dividends of listed firms with descriptive statistics. Section 5.7 uses logit regression to test the significance of the findings in relation to the propensity to pay dividends of firms and change in the characteristics of firms. Finally, section 5.8 is a conclusion to this chapter. 245 Chapter 5: Analysis of Data and Findings Figure 5.1: The Structure of Chapter Five 5.1 5.2 5.2.1 Introduction Descriptive Statistics Descriptive Statistics Definitions 5.2.2 A Description of the sample 5.3 An Initial analysis of Descriptive Analysis 5.4 Time Trends in Cash dividend 5.4.1 Payers and their dividends yield 5.5 5.5.1 5.2.3 5.4.2 Characteristics of dividend payers Firm Characteristics (Descriptive Statistics Approach) 5.5.2 ANOVA test 5.5.4 Confirmation from Logit Regressions Review of Descriptive Statistics Results 5.5.6 5.5.5 Dividend Pattern Hypothesis Testing (1) Average coefficient 5.6 5.7 Hypothesis Testing (2) Group coefficient The Propensity to pay dividends (Descriptive approach) Changes in characteristics and propensity to pay (Logit Regressions approach) 5.8 Conclusion Source: Developed for this research 246 Chapter 5: Analysis of Data and Findings 5.2 DESCRIPTIVE STATISTICS This section describes the descriptive statistics which are used in this tool in the study. 5.2.1 DESCRIPTIVE STATISTICS DEFINITIONS Statistics is the science of collecting, organising, presenting, analysing, and interpreting data (Mason, Lind and Marchal 1999, p. 3). In business, statistics is often used for improving the quality of investment decisions (Mason, Lind and Marchal 1999). Statistics can be presented in two ways, namely: (1) numerically and (2) graphically (Mason, Lind and Marchal 1999). The statistical findings of this study, in both of these forms is discussed in the later part of this chapter. There are two main types of statistics: (1) inferential statistics and (2) descriptive statistics. The objective of inferential statistics is to observe a phenomenon from a sample of a population (Mason, Lind and Marchal 1999). Descriptive statistics are used for organising, summarising and presenting data in a simple, but informative way (Mason, Lind and Marchal 1999; Ticehurst and Veal 2000; Zikmund 2000). Descriptive statistics enable a researcher to better understand a set of data before undertaking more advanced statistical analysis (Huck, Cormier and Bounds 1974). Frequencies and means are two of the most common descriptive statistics. Frequencies show the ‘counting and percentages for each variable’ while the means is an average of the total population or sample (Ticehurst and Veal 2000, p. 181). 5.2.2 A DESCRIPTION OF THE SAMPLE As discussed in chapter 4, the sample used in this study consists of firms listed on the Stock Exchange of Thailand between 1990 and 2002. Listed firms are classified into thirty-one sectors, or industries. The names of these industries are as follows: (1) Agribusiness (2) Banking (3) Commerce (4) Communication (5) Construction, building materials and furniture (6) Electronic components (7) Electrical and computer 247 Chapter 5: Analysis of Data and Findings parts (8) Energy (9) Entertainment (10) Finance (11) Food and Beverage (12) Health Care Services (13) Hotels (14) Insurance (15) Jewelry (16) Mining (17) Packaging (18) Pharmaceutical (19) Printing (20) Professional Services (21) Property (22) Pulp and Paper (23) Textiles (24) Transportation (25) Vehicles (26) Warehouse (27) Household (28) Machinery (29) Chemicals (30) Others (31) Rehabilitation. An objective of this study is to identify and explain the pattern of cash dividends from 1990 to 2002. As discussed earlier, this study will explore the effect of the Asian Economic Crisis on the payment of dividends. The study focuses on three periods of time, (1) Step A, pre-crisis (2) Step B, post-crisis and (3) Step C, all times (1990-2002). The ‘pre-crisis’ period contains data from 1990 to 1996 and represents the period before the Asian Economic Crisis. The ‘post-crisis’ is represented by the period 1997 to 2002 and finally the ‘all times’ period focuses on the period 1990 to 2002. The data will be classified according to the framework and will be analysed with descriptive statistics and logistic regression. A number of listed firms were deleted from the sample because their financial statements were incomplete. SEC requires that the financial statement of listed firms should be reviewed by an auditor of SET. However, as discussed in chapter 4, a number of listed firms provided insufficient information on their financial statements and have not been audited. If a firm’s financial statement has not been provided, it will be forced to cease trading (a ST, or trading suspended sign will be placed on its securities). If a firm fails to provide an annual financial statement for 2 years, or only provides a quarterly financial statement, it will be delisted. The Minister of Finance (MOF) and the Bank of Thailand (BOT) forced a number of financial institutions to cease operating, or present them with plans for a merger or acquisition, during the Asian Economic Crisis. Some of these financial institutions survived a while, some closed. In short, the financial statements and trading information of firms which closed are incomplete. To ensure the analysis is accurate these firms have been deleted from the sample. 248 Chapter 5: Analysis of Data and Findings 5.3 AN INITIAL ANALYSIS OF DESCRIPTIVE STATISTICS The percentage of listed firms paying cash dividends in Thailand (SET) fell from 84.2 percent in 1990, to 46.4 percent in 2002, a decline of about 40 percent (table 5.1). As stated earlier in chapter 3, the payment of dividends has also declined in the US (Fama and French 2001). These researchers suggested that this trend resulted from a change in the characteristics and the propensity to pay dividend of firms and their results are supported by a number of studies (La Porta et al. 2000; DeAngelo, DeAngelo and Skinner 2002; Baker and Wurgler 2003a). To explain the phenomenon disappearing dividends in Thailand, samples of listed firms from 1990 to 2002 have been collected. This section compares the results of this study with Fama and French (2001). These results include changes in the characteristics and the propensity to pay dividends of listed firms in developed and emerging capital markets. 5.4 TIME TRENDS IN CASH DIVIDENDS Generally, dividends are taxed at a higher rate than capital gains and they are less valuable for investors (Fama and French 2001). Similar taxation policies are applied in Thailand because dividends are taxed at 10 percent, although a tax credit of 3/7 is received at the end of the year. Consequently, the real tax rate of dividends is 5.71 percent in Thailand. Capital gains are currently exempt from taxation (Cholviroj 1989; SEC 2002). As discussed in chapter 3, firms paying dividends appear to be disadvantaged (they experience a higher tax rate and low investment opportunities) as their equity costs are higher than those of firms that do not pay dividends. However, some firms continue to pay dividends and investors and academics have found it difficult to explain this puzzle (Fama and French 2001). This section focuses on trends in the payment of dividends between 1990 and 2002. The sample firms are classified as payers, or non-payers of dividends in the present 249 Chapter 5: Analysis of Data and Findings time (t). As explained in chapter 4, payers and non-payers are classified according to whether they paid dividends in the past period (t-1) and present period (t). Table 5.1: Counts and Percent of SET firms in Different Dividend Groups Counts of SET firms All Firms New Lists 1990 209 1991 228 45 1992 254 38 1993 288 33 1994 328 38 1995 344 23 1996 378 34 1997 393 4 1998 388 3 1999 385 4 2000 368 11 2001 376 7 2002 386 20 Percent of SET firms Payers Non-Payer Former-Payer Never-Paid New Lists 1990 84.2 15.8 1991 84.6 15.4 10.1 5.3 19.7 1992 90.6 9.4 5.1 4.3 15.0 1993 85.1 14.9 8.0 6.9 11.5 1994 83.8 16.2 9.5 6.7 11.6 1995 84.3 15.7 12.8 2.9 6.7 1996 81.0 19.6 14.6 5.0 9.0 1997 71.5 28.5 22.9 5.6 1.0 1998 28.4 71.6 65.2 6.4 0.8 1999 29.9 70.1 62.9 7.3 1.0 2000 35.3 64.9 56.5 8.4 3.0 2001 40.7 59.3 48.7 10.6 1.9 2002 46.4 53.6 40.2 13.5 5.2 82.2 73.7 51.5 52.6 82.6 73.5 50.0 33.3 0.0 27.3 57.1 20.0 New Lists that Pay Note: Payers paid dividends in time t while non-payers did not. Non-payers are dividend into two subgroups, namely, former payers (firms that do not pay in time t but paid in the previous year) and firms that have never paid dividends. New lists are firms which were listed in time t and new lists that paid dividends are classified as payers. Source: SET (1990, 1993, 1996, 1999, 2002a) Patterns in the payment of dividends are explored from 1990 to 2002 with statistics from the SIMS database. In 1990, 84.2 percent of listed firms paid dividends (see table 5.1). The percentage of payers rose to its peak of 90.6 percent in 1992 and fell gradually to 71.5 percent in 1997. The percentage of payers fell to 28.4 percent in 1998. This figure was recorded a year after the Asian crisis. However, the percentage of firms paying dividends recovered slightly in 1999 (29.9 percent) and rose to 46.4 percent by 2002. However, the percentage of firms paying dividends in 2002 (46.4 percent) was well below the 1990 level (84.2 percent). To better understand changes in the pattern of dividend payments, the sample firms were classified by industry according to SET’s classification. Table 5.2 shows the number of firms in each industry group from 1990 to 2002. 250 Chapter 5: Analysis of Data and Findings Table 5.2: Number of Sample Firms in Each Industry 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 AGRIBUSINESS Number of firms in each industry 11 15 17 24 28 27 28 26 24 21 21 20 20 BANKING 15 14 10 10 10 9 10 11 10 13 14 14 14 BUILDING AND FURNISHING MATERIALS 13 19 21 24 25 27 30 33 29 25 23 24 17 CHEMICALS AND PLASTICS 5 7 8 8 10 13 15 15 14 13 13 12 12 COMMERCE 9 9 12 12 14 13 14 15 14 14 12 13 14 1 2 3 5 9 9 10 10 10 10 10 11 12 13 2 10 10 11 11 10 7 22 42 43 51 49 ELECTRICAL PRODUCTS AND COMPUTER 6 7 8 9 10 11 12 13 11 9 11 8 10 ELECTRONIC COMPONENTS 2 2 2 3 4 6 8 8 8 8 7 8 8 ENERGY 3 3 3 5 7 8 9 9 9 9 10 10 11 COMMUNICATION COMPANIES UNDER REHABILITATION 1 1 1 1 2 3 6 7 7 8 7 7 10 FINANCE AND SECURITIES 24 18 13 13 14 17 20 21 20 21 19 21 26 FOODS AND BEVERAGES 15 15 16 18 20 20 21 24 24 23 21 22 22 HEALTH CARE SERVICES 4 6 8 9 11 11 13 13 13 13 11 10 10 10 11 11 12 12 12 12 12 12 12 12 12 10 2 4 6 8 8 8 8 8 8 6 7 7 17 18 19 19 19 20 22 22 22 21 21 21 ENTERTAINMENT AND RECREATION HOTELS AND TRAVEL SERVICES HOUSEHOLD GOODS INSURANCE JEWELRY AND ORNAMENTS 13 2 MACHINARY AND EQUIPMENT MINING OTHERS PACKAGING PHARMACEUTICAL PRODUCTS AND COSMETICS 1 4 4 4 4 4 4 4 3 2 2 2 2 1 2 2 3 4 6 6 5 4 4 3 3 1 1 1 1 1 1 1 1 1 1 1 1 1 2 4 4 5 5 5 6 5 4 5 5 5 11 13 12 13 17 17 17 17 15 14 14 12 13 2 2 2 2 2 2 2 2 2 2 2 2 2 PRINTING AND PUBLISHING 4 6 7 8 8 7 10 10 10 8 7 8 8 PROFESSIONAL SERVICES 1 2 2 2 2 2 2 2 2 2 2 2 2 PROPERTY DEVELOPMENT 6 11 15 21 25 28 32 37 35 26 21 22 29 3 2 2 2 2 3 5 5 5 5 5 4 4 24 22 25 27 28 28 27 28 25 25 24 24 24 TRANSPORTATION 2 3 3 4 5 6 7 7 8 8 8 8 8 VEHICLES AND PARTS 3 6 6 6 8 9 10 10 10 8 8 8 9 WAREHOUSE AND SILO 4 4 4 4 4 4 4 4 5 5 4 4 4 PULP AND PAPER TEXTILES, CLOTHING AND FOOTWEAR Source: SET (1990, 1993, 1996, 1999, 2002a) Table 5.2 indicates that the number of listed firms has increased in most sectors since 1990. The highest growth has been in rehabilitation, electronic components, energy, entertainment and recreation, transportation vehicles and parts, household goods and property development. The number of firms in these industries more than tripled between 1990 and 2002. To comply with the research problem stated in chapter 1, trends in the number of dividend payers in each industry group are analysed. 251 Chapter 5: Analysis of Data and Findings Table 5.3: Number of Dividend Payers in Each Industry Number of payer in each industry 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 9 12 17 20 25 23 22 20 9 11 13 18 16 BANKING 11 13 9 9 10 9 10 9 7 0 0 0 0 BUILDING AND FURNISHING MATERIALS 12 16 18 21 18 19 24 19 2 2 2 3 6 CHEMICALS AND PLASTICS 3 3 8 8 9 11 9 10 3 5 6 6 7 COMMERCE 9 8 12 12 13 13 13 13 4 6 7 7 8 COMMUNICATION 0 2 3 3 5 6 6 8 2 1 0 2 3 COMPANIES UNDER REHABILITATION ELECTRICAL PRODUCTS AND COMPUTER 9 2 5 4 5 5 4 1 0 0 0 0 0 6 7 7 8 9 9 10 9 3 3 4 4 7 ELECTRONIC COMPONENTS 2 2 1 2 4 5 6 7 2 3 2 3 4 ENERGY 3 3 3 2 5 6 7 7 4 5 6 5 8 ENTERTAINMENT AND RECREATION 1 1 1 1 2 3 5 7 4 3 3 3 3 FINANCE AND SECURITIES 21 17 13 12 14 17 19 18 2 4 6 4 7 FOODS AND BEVERAGES 13 11 15 15 15 18 18 19 11 12 13 13 13 HEALTH CARE SERVICES 4 6 8 8 10 10 11 10 2 1 4 5 6 HOTELS AND TRAVEL SERVICES 7 11 9 10 11 10 10 9 3 5 7 9 9 2 4 5 6 7 5 8 4 3 1 4 4 11 16 16 17 17 17 19 19 14 14 13 17 17 2 4 4 4 4 4 3 2 1 1 1 1 2 0 2 2 3 4 6 6 1 1 1 1 1 0 AGRIBUSINESS HOUSEHOLD GOODS INSURANCE JEWELRY AND ORNAMENTS MACHINARY AND EQUIPMENT MINING 1 1 1 1 1 0 0 0 0 0 0 0 OTHERS 0 2 1 2 3 3 3 5 2 2 3 4 3 PACKAGING PHARMACEUTICAL PRODUCTS AND COSMETICS 7 10 11 11 12 14 13 11 4 4 5 4 7 1 2 1 2 2 2 2 2 0 1 1 1 1 PRINTING AND PUBLISHING 4 5 7 8 8 6 8 8 4 4 5 5 5 PROFESSIONAL SERVICES 1 2 2 2 2 2 2 2 2 2 2 2 2 PROPERTY DEVELOPMENT 5 9 12 15 20 25 25 16 1 2 1 3 6 2 2 2 2 2 2 4 2 0 0 0 2 2 23 13 25 24 24 23 23 19 11 15 15 16 17 TRANSPORTATION 2 3 3 3 4 5 6 4 2 0 1 1 4 VEHICLES AND PARTS 3 6 6 6 8 8 9 7 3 2 5 6 7 WAREHOUSE AND SILO 4 2 4 4 4 4 4 4 3 3 3 4 4 PULP AND PAPER TEXTILES, CLOTHING AND FOOTWEAR Source: SET (1990, 1993, 1996, 1999, 2002a) Table 5.3 indicates that the number of firms which pay dividends expanded in most industry sectors between 1990 and 2002. The number of payers declined in eight industries. These industries include agribusiness, banking, building, commerce, rehabilitation, finance, mining and textiles. As discussed earlier, table 5.2 indicates that the number of listed firms increased between 1990 and 2002. Therefore, it is not surprising that the number of payers in each industry has also increased over this period. However, table 5.4 presents the percentage of firms which pay dividends in each industry as a standard measurement of 252 Chapter 5: Analysis of Data and Findings the increase, or decrease in the proportion of dividend payers in each industry over time. Table 5.4: Percentage of Payers to Total Number of Listed Firms in Each Industry Percent of payers to total number of firms in each industry 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 AGRIBUSINESS 81.8 80.0 100.0 83.3 89.3 85.2 78.6 76.9 37.5 52.4 61.9 90.0 80.0 BANKING BUILDING AND FURNISHING MATERIALS 73.3 92.9 90.0 90.0 100.0 100.0 100.0 81.8 70.0 0.0 0.0 0.0 0.0 92.3 84.2 85.7 87.5 72.0 70.4 80.0 57.6 6.9 8.0 8.7 12.5 35.3 CHEMICALS AND PLASTICS 60.0 42.9 100.0 100.0 90.0 84.6 60.0 66.7 21.4 38.5 46.2 50.0 58.3 100.0 88.9 100.0 100.0 92.9 100.0 92.9 86.7 28.6 42.9 58.3 53.8 57.1 0.0 100.0 100.0 60.0 55.6 66.7 60.0 80.0 20.0 10.0 0.0 18.2 25.0 45.5 45.5 40.0 14.3 0.0 0.0 0.0 0.0 0.0 COMMERCE COMMUNICATION COMPANIES UNDER REHABILITATION ELECTRICAL PRODUCTS AND COMPUTER 69.2 100.0 50.0 40.0 100.0 100.0 87.5 88.9 90.0 81.8 83.3 69.2 27.3 33.3 36.4 50.0 70.0 ELECTRONIC COMPONENTS 100.0 100.0 50.0 66.7 100.0 83.3 75.0 87.5 25.0 37.5 28.6 37.5 50.0 ENERGY ENTERTAINMENT AND RECREATION 100.0 100.0 100.0 40.0 71.4 75.0 77.8 77.8 44.4 55.6 60.0 50.0 72.7 100.0 100.0 100.0 100.0 100.0 100.0 83.3 100.0 57.1 37.5 42.9 42.9 30.0 FINANCE AND SECURITIES 87.5 94.4 100.0 92.3 100.0 100.0 95.0 85.7 10.0 19.0 31.6 19.0 26.9 FOODS AND BEVERAGES 86.7 73.3 93.8 83.3 75.0 90.0 85.7 79.2 45.8 52.2 61.9 59.1 59.1 HEALTH CARE SERVICES 100.0 100.0 100.0 88.9 90.9 90.9 84.6 76.9 15.4 7.7 36.4 50.0 60.0 70.0 100.0 81.8 83.3 91.7 83.3 83.3 75.0 25.0 41.7 58.3 75.0 90.0 0.0 100.0 100.0 83.3 75.0 87.5 62.5 100.0 50.0 37.5 16.7 57.1 57.1 HOTELS AND TRAVEL SERVICES HOUSEHOLD GOODS INSURANCE JEWELRY AND ORNAMENTS 84.6 94.1 88.9 89.5 89.5 89.5 95.0 86.4 63.6 63.6 61.9 81.0 81.0 100.0 100.0 100.0 100.0 100.0 100.0 75.0 50.0 33.3 50.0 50.0 50.0 100.0 33.3 0.0 0.0 100.0 100.0 100.0 100.0 100.0 100.0 20.0 25.0 25.0 33.3 MINING 100.0 100.0 100.0 100.0 100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 OTHERS 0.0 100.0 25.0 50.0 60.0 60.0 60.0 83.3 40.0 50.0 60.0 80.0 60.0 63.6 76.9 91.7 84.6 70.6 82.4 76.5 64.7 26.7 28.6 35.7 33.3 53.8 50.0 MACHINARY AND EQUIPMENT PACKAGING PHARMACEUTICAL PRODUCTS AND COSMETICS 50.0 100.0 50.0 100.0 100.0 100.0 100.0 100.0 0.0 50.0 50.0 50.0 PRINTING AND PUBLISHING 100.0 83.3 100.0 100.0 100.0 85.7 80.0 80.0 40.0 50.0 71.4 62.5 62.5 PROFESSIONAL SERVICES 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 PROPERTY DEVELOPMENT 83.3 81.8 80.0 71.4 80.0 89.3 78.1 43.2 2.9 7.7 4.8 13.6 20.7 PULP AND PAPER TEXTILES, CLOTHING AND FOOTWEAR 66.7 100.0 100.0 100.0 100.0 66.7 80.0 40.0 0.0 0.0 0.0 50.0 50.0 95.8 59.1 100.0 88.9 85.7 82.1 85.2 67.9 44.0 60.0 62.5 66.7 70.8 TRANSPORTATION 100.0 100.0 100.0 75.0 80.0 83.3 85.7 57.1 25.0 0.0 12.5 12.5 50.0 VEHICLES AND PARTS 100.0 100.0 100.0 100.0 100.0 88.9 90.0 70.0 30.0 25.0 62.5 75.0 77.8 WAREHOUSE AND SILO 100.0 50.0 100.0 100.0 100.0 100.0 100.0 100.0 60.0 60.0 75.0 100.0 100.0 Source: SET (1990, 1993, 1996, 1999, 2002a) Table 5.4 indicates that the percentage of payers have declined in 22 out of 31 industries. The number of listed dividend payers has therefore declined in most industries in Thailand. 253 Chapter 5: Analysis of Data and Findings 5.4.1 PAYERS AND THEIR DIVIDEND YIELD SET publishes data on the annual dividend yield of listed firms that paid dividends in time t. The dividend yield is the total common stock dividend of the last 12 months divided by the market capitalisation of the stock and multiplied by 100 (SET 2002a). During the study period (1990 to 2002), the average market dividend yield has fluctuated due to changes (decreased and increased) in the number of payers and stock prices. Figure 5.2 shows the pattern of the annual average yields for the sample firms between 1990 and 2002. However, there could be some bias when using the average dividend yield for monitoring the pattern of dividend yields of common stocks in Thailand. The dividend yield is the annual cash dividend in time t divided by the rate of market capitalisation in Thai baht (SET 2003). This is not a standardised tool because firms could have a different number of common stocks outstanding and stock prices will make the dividend yield unstable. The Weighted Average Dividend Yield could be more accurate because it is the dividend yield multiplied by the percentage of market capitalisation rate of a firm to total market capitalisation and divided by 100. Figure 5.2 shows trends in the weighted average yield of dividends. 254 Chapter 5: Analysis of Data and Findings Figure 5.2: Annual Average Dividend Yields and the Pattern in the Yield of the Weighted Average Dividend Average Dividend Yield and Weighted Average Dividend Yield Average Dividend Yield Weighted Dividend Yield 18.00% 16.00% 14.00% Percent 12.00% 10.00% 8.00% 6.00% 4.00% 2.00% 0.00% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Source: SET (1990, 1993, 1996, 1999, 2002a) Figure 5.2 indicates that the yield of the weighted average dividend fluctuated less than the yield of the arithmetic average dividend, especially, during the Asian Economic Crisis. It is also observed that the market dividend yield for the average dividend yield and the weighted dividend yield rarely exceeded 4 percent after 1990. Considering this, it is interesting to explore the range of the dividend yield paid to shareholders. SET noted there are four ranges for the yield of dividends (1) lower than 3 percent (2) 3 to 7 percent (3) 7 to 12 percent and (4) greater than 12 percent (SET 2003). Table 5.5 presents the percentage of firms that paid dividends in the four ranges between 1990 and 2002. 255 Chapter 5: Analysis of Data and Findings Table 5.5: Annual Dividend Yield by Range Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 <3% 40.91% 38.86% 46.09% 58.89% 47.31% 30.34% 22.55% 12.81% 24.55% 20.00% 7.69% 13.55% 15.08% Dividend Yield (Annual) 3%-7% 7%-12% 43.18% 14.77% 49.74% 8.29% 46.52% 6.52% 35.57% 4.74% 41.22% 10.39% 47.59% 18.97% 36.60% 29.41% 21.35% 22.78% 39.09% 18.18% 32.17% 34.78% 33.85% 36.92% 42.58% 32.90% 55.87% 23.46% >12% 1.14% 3.11% 0.87% 0.79% 1.08% 3.10% 11.44% 43.06% 18.18% 13.04% 21.54% 10.97% 5.59% Source: SET (1990, 1993, 1996, 1999, 2002a) However, this method which separates the sample firms into four groups based on the yield on dividends does not reveal a clear trend in dividend yields in the market. Therefore, by taking the mid point of the four ranges, table 5.6 presents only two ranges of dividend yields and the percentages of payers according to their new and wider ranges (lower than 7 percent and higher than 7 percent). The results are illustrated in figure 5.3. Table 5.6: Percentage of Payers in Two Ranges of Dividend Yield Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Dividend Yield (Annual) <7% >7% 84.09% 15.91% 88.60% 11.40% 92.61% 7.39% 94.47% 5.53% 88.53% 11.47% 77.93% 22.07% 59.15% 40.85% 34.16% 65.84% 63.64% 36.36% 52.17% 47.83% 41.54% 58.46% 56.13% 43.87% 70.95% 29.05% Source: SET (1990, 1993, 1996, 1999, 2002a) 256 Chapter 5: Analysis of Data and Findings Figure 5.3: Trends in the Dividend Yield 1990-2002 Percent of Dividend Payers with reported dividend yield <7% Percent of Payers >7% 100.00% 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Source: Developed for this research Figure 5.3 shows that more than 80 percent of the dividends of listed firm’s yield less than 7 percent before the crisis. However, during the crisis, the percentage of payers with high yield dividends increased sharply to 65.85 percent and this figure fluctuated strongly throughout the remainder of the study period. However, the percentage of firms with low yield dividends has increased sharply since 2000. In 2002, over 70 percent of firms paid dividends with a yield of less than 7 percent. Low yields may result from the following: (1) the payment of high dividends by large and mature firms could appear small when compared to their stock price, therefore, dividend yields could be small or (2) listed firms really make small dividend payments which are unrelated to the size of the firm. As discussed earlier, the dividend yield is related to market capitalisation, or change in the price of stocks. The unstable pattern of dividends could be explained by observing the percentage change in stock prices during the period of investigation. 257 Chapter 5: Analysis of Data and Findings Table 5.7: Average Dividend Yield, Average Weighted Dividend Yield, Average Percentage Changes in Security Prices and Percentage of Market Capitalisation of Payers to Total Firms. 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Average Dividend Yield 3.69% 3.72% 3.26% 2.73% 3.26% 4.40% 5.73% 16.64% 2.00% 2.22% 3.17% 2.97% 2.78% Average Weighted Dividend Yield 3.53% 3.32% 2.45% 1.32% 1.54% 1.78% 2.81% 4.89% 1.86% 0.53% 1.53% 1.92% 2.50% Average Percentage Changes in Security Prices -26.01% -4.44% 8.15% 25.47% -24.37% -19.71% -24.39% -44.25% 6.94% 51.27% -15.33% 33.72% 27.31% Percentage of Market Cap. Of Payers to Total Firms 89.78% 91.00% 92.89% 79.74% 81.39% 85.46% 84.37% 88.02% 53.37% 13.68% 28.35% 43.71% 62.33% Source: SET (1990, 1993, 1996, 1999, 2002a) With a similar number of shares outstanding, the lower the stock price, the higher the dividend yield. Therefore, the high payment of dividends in 1997 could be a direct result of low stock prices due to the Asian Economic Crisis. This view is consistent with the trend, which is shown in table 5.7. The table indicates that the average percentage change in security prices was largest in 1997 (-44.25 percent). Therefore, in 1997 the average dividend yield was 16.64 percent and the average weighted dividend yield was 4.89% (figure 5.2) and these figures were the highest recorded during the study period. The Asian economic crisis influenced the dividend yields by making the figures unstable over time but how the crisis influenced the characteristics, the propensity to pay dividends of payers and the phenomenon of disappearing dividends is to be investigated in the later sections. 258 Chapter 5: Analysis of Data and Findings This section compares trends in the payment of dividends and dividend. This analysis explores the extend at which dividends are disappearing from the capital market of Thailand. The study aims to explain changes in the pattern of dividend payments before, during and after the Asian Economic Crisis. The changes in the characteristics of listed firms and the propensity to pay dividends are explored in the following analysis. 5.4.2 DIVIDEND PATTERN Figure 5.4 shows the total number of listed firms between 1990 and 2002 that (1) pay dividends (2) do not pay dividends (3) formerly paid dividends and (4) have never paid dividends. Figure 5.5 shows the percentage of the sample of listed firms which are presented in each of the four dividend groups. Figure 5.4: The Number of SET Firms in Each Group Number of SET firms in different dividend groups All firms Payers Non-Payer Former Payers Never-Paid Number of SET firms 450 400 350 300 250 200 150 100 50 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Source: Developed for this research 259 Chapter 5: Analysis of Data and Findings Figure 5.5: Percentage of All SET Firms in Different Dividend Groups Percent of all SET firms in different dividend groups Percent Payers Non-Payer Former-Payer Never-Paid 100 90 80 70 60 50 40 30 20 10 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Source: Developed for this research Figure 5.4 indicates that the number of listed firms increased sharply from 200 firms in 1990 to 400 firms in 1996 (pre-crisis). Table 5.1 indicates that the number of listed firms in the sample declined to below 400 immediately after the crisis. This figure has recovered slightly and stood at 397 firms in 2002. However, current number of listed firms is still below the figure for 1996 and 1997. The number of firms paying dividends increased from almost 200 in 1990 to greater than 300 firms in 1996. However, the rate of increase in the number of payers has declined. The number of firms which pay dividends has partly recovered although it was still below 200 firms in 2002. There are four possible explanations for the decline in the number of firms paying dividends in the short-term and the long-term. Firstly, existing and new listed firms exhibit a lower propensity to pay dividends, or are less willing to pay dividends. Secondly, firms have recently listed with the characteristics of non-payers tend to avoid paying dividends. Thirdly, the characteristics of mature firms appear to have changed and many of these firms have ceased paying dividends. Fourthly, many firms have been distressed since the crisis (both new lists and existing firms) and ceased paying dividends. This will be discussed more in detail later in the study. 260 Chapter 5: Analysis of Data and Findings The number of firms which did not pay dividends before the crisis (the figure was less than 100) was consistently less than the number of payers. However, the number of firms which did not pay dividends rose sharply from 112 in 1997, to a peak of 280 in 1998. This figure has declined slowly since 1998, although it remained slightly above 200 in 2002. Fama and French (2001) suggested that firms which do not pay dividends can be classified into the two groups, former payers and firms which have never paid dividends. Former payers are firms that do not pay dividends in the present year (time t) but paid dividends in a previous year. Figure 5.4 indicates that the number of former payers (before and after the crisis) is similar to the number of non-payers. In addition, the number of firms which have never paid dividends was relatively small in 1990. The number of non-payers and former payers increased sharply in 1997 and peaked in 1998 (see figure 5.4). This increase in the number of non-payers and former payers can be attributed to the Asian Economic Crisis. Indeed, a strong decline in the number of firms which pay dividends also appears to result from the crisis. The number of non-payers declined from about 300 in 1998 to about 200 in 2002. In addition, the number of former payers declined from 250 in 1998 to about 150 in 2002. A trend that has emerged since the crisis is that the number of firms that have never paid dividends accounts for an increasingly larger share of the non-payer group (see figure 5.4). It appears that (1) firms that were listed between 1997 and 2002 are more likely to have never paid dividends and (2) most of Thailand’s listed firms tend to pay dividends if they are capable of doing so. Therefore, most of the non-payers (former payers are the majority of this group) were forced to cease paying dividends due to financial distress flowing from the crisis. This view is consistent with the findings of Fama and French (2001). Former payers are likely to be distressed and have not paid dividends in time t. However, it remains to be seen whether non-payers and former payers will pay dividends in the future. This issue will be addressed later in this section. As stated earlier, the percentage of firms which paid dividends decreased from 84.2 percent in 1990 to 46.4 percent in 2002. Before the Asian Economic Crisis (19901996), the percentage of payers declined slightly from 84.2 percent in 1990 to 81.0 261 Chapter 5: Analysis of Data and Findings percent in 1996. However, the percentage of payers fell sharply from 71.5 percent to 28.4 percent during 1997 and 1998. In 1999, the percentage of payers increased slightly to 29.9 percent. Indeed, the percentage of payers has continued to rise and stood at 46.4 percent in 2002. This figure is about 20 percent higher than the figure recorded during the crisis, although it remains about 40 percent lower than 1990. The trends in the percentage of payers indicate (1) in the long-term, the payment of dividends appears to be disappearing because the percentage of listed payers has declined, and (2) in the short-term (the pre-crisis period), the percentage of payers declined slightly. This outcome may be an early indicator of the financial collapse which occurred in 1997, and (3) in the short-term (post-crisis period), the percentage of firms paying dividends has rebounded from the low point recorded during the crisis and may rise to pre-crisis levels in the future. The decline in the percentage of dividend payers may also result from growth in the non-payer group due to new listings. Before the crisis, the number of firms in most sectors expanded by more than 10 percent per annum due to new listings (see table 5.1). However the percentage of new listings declined to less than one percent of the total number of listed firms during the crisis and currently stands at about 5.2 percent. Table 5.1 indicates that some newly listed firms paid dividends, although the percentage of these firms fluctuated between 1990 and 2002. The percentage of new firms that paid dividends stood at 82.2 percent (1991), 73.3 percent (1992), 51.5 percent (1993), 52.6 percent (1994), 82.6 percent (1995), 73.5 percent (1996) and 50 percent in 1997. After the crisis, less than 30 percent of new lists paid dividends. In addition, the percentage of new firms that paid dividends was lower than the total percentage of payers in all years with the exception of 1998 and 2001 (see table 5.1). Therefore, it appears that new firms are less likely to pay dividends and this partly explains the decline in the percentage of payers between 1990 and 2002. Fama and French (2001) found that the percentage of listed firms which paid dividends decreased sharply from almost 70 percent in 1978 (a period of recession) to 20.8 percent in 1999. As stated earlier, Fama and French (2001) averaged the figures for the period preceding 1978 and explored the decline in the payment of dividends after 1978. 262 Chapter 5: Analysis of Data and Findings However, Thailand’s capital market is relatively new and has only experienced one major economic downturn (1997-1998). As stated earlier the percentage of listed payers in Thailand’s market has recovered slightly since the crisis. The percentage of firms paying dividends in the United States has continued to decline. Fama and French (2001) did not indicate whether there was any change in the percentage of payers in the US market after 1999. These researchers presented their data over several time periods and did not investigate trends from one year to the next. Therefore, more research is needed to determine whether the decline in the number of payers in the US has been sustained. This study presents the pattern of dividends year by year from 1990 to 2002 and in several time periods (range of years). This analysis is similar to Fama and French (2001). This study found that the percentage of payers in Thailand’s capital market increased from 20.8 percent in 1999 to 50.4 percent in 2002 (see figure 5.5). Table 5.8 presents the dividend behaviour of payers in year t-1 and non-payers in year t-1 undertake in time t. The table is divided into three parts. The first part presents behaviour in year t of firms that paid dividends during the previous year. These firms continue to pay dividends, ceased paying or were forced to delist. The second part presents the behaviour of firms that were non-payers in time t-1. These firms commenced paying dividends, remained a non-payer, or were delisted from the stock exchange. The last part shows the percentage of non-payers that commenced paying dividends in time t. The table also indicates if these firms were former payers, or if they have never paid dividends. 263 Chapter 5: Analysis of Data and Findings Table 5.8: What Happens in Year t to SET Firms that Do and Do not Pay Dividends in Year t-1? What Happens in Year t to Firms that Pay Dividends in Year t-1 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 78.4 88.1 93.0 93.5 91.3 91.4 81.7 38.4 75.5 89.6 92.3 95.4 Continue to Pay 13.1 6.7 5.7 6.1 6.2 6.6 17.0 59.4 23.6 5.2 6.9 3.9 Stop Paying 0.6 0.5 0.0 0.0 0.4 0.3 0.0 3.0 0.0 3.5 0.0 0.0 Delist What Happens in Year t to Firms that Do Not Pay Dividends in Year t-1 54.5 91.4 45.8 55.8 35.8 16.7 24.3 Start Paying 12.1 2.9 54.2 44.2 62.3 83.3 64.9 Do not Pay 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Delist 0.9 94.6 0.9 11.2 84.2 1.4 8.9 83.3 4.1 11.7 84.9 1.3 12.1 82.5 2.2 Percent of Non-Payers in Year t-1 that Start Paying in Year t-1 54.5 91.4 45.8 55.8 35.8 All Non-Payers (t-1) 63.6 11.4 29.2 7.0 15.1 Former Payers 33.3 20.0 70.8 37.2 1.9 Never Paid 0.9 27.7 0.0 11.2 8.6 0.0 8.9 10.0 0.4 11.7 9.6 1.7 12.1 0.0 0.0 16.7 20.4 13.0 24.3 1.4 0.0 Source: SET (1994, 1997, 2000, 2002a) Note: Firms that continue to pay were payers in year t-1 and continue to pay dividends in time t. Firms that cease paying dividends are firms that were payers in time t-1, but non-payers in time t. Firms that delist are firms that were payers in time t-1 but delisted in time t. Numbers shown in percentage. Table 5.8 shows the likelihood that payers and non-payers of the last period (t-1) will continue to pay, commence or cease paying dividends in this year. Firms which paid dividends last year (t-1) tend to continue paying dividends in year t. Before the crisis, almost 90 percent of the payers continued to pay dividends (table 5.8). In 1997, however, the percentage of payers in the previous year (t-1) that continued to pay (in t) fell to 81.7 percent and to 38.4 percent during 1998. After the crisis, the percentage of payers in the previous period, that continued to pay in the next period, increased to 75.5 percent in 1999, 89.6 percent in 2000, 92.3 percent in 2001 and finally 95.4 percent in 2002. Although the number of firms that continue to pay dividends has increased to more than 90 percent, it should be stated that the figures after the crisis, are compared with the lowest percentage of payers that continued to pay dividends in 1998 (38.4 percent). Therefore, after 1998, the denominator (the total number of payers in the last period) is less than the figure observed in the pre-crisis period. The question which arises is what has occurred to the remaining 60 percent of firms which are firms that ceased paying dividends during the crisis? 264 Chapter 5: Analysis of Data and Findings Firms which paid dividends before the crisis, but did not pay during the crisis, ceased paying dividends at the rate of 59.4 percent and delisted at the rate of 3 percent in 1998. Before the crisis, the percentage of payers that ceased paying dividends was about 10 percent per annum. However, this figure rose to 17 percent in 1997 and peaked at 59.4 percent in 1998. In addition, the percentage of payers that delisted in the next period (t) was less than one percent per annum before the crisis, but rose to 3 percent in 1998. Finally, after the crisis, a greater proportion of firms which paid dividends continued doing so while less ceased paying and delisted. This indicates that the market is recovering. Firms that did not pay dividends in the last period commenced paying at the rate of 91.4 percent in 1991 and 45.8 percent in 1992. This indicates that the vast majority of firms possessed a high propensity to pay, or resumed paying dividends at this time. However, this figure fell to 35.8 percent in 1994 and to less than one percent in 1998. Finally, after the crisis, this figure increased to above 10 percent, although, it is still lower than the percentage of non-payers that commenced paying dividends before the crisis. In summary, less firms re-commenced paying dividends after the crisis. Before the crisis, non-payers appeared less willing to commence paying dividends in the next period (decreasing their propensity to pay dividends). The percentage of nonpayers that continued to avoid this payment in the next period also supports this finding (see table 5.8). After 1993, a large proportion of non-payers continued to avoid paying dividends, even though it was still several years before the crisis. In addition, only 0.9 percent of non-payers commenced the payment of dividends in 1998. This figure increased to 12.1 percent in 2002, indicating that the market was recovering and a high propensity to pay dividends by former non-payers. The last part of table 5.8 indicates that, former payers resumed the payment of dividends more frequently than firms that have never paid dividends. This view supports the earlier discovery that former payers omit the payment of dividends but do not intend to cease the payment on a permanent basis. The descriptive statistical analysis of Fama and French (2001) suggests that former payers are distressed firms 265 Chapter 5: Analysis of Data and Findings which ceased paying dividends to preserve their cash. The relationship between former payers and distress will be analysed for Thailand with the aid of a variety of ratios. This analysis is presented in section 5.5. A preliminary analysis of the descriptive statistics reveals that the percentage of firms which paid dividends started to decline before the crisis, and fell sharply during the crisis. The percentage of payers rose slightly after the crisis, but remains much lower than the level recorded before the crisis. It also appears that payers were less likely to continue the payment of dividend during the crisis. The analysis suggests that the propensity to pay dividends of existing payers is lower. However, this suggestion will be confirmed in the later sections of the chapter. As stated earlier, the percentage of non-payers has increased over time and increased sharply immediately after the crisis. Most of the non-payers were former payers which were forced to avoid the payment of dividends due to the crisis. In addition, it appears that non-payers were less willing to re-commence paying dividends. This supports the proposed explanation for the observed decline in the percentage of payers; non-payers are less likely to re-commence paying dividends. Consequently, the total number of payers is smaller and represent a smaller proportion of the total number of firms. Many new firms were listed at SET before the crisis and large proportion of these firms did not pay dividends. The percentage of new firms that paid dividends was less than the percentage of listed firms that paid dividends, in most years (see table 5.1). Therefore, new lists usually do not pay dividends and the denominator (total number of listed firms) is increasing over time, while the numerator (the total number of payers) is becoming smaller leading to a decline in the percentage of payers. The findings of this section have raised several questions which require further investigation. (1) What happened to listed firms in 1997 to 1998? Why do payers tend to cease paying dividends? What are the characteristics of firms which pay dividends? Has the propensity to pay dividends fallen? Have characteristics of these firms changed? (2) What are new lists’ characteristics? Are the characteristics of new lists 266 Chapter 5: Analysis of Data and Findings different from mature firms? Is the decline in percentage of dividend payers due to new listing? Section 5.5 presents the descriptive statistics on the characteristics of the dividend groups and provides answers to these questions. 5.5 CHARACTERISTICS OF THE DIVIDEND PAYERS The preliminary discussion of the characteristics of dividend payers, non-payers and other dividend found that payers and non-payers differ in terms of (1) profitability (2) investment opportunities and (3) size. Section 5.5.2 and 5.5.4 will confirm the findings of the descriptive statistics with the aid of ANOVA (An Analysis of Variance) and logit regression. As stated earlier, Fama and French (2001) suggested there are three characteristics of firms which pay dividends (1) profitability (2) investment opportunities and (3) size. Each characteristic will be measured and compared with the aid of a variety of ratios. The analysis will compare the changes in these characteristics during the pre-crisis and post-crisis periods by using descriptive statistics. In addition, each characteristic will be compared with the findings of the literature. Finally, this analysis will test differences in the mean of these characteristics to determine whether they are significantly different. 5.5.1 FIRM CHARACTERISTICS DESCRIPTIVE STATISTICS APPROACH 5.5.1.1 PROFITABILITY Table 5.9 presents the characteristics of firms in a variety of dividend groups. The profitability of the dividend groups is expressed by two ratios ‘(1) Et/At (the ratio of aggregate earnings before interest to aggregate assets) (2) Yt/BEt (the ratio of aggregate common stock earnings over aggregate book equity)’ (Fama and French 2001, p. 7). The discussion is divided into two parts, in line with the two ratios for measuring profitability. 267 Chapter 5: Analysis of Data and Findings Table 5.9: Average Values of Six Characteristics Ratios 19912002 19911996 19972002 1991 1992 1993 1994 1995 1996 6.16 Et/At (percent): The Ratio of Aggregate Earnings Before Interest to Aggregate Assets All Firms 3.64 8.71 -1.43 11.16 10.14 8.96 8.07 7.78 Payers 6.68 9.47 3.89 10.97 10.59 9.88 8.97 8.84 7.53 -0.45 4.35 -5.24 10.75 5.84 3.73 3.37 2.08 0.31 Non-Payers Never Paid Former Payers 0.48 6.22 -5.25 11.41 8.66 6.33 7.18 2.78 0.97 -0.96 3.25 -5.16 12.63 3.45 0.75 0.67 1.92 0.08 All New Lists 4.62 11.77 -2.53 11.83 10.86 12.07 20.12 10.06 5.67 Payers 9.95 11.19 8.71 11.62 11.3 13.9 12.41 11.6 6.33 0.4 7.95 -7.14 12.82 9.63 10.13 8.57 2.74 3.84 19912002 19911996 19972002 1991 1992 1993 1994 1995 1996 Non-Payers Yt/BEt (percent): The Ratio of Aggregate Common Stock Earnings Over Aggregate Book Equity -108.62 8.65 -225.89 16.64 14.44 13.01 -8.42 10.57 5.64 -9.21 13.21 -31.63 16.67 16.26 12.79 13.17 11.92 8.44 Non-Payers -164.63 -15.91 -332.83 16.48 -3.01 14.28 -3.51 3.31 -6.1 Never Paid -43.23 5.85 -92.31 14.64 7.88 6.04 5.58 -11.52 12.51 0.08 All Firms Payers Former Payers -0.96 3.25 -5.16 12.63 3.45 0.75 0.67 1.92 All New Lists -10.1 14.63 -34.83 18 17.75 13.1 25.23 11.15 2.56 10.5 14.07 6.92 17.96 20.63 12.54 16.11 15.07 2.14 -18.34 8 -44.67 18.19 9.72 13.69 10.13 -7.48 3.72 19912002 19911996 19972002 1991 1992 1993 1994 1995 1996 Payers Non-Payers dAt/At (percent): The Growth Rate in Assets All Firms -8.98 23.15 -41.1 29.09 30.58 23.52 23.78 19.25 12.65 Payers 11.95 22.38 1.51 31.75 28.66 19.22 21.07 20.45 13.16 Non-Payers -18.49 28.72 -65.7 14.44 48.98 48.05 37.86 12.8 10.22 Never Paid 47.6 69.02 26.18 63.29 95.88 72.83 88.28 46.75 47.11 -39.78 3.74 -83.29 -11.05 9.29 19.55 2.08 5.09 -2.53 All New Lists 96.39 113.05 79.74 100 97.87 97.13 186.24 100 97.06 Payers 87.57 98.89 76.25 100 97.11 100 96.24 100 100 Non-Payers 87.57 97.16 77.97 100 100 94.07 100 100 88.89 19912002 19911996 19972002 1991 1992 1993 1994 1995 1996 1.18 Former Payers Vt/At: The Ratio of the Aggregate Market Value to the Aggregate Book Value of Assets All Firms 1.77 2.15 1.39 4.35 1.92 2.42 1.67 1.38 Payers 1.55 2.09 1 4.02 1.9 2.41 1.65 1.39 1.18 Non-Payers 2.05 2.5 1.6 6.16 2.08 2.46 1.81 1.36 1.11 Never Paid 1.85 2.52 1.18 4.02 2.36 2.9 2.27 2.05 1.51 Former Payers 2.06 2.46 1.66 7.28 1.85 1.97 1.48 1.2 0.98 All New Lists 1.27 2.09 0.46 3.94 1.89 1.94 2.27 1.6 0.89 Payers 1.95 2.75 1.15 4.79 2.56 3.77 2.27 1.93 1.21 Non-Payers 2.01 2.69 1.33 3.29 2.48 3.68 2.4 2.35 1.92 268 Chapter 5: Analysis of Data and Findings Table 5.9: Average Value of Six Characteristic Ratios (Continued) 19912002 19911996 19972002 1991 1992 1993 1994 1995 1996 All Firms 18107.71 14816.16 21399.25 12497.04 12378.22 13372.01 14835.83 17179.76 18634.12 Payers 16160.28 16593.41 15727.15 14185.74 13086.91 14633.33 16802.84 19275.35 21576.29 Non-Payers 14819.71 5246.12 24393.3 3185.04 5586.69 6185.43 4629.64 5925.68 5964.26 Never Paid 13830.14 8507.63 19152.64 1250.8 2546.08 4359.1 7971.55 19891.7 15026.57 Former Payers 15355.32 4747.11 25963.53 4194.21 8159.52 8285.72 2257.97 2751.59 2833.64 6936.83 5440.35 8433.31 3041.72 2628.18 3226.89 8066.69 10383.35 5295.26 3196.8 3987.73 2405.87 3335.83 2588.8 1572.31 2273.34 8989.73 5166.41 7869.11 13987.61 1750.6 15428.2 7248.63 1670.57 2525.93 42701.22 14351.13 19912002 19911996 1991 1992 1993 1994 1995 1996 At: Total Assets All New Lists Payers Non-Payers 19972002 Lt/At: The Ratio of Total Liabilities to Total Assets All Firms 0.73 0.53 0.93 0.53 0.51 0.55 0.5 0.53 0.56 Payers 0.45 0.52 0.39 0.5 0.5 0.55 0.5 0.52 0.55 Non-Payers 0.92 0.58 1.25 0.67 0.57 0.54 0.53 0.6 0.58 Never Paid 0.63 0.45 0.81 0.29 0.54 0.39 0.41 0.57 0.49 1 0.67 1.33 0.86 0.59 0.72 0.61 0.61 0.61 All New Lists 0.62 0.53 0.72 0.44 0.53 0.57 0.67 0.47 0.49 Payers 0.41 0.51 0.31 0.46 0.53 0.72 0.33 0.45 0.54 Non-Payers 0.61 0.43 0.8 0.32 0.55 0.42 0.37 0.56 0.36 Former Payers Source: Developed for this research Note: At is total asset, BEt is the book value for common equity, MEt is the market value of common equity, Lt is the book liabilities, Vt is the market value of the entire firm, Et is the earnings before interest but after taxes, Yt is the after-tax earnings to common stock and dAt is the changes in the assets from time t and t-1. Results for ranges of years are grouped and averaged according to the dividend groups. Et/At, Yt/Bet, dAt/At are shown in percentages. Vt/At and Lt/At are shown in ratio values. At is shown in Baht values. 269 Chapter 5: Analysis of Data and Findings Table 5.9: Average Values of Six Characteristic Ratios (Continued) 1997 1998 1999 2000 2001 2002 Et/At (percent): The Ratio of Aggregate Earnings Before Interest to Aggregate Assets All Firms -9.43 4.11 -3.25 -3.92 -1.55 Payers -5.36 9.56 8.84 8.5 -7.28 9.08 Non-Payers -19.64 2.52 -8.4 -10.66 2.38 2.34 Never Paid -22.57 1.8 -6.53 -5.75 -1.27 2.8 Former Payers -18.92 2.6 -8.62 -11.39 3.17 2.18 -9.72 -1.11 -19.39 -7.39 12.51 9.9 -10.18 19.99 0 7.17 20.72 14.56 -9.26 -11.67 -19.39 -12.85 1.57 8.73 1997 1998 1999 2000 2001 2002 All New Lists Payers Non-Payers 5.47 Yt/BEt (percent): The Ratio of Aggregate Common Stock Earnings Over Aggregate Book Equity All Firms -289.65 -706.33 -63.82 -43.08 -259.24 6.76 Payers -246.84 10.54 10.87 10.92 12 12.71 Non-Payers -397.05 -988.27 -95.63 -72.28 -445.34 1.61 Never Paid -61.75 -76.07 -354.55 -44.79 -10.46 -6.27 Former Payers -18.92 2.6 -8.62 -11.39 3.17 2.18 All New Lists -52.13 -9.7 -147.88 -30.99 18.11 13.65 Payers -42.88 21.88 0 7.71 30.79 24.04 Non-Payers -61.39 -25.5 -147.88 -45.51 1.2 11.05 1997 1998 1999 2000 2001 2002 -2.98 dAt/At (percent): The Growth Rate in Assets -1.23 -88.14 -5.78 -19.64 -128.86 0.14 -2.36 5.43 3.14 -2.02 4.73 Non-Payers -4.66 -121.49 -10.56 -31.96 -215.88 -9.64 Never Paid 65.87 10.85 11.58 13.64 25.89 29.27 Former Payers -21.9 -134.57 -13.12 -38.75 -268.72 -22.7 All New Lists 100 54.16 49.82 90.38 88.6 95.49 Payers 100 100 0 100 80.05 77.47 Non-Payers 100 31.24 49.82 86.77 100 100 1997 1998 1999 2000 2001 2002 All Firms Payers Vt/At: The Ratio of the Aggregate Market Value to the Aggregate Book Value of Assets All Firms 1.04 1.88 1.26 1.27 1.59 Payers 1.05 1.01 1 0.87 0.99 1.34 1.11 Non-Payers 1.02 2.22 1.36 1.48 1.99 1.54 Never Paid 0.99 0.93 1.27 1.16 1.27 1.48 Former Payers 1.02 2.34 1.37 1.52 2.15 1.57 All New Lists 0.76 0.45 0 0.23 1.08 0.27 Payers 1.52 1.34 0 0.83 1.89 1.33 Non-Payers 0.81 1.2 1.43 1.41 1.46 1.66 270 Chapter 5: Analysis of Data and Findings Table 5.9: Average Values of Six Characteristic Ratios (Continued) 1997 1998 1999 2000 2001 2002 22974.21 At: Total Assets All Firms 20729.08 20147.08 21303.15 22159.39 21082.6 Payers 25384.91 47848.21 3034.98 3893.16 5289.25 8912.41 Non-Payers 9047.92 9173.24 29084.04 32002.28 31918.4 35133.93 Never Paid 17754.73 17074.25 26090.25 23752.6 18282.7 11961.32 Former Payers 6919.59 8392.5 29430.43 33231.79 34898.88 42907.97 All New Lists 2272.65 7479.3 27672.49 6902 2670.88 3602.55 Payers 2772.81 1192.13 0 1436.5 4003.96 5029.8 Non-Payers 2772.81 596.07 0 538.69 5338.61 1257.45 1997 1998 1999 2000 2001 2002 Lt/At: The Ratio of Total Liabilities to Total Assets All Firms 0.68 1.34 0.81 0.93 1.11 0.71 Payers 0.63 0.39 0.34 0.32 0.33 0.33 Non-Payers 0.82 1.72 1 1.26 1.65 1.05 Never Paid 0.82 0.75 0.84 0.92 0.86 0.71 Former Payers 0.82 1.81 1.02 1.31 1.82 1.16 All New Lists 0.68 0.8 1.29 0.8 0.36 0.37 Payers 0.65 0.09 0 0.29 0.4 0.44 0.7 1.15 1.29 0.99 0.31 0.35 Non-Payers Source: Developed for this research Note: At is total asset, BEt is the book value for common equity, MEt is the market value of common equity, Lt is the book liabilities, Vt is the market value of the entire firm, Et is the earnings before interest but after taxes, Yt is the after-tax earnings to common stock and dAt is the changes in the assets from time t and t-1. Results for ranges of years are grouped and averaged according to the dividend groups. Et/At, Yt/Bet, dAt/At are shown in percentages. Vt/At and Lt/At are shown in ratio values. At is shown in Baht values. 271 Chapter 5: Analysis of Data and Findings The ratio of aggregate earnings before interest to aggregate assets (Et/At) Before the Asian Economic Crisis (1991-1996) payers, on average, have higher profitability (9.47 percent) than non-payers (4.35 percent), firms which have never paid dividends (6.22 percent) and former payers (3.25 percent). All samples before the crisis indicate that firms which pay dividends are the most profitable (8.71 percent). This result is consistent with the findings of Fama and French (2001). Therefore, when using the Et/At ratio for measuring profitability, payers have higher profitability than nonpayers. When comparing the two groups of non-payers, firms that have never paid dividends were more profitable (6.22 percent) than former payers (3.25 percent). As explained in chapter 2, during the Asian Economic Crisis almost 70 percent of firms listed in Thailand suffered large losses (SET 2002a). Therefore, payers were less profitable during and after the crisis (1997-2002) although they remained the most profitable (3.89 percent) of the four dividend groups. While payers experienced positive profitability, non-payers experienced losses (-5.24 percent). Within the non-payer group, firms that have never paid dividends suffered slightly higher losses (-5.25 percent) than former payers (-5.16 percent). The average profitability (Et/At) of all samples after the crisis was -1.43 percent. This was lower than the profitability of dividend payers, but higher than the figure recorded for non-payers and their subgroups. This finding is consistent with the literature (Fama and French 2001). During the period 1991 to 2001, payers of dividends were the most profitable (6.68 percent) followed by firms that have never paid dividends (0.48 percent). Non-payers of dividends were more profitable than former payers. Non-payers experienced losses of -0.45 percent while former payers experienced losses of -0.96 percent. The all firm average level of profitability was 3.64 percent. Therefore, the Et/At ratio indicates that payers are more profitable than non-payers during favourable and poor economic conditions. Former payers tend to be distressed 171 Chapter 5: Analysis of Data and Findings because they are the least profitable in good times and experience large losses in bad times. Hence, these firms often cease paying dividends. The descriptive statistics have shown that when non-payers are divided into two sub-groups (never paid and former payers), firms which have never paid dividends perform better than former payers. This view is consistent with the findings of La Porta et al. (2000), Benito and Young (2001), DeAngelo, DeAngelo and Skinner (2001), Fama and French (2001), and Aivazian, Booth and Cleary (2003). Generally, firms that were newly listed on the market before the crisis were highly profitable (11.7 percent) and more profitable than the average of the existing listed firms (see table 5.9). The profitability of the new lists that paid dividends was also high (11.19 percent), while the profitability of new lists that did not pay dividends was lower (7.95 percent). All figures for profitability measured by Et/At were higher for new lists than for existing listed firms before the crisis period. After the crisis, the losses of the new lists increased slightly (-2.03) percent, when compared to the figure recorded for the existing listed firms (table 5.5). However, new lists that paid dividends were highly profitable (8.71 percent) when compared to all existing listed firms that pay dividends. For the whole period of investigation, new lists were relatively profitable (4.62 percent) and generally more profitable than existing firms during the same period. New lists that paid dividends were highly profitable (9.95 percent) and this figure is higher than existing firms which paid dividends. 172 Chapter 5: Analysis of Data and Findings Fama and French (2001) found that new lists tended to be highly profitable, early on, although their profitability declined after the crisis in 1978 (Fama and French 2001). Therefore, Fama and French (2001) concluded that the decline in the percentage of payers after 1978 is partly due to the new characteristics of the new lists (low profitability, high investment opportunities and small size). However, in the case of Thailand, a clear trend in the disappearing of dividends is not evident. It appears that Thai firms have resumed paying dividends after a crisis. Therefore, new lists with higher profitability than mature firms are more common in Thailand. There is no clear explanation for this result although it appears that changes in the characteristics of new lists and mature firms are very small and the Thai capital market has not reached its turning point. 173 DBA Thesis: Malinee Ronapat The ratio of aggregate common stock earnings over aggregate book equity (Yt/BEt) Fama and French (2001) suggested that Et or earnings before interest may not be the relevant tool affecting firms’ dividend payment decisions. These researchers suggested that Yt, or earnings available for common stocks could be a better tool. Table 5.9 indicates that the gap between the profitability of payers and non-payers tends to be wider when measured by the Yt/BEt ratio. This view is consistent with the findings of Fama and French (2001). During the pre-crisis period dividend payers were more profitable (13.21 percent) than the other dividend groups. When focusing on non-payers, firms that have never paid dividends were more profitable (5.85 percent) than former payers (3.25 percent). The average level of profitability for all firms was 8.65 percent, which is lower than that recorded for payers, but higher than the other dividend groups. This finding is similar to the findings of Fama and French (2001). During and after the Asian Economic Crisis, all firms, on average, experienced losses of -225.89 percent. The largest losses were recorded in 1997 and 1998 (table 5.9). Furthermore, when analysing each dividend group, the data indicates that every dividend group experienced losses. In relation to the period 1991 to 2002, the average Yt/BEt ratio for payers was -9.21 percent, while firms which have never paid dividends experienced greater losses of -43.23 percent. Non-payers and the sub-group of former payers, experienced very large losses of -164.63 percent and -43.23 percent respectively. The average level of profitability of all firms was -108.62 percent. 174 DBA Thesis: Malinee Ronapat When using Yt/BEt as a measure of profitability, payers are more profitable than nonpayers during positive economic conditions. This finding is consistent with La Porta et al. (2000), Benito and Young (2001), DeAngelo, DeAngelo and Skinner (2001), Fama and French (2001), and Aivazian, Booth and Cleary (2003). During difficult times, payers experienced smaller losses than non-payers. The profitability of new lists was 14.63 percent before the crisis (using Yt/BEt). The average profitability of new lists that paid dividends was 14.07 percent, which is much higher than the level experienced by the existing firms which paid dividends. During and after the crisis, new lists made losses of -34.83 percent. However, new lists that paid dividends were highly profitable and this performance was superior to the results achieved by listed firms that pay dividends. For the whole period of investigation (1990-2002), new lists experienced losses of -10.10 percent, although the payer groups were more profitable (10.5 percent) than the returns achieved by existing listed firms that paid dividends (see table 5.9). These findings are similar to Fama and French (2001). However, Fama and French (2001) stated that the characteristics of newly listed firms changed from high to low profitability, after a strong downturn in the US in 1978. It should be noted that Fama and French (2001) concluded that this outcome is not consistent, over time, for new lists that paid dividends and mature firms that paid dividends. Mature firms which pay dividends are less profitable than new listed payers when using both ratios (Et/At and Yt/BEt) in the US. This finding is similar to the results of this study on the Thai stock market (see table 5.9). The study focuses on the characteristics of the dividend payers. Therefore, it is important to highlight the issue that new lists which pay dividends are more profitable than mature firms that do not pay dividends in Thailand. Furthermore, the profitability of new lists has been consistently higher than mature firms over the period of investigation. 175 DBA Thesis: Malinee Ronapat 5.5.1.2 INVESTMENT OPPORTUNITIES The investment opportunities of firms are measured by the three ratios which were recommended by Fama and French (2001) (1) dAt/At (growth rate in assets) (2) Vt/At (the ratio of the aggregate market value to the aggregate book value of assets) (3) RDt/At (research expenditure) in table 5.9. However, this study will apply the first two ratios because data on research and development is not available in Thailand. This information is not required by SEC and need not be presented on the financial statements of listed firms (SET 2003b). The two ratios will be calculated and analysed separately in this study. The growth rate in assets (dAt/At) Before the Asian Economic Crisis, the percentage rate of growth in assets for payers was 22.38 percent, however, this was lower than that achieved by non-payers (28.72 percent). This finding is consistent with Fama and French (2001). These researchers suggested that payers have poorer investment opportunities than non-payers, because they pay a proportion of their earnings as dividends (Fama and French 2001). Former payers and firms that have never paid dividends experienced growth in their assets of 3.74 percent and 69.02 percent respectively. The average rate of growth in assets of all samples was 23.15 percent. Firms that pay dividends out of their earnings have fewer opportunities for investing in projects with good returns. Hence, firms which pay dividends have a low rate of growth in assets (as a proxy of investment opportunities) than firms which retain their earnings and invest in assets (Fama and French 2001). When the Asian Economic Crisis commenced in 1997, every dividend group experienced a negative rate of growth in their assets, with the exception of the payer group (1.51 percent) and firms which have never paid dividends (26.18 percent). After the crisis, these two groups continued to experience significant growth in their assets, while the other groups continued to record negative growth in their assets. Negative 176 DBA Thesis: Malinee Ronapat growth rates could be due to downsizing, or the liquidation of some of the assets owned by non-payers and former payers. During the period 1990 to 2002, the rate of growth of assets for payers was positive (11.95 percent). The highest rate achieved among the four dividend groups was achieved by firms that have never paid dividends (47.6 percent). The asset growth rate of non-payers was -18.49 percent, while the former payers experienced losses of -39.78 percent. Firms that have never paid dividends recorded the highest growth in their assets followed by the payers. This finding is consistent with the literature (Fama and French 2001). Firms that have never paid dividends and non-payers exhibit higher investment opportunities than payers before and after the Asian Economic Crisis. The securities of these firms are therefore likely to be real growth stocks. Payers maintained their asset sizes at levels recorded prior to the crisis (table 5.9). After the crisis, payers continued to have positive investment opportunities, while the other listed firms engaged in downsizing. Former payers were distressed and experienced low investment opportunities and growth. Overall, payers tend to have a lower rate of asset growth, than firms that have never paid dividends (see table 5.9). However, payers experienced a higher rate of growth of their assets than non-payers and former payers. During 1991-1996, newly listed firms experienced high growth in investment opportunities (113.05 percent) and the new lists which paid dividends experienced lower asset growth than the new lists which did not pay dividends. Therefore, after 1991 new lists experienced high growth and tended to have similar characteristics to non-payers. This result is consistent with Fama and French (2001). The ratio of the aggregate market value to the aggregate book value of assets (Vt/At) During the pre-crisis period, investors perceived that the assets of firms which pay dividends, should be valued highly (2.09 times), but lower than that of non-payers (2.05 177 DBA Thesis: Malinee Ronapat times). This implies that (1) the market value of payers’ stocks could be undervalued or (2) investors are willing to pay more for the stocks of firms that do not pay dividends. An appreciation in stock prices is therefore more likely for firms that do not pay dividends. The Vt/At ratio for never paid firms is 2.52 times and 2.46 times for former payers. The average Vt/At ratio for all firms was 2.15 times. The payer group had the lowest Vt/At ratio. In the post-crisis period, the assets of payers valued more lowly (1.00 times) than other dividend groups. The average Vt/At for all firms was 1.39 times. These results were also consistent with Fama and French (2001). Over the period 1990 to 2002, the Vt/At ratio of payers has averaged 1.77 times. At present, payers have a lower Vt/At than non-payers and other dividend groups. In summary, it is observed that payers experience lower growth than non-payers because they tend to maintain the size of their assets during recessions, while nonpayers, never-paid and former-payers accounted for more than 50 percent of the firms engaged in downsizing (BOT 1998). Over the period 1990 – 2002 payers had lower investment opportunities than the other dividends groups. Former payers were distressed firms with low profitability and low investment opportunities. This result is also consistent with the findings of Fama and French (2001). Furthermore, Fama and French (2001) suggested that in relation to investment opportunities (1) the understatement of profitability (measured by Et/At) by growth firms, could take some time to reach its maximum level and (2) profitability (measured by Et/At) of payers could be overstated due to inflation (Fama and French 2001). 178 DBA Thesis: Malinee Ronapat 5.5.1.3 SIZE The size of firms can be measured with two ratios (1) At (total assets) and (2) Lt/At (the ratio of total liabilities to total assets) in (table 5.9). In line with Fama and French 2001, each ratio will be analysed separately. Total assets (At) From 1991 to 1996, the total assets of payers was 16593.41 million Baht. This was higher than the figure for non-payers’ (5246.12 Million Baht), never paid (8507.63 Million Baht) and former payers (4747.11 Million Baht). Therefore, the assets of payers tend to be larger than the other dividend groups and the assets of former payers were larger than non-payers and firms which have never paid dividends. This result is consistent with the findings of Fama and French (2001). After 1997, the total assets of payers were maintained at 15727.15 million Baht, while the assets of non-payers increased to 24393.3 million Baht. The asset sizes of never paid and former payers more than doubled, (19152.64 and 25963.53 Million Baht respectively) and the assets of the payer group was smaller than the assets of nonpayers. This outcome is due to (1) a reduction in number of payers (2) an increase in the total number of listed firms (3) the new lists were smaller in size than the mature firms and (4) a reduction in the percentage of new lists that paid dividends (see table 5.1 and 5.9) However over the period 1991 to 2002, the total assets of payers’ was maintained at 16160.28 mil. Baht. Indeed, the total assets of the payer group was less than all other dividend groups. This is consistent with Fama and French’s (2001). 179 DBA Thesis: Malinee Ronapat In terms of total assets, firms that paid dividends were larger firms before and after the crisis. These firms also maintained their size. In addition, more firms became nonpayers during the period and the asset size of non-payers also increased. The ratio of total liabilities to total asset (Lt/At) This ratio represents the value of a firm’s liabilities to its total assets in baht. Therefore, it represents the leverage of the firm. Fama and French (2001) used Lt/At to measure the size of firms and assumed that the higher a firm’s liabilities, the higher is its ability to access funds and its size. Before the crisis, the Lt/At ratio for payers was 0.52 times. This was higher than the ratio for never paid firms (0.45 times) and indicates that payers were well balanced in their debt and equity capital. The ratios of non-payers and former payers were higher at 0.58 and 0.67 respectively. After the crisis, the ratio of most dividend groups was large, indicating that the debt levels of all firms had increased. Only the group that paid dividends maintained its Lt/At ratio. This indicates that these firms were not experiencing high levels of debt or financial difficulties. On average, between 1997 and 2001, the lowest Lt/At ratio (0.39 times) was achieved by payers, followed by never paid firms (0.81 times) and former payers (1.33 times). The non-payer group averaged 1.25 times. From 1991 to 2002, the payer group recorded the lowest Lt/At, or highest level of leverage (0.45 times). The ratio of all firms’s was 0.73, non-payers 0.92, never paid 0.63 and former payers 1.00 times. After the economic crisis, a large number of firms were counted as non-payers, never paid and former payers. Therefore, the total assets of these groups increased after the crisis. 180 DBA Thesis: Malinee Ronapat In term of leverage, firms which pay dividends have high leverage and they tend to be well balanced between debt and equity capital, while non-payers are burdened with debt. When measuring total assets, firms which paid dividends were larger than the other dividend groups, before the crisis, although the assets of some of these firms were maintained after the crisis, while the value fell for some payers. The number of former payers and non-payers grew after the Asian Economic Crisis. However, the ratio of total liabilities to total assets was small for payers over time and the total asset size of the payer group remained higher than the other dividend groups. The debt and equity of new lists was well balanced before the crisis as indicated by a Lt/At ratio of 0.53 times. After the crisis, this ratio increased to 0.72 times. This figure was greater than the figure recorded for the existing listed firms. This indicates that the new lists were greater issuers of debt than existing firms. Table 5.10 summarises the characteristics of the dividend groups after the analysis using descriptive statistics. Table 5.10: Summary of the Characteristics of the Listed Firms in Thailand from Descriptive Statistics Approach CHARACTERISTICS OF THE LISTED FIRMS All times Payer Profit (Highest) Highly Positive Before Crisis Investment Size Highly Positive Largest After Crisis Profit (Highest) Highly Positive Investment Size Positive Largest Profit (Highest) Slightly Positive Investment Highly Negative Non-Payer Slightly Negative Negative Slightly Positive Highly Positive Highly Negative Never Paid Slightly Positive (Highest) Highly Positive Slightly Positive (Highest) Highly Positive Highly Negative Former Payer (Lowest) Slightly Negative Highly Negative (Lowest) Slightly Positive Slightly Positive Highly Negative Size Positive (Highest) Highly Positive (Lowest) Highly Negative Source: Developed for this research 181 Largest DBA Thesis: Malinee Ronapat Table 5.11 shows the percentage aggregate value accounted attributed to firms paying dividends. Table 5.11 shows the real effect of the numerator and denominator for the six ratios used to measure the characteristics of firms. Table 5.11: Percent of Aggregate Values Accounted for by Firms Paying Dividends Yt Et dAt At Vt BEt MEt Lt 19912002 91.9 47.5 56.0 63.5 62.6 70.4 67.1 62.7 19911996 97.9 96.6 90.0 94.7 92.2 89.0 86.0 95.9 19972001 85.8 -1.6 22.0 32.3 33.0 51.9 48.3 29.5 1991 90.4 95.4 95.2 96.1 95.3 92.2 91.0 96.7 1992 97.0 96.3 92.4 95.7 95.0 94.2 92.9 96.0 1993 96.7 95.0 83.0 93.1 88.2 86.4 80.7 94.5 1994 98.0 97.0 86.8 95.0 90.6 86.3 81.3 96.9 1995 100.0 97.4 91.1 94.6 92.1 88.0 85.4 96.1 1996 105.2 98.6 91.7 93.7 92.4 86.7 84.4 95.3 1997 51.3 -66.5 86.6 87.6 88.1 82.4 88.0 88.1 1998 64.3 65.7 31.4 67.3 66.6 58.2 53.4 68.6 1999 -5.8 -104.5 10.1 4.3 4.4 23.3 13.7 1.8 2000 -20.0 38.9 29.0 6.2 6.1 39.3 28.4 2.6 2001 333.7 9.7 -52.8 10.2 11.5 48.7 44.1 5.1 2002 91.5 47.4 27.9 18.0 21.0 59.3 62.3 10.8 Source: Developed for this research Note: At is asset, BEt is the book value for common equity, MEt is the market value of common equity, Lt is the book liabilities, Vt is the market value, Et is the earnings before interest but after taxes, Yt is the after-tax earnings to common stock and dAt is the change in assets in time t and t-1. Results for each time period are grouped and averaged according to the dividend groups. The aggregate values of both Yt and Et, are used for presenting the profitability of firms. Firms which paid dividends accounted for almost 100 percent of the total Yt in the market before the crisis. This confirms the findings of section 5.5 and payers are more profitable than the other dividend groups. After the crisis, the figure for Yt and Et decreased to 85.8 percent. This may result from the smaller number of payers in that year. Indeed, most of the firms, including those that experienced high profitability, suspended the payment of dividends in 1998. In 2001 and 2002, Yt increased to 91.5 percent because a large number of firms paid dividends (see table 5.1). The dAt which represents the growth in assets, or investment opportunities displayed a very high aggregate value ratio from 1991 to 1996 (table 5.11) although this declined to only 22 percent after the crisis. The total assets (At) accounted for by payers was a very high ratio between 1991 and 1996. However, after the crisis At fell to 4 percent, before 182 DBA Thesis: Malinee Ronapat rebounding to 18 percent in recent years due to the recovery. A similar pattern has occurred with Vt, BEt, MEt and Lt (table 5.10). This outcome has resulted from (1) a decline in the numbers of payers (2) downsizing of firms during the crisis (3) an increase in the number of listed firms and (4) the fact that new lists are more likely to be growth rather than value firms (table 5.1, 5.9). 5.5.2 ANOVA TEST An analysis of variance (ANOVA) was undertaken to confirm that the groups are sampled from different populations. ANOVA can also be used for comparing more than two means at the same time (Moore and McLabe 1998; Ticehurst and Veal 2000; Tabachnick and Fidell 2001). ANOVA is used for calculating the F statistic. The F statistic is the ratio of the magnitude of the difference between the dividend groups to the magnitude of the differences within these groups (Mason, Lind and Marchal 1999). An ANOVA test was used to statistically confirm the results which are summarised in table 5.9, 5.10 and 5.11. The first ANOVA test will (1) analyse the variance of the payer and non-payer groups to statistically confirm the characteristics of the payers from 1991 to 2002 and (2) analyse three groups: payers, former payers and firms that have never paid dividends to statistically confirm the characteristics of these groups from 1991 to 2002. Table 5.12 identified the year and names of the ratios (presented earlier as characteristics) which differed significantly between payers and non-payers for every year over the period 1991 to 2002. The analysis was based on the F-statistic at the 0.05 level of significance. Table 5.12 summarises the characteristics of payers in a general form. 183 DBA Thesis: Malinee Ronapat Table 5.12: ANOVA F-statistics Test of Characteristics from 1991 to 2002 ANOVA Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 NONE - Sum of Squares - df - Mean Square - F - Sig. - Et/At Between Groups 0.049 1 0.049 7.941 0.005 Yt/BEt Between Groups 0.807 1 0.807 14.172 0.000 dAt/At Between Groups 0.897 1 0.897 5.570 0.019 Et/At Between Groups 0.140 1 0.140 13.359 0.000 dAt/At Between Groups 3.022 1 3.022 12.842 0.000 Et/At Between Groups 0.118 1 0.118 29.695 0.000 Yt/BEt Between Groups 76.584 1 76.584 6.444 0.012 Et/At Between Groups 0.208 1 0.208 63.380 0.000 Yt/BEt Between Groups 0.338 1 0.338 3.929 0.048 Lt/At Between Groups 0.350 1 0.350 7.895 0.005 0.000 Et/At Between Groups 0.303 1 0.303 51.805 Yt/BEt Between Groups 1.261 1 1.261 13.242 0.000 Et/At Between Groups 1.655 2 0.827 11.861 0.000 Lt/At Between Groups 3.055 2 1.528 12.980 0.000 At Between Groups 117807901049 1 117807901049 11.211 0.001 0.000 Et/At Between Groups 2.397 1 2.397 17.776 dAt/At Between Groups 2.062 1 2.062 8.638 0.003 Vt/At Between Groups 10.566 1 10.566 6.267 0.013 AT Between Groups 54724905688 1 54724905688 5.485 0.020 Lt/At Between Groups 35.904 1 35.904 36.901 0.000 Et/At Between Groups 3.099 1 3.099 6.728 0.010 Vt/At Between Groups 32.035 1 32.035 6.657 0.010 0.014 At Between Groups 67067466395 1 67067466395 6.127 Lt/At Between Groups 75.784 1 75.784 15.778 0.000 Vt/At Between Groups 91.620 1 91.620 5.912 0.016 At Between Groups 64346225910 1 64346225910 6.093 0.014 Lt/At Between Groups 159.137 1 159.137 10.592 0.001 Et/At Between Groups 0.437 1 0.437 12.163 0.001 Vt/At Between Groups 18.053 1 18.053 7.130 0.008 At Between Groups 66001204222 1 66001204222 5.775 0.017 Lt/At Between Groups 49.175 1 49.175 24.205 0.000 Source: Developed for this research Note: Et/At is the ratio of aggregate earnings before interest to aggregate assets, Yt/BEt is the ratio of aggregate common stock earnings over aggregate book equity, dAt/At is the growth rate in assets, Vt/At is the aggregate market value to the aggregate book value of assets, At is total assets, Lt/At is the ratio of total liabilities to total assets. Statistic is significant at 0.05 level. See appendix A for details. 184 DBA Thesis: Malinee Ronapat Table 5.13: Summary of the Payers’ Characteristics Confirmation Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Characteristics confirmed Investment Profitability Size Opportunities Et/At Yt/BEt dAt/At Vt/At At Lt/At Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Source: Developed for this research Note: Y is ‘Yes’ for confirming the characteristics in each year from 1991 to 2002. Et/At is the ratio of aggregate earnings before interest to aggregate assets, Yt/BEt is the ratio of aggregate common stock earnings over aggregate book equity, dAt/At is the growth rate in assets, Vt/At is the aggregate market value to the aggregate book value of assets, At is the total asset, Lt/At is the ratio of total liabilities to total asset. See appendix A for details. After conducting a one-way analysis of variance, firms which paid dividends displayed significantly higher profitability (as measured by Et/At) (table 5.12) than non-payers in 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000 and 2002. In 1998 and 2001, there is no significant difference between the means of the two groups. However, most firms made losses in these years. The Yt/BEt (the ratio of aggregate common stock earnings over aggregate book equity) of payers indicated significantly higher profitability for firms which paid dividends in 1992, 1994, 1995 and 1996. It appears that profitability is an important characteristic of firms which paid dividends before and after the crisis. The investment opportunities (this was measured by dAt/At which is an indication of the growth rate of assets) of the payers was significantly lower than non-payers in 1992, 1993, 1997 and 1999. Vt/At (market value of stocks to book value of assets) identified significant variation between payers and non-payers in 1999, 2000, 2001 and 2002. Firms which pay 185 DBA Thesis: Malinee Ronapat dividends have a lower Vt/At than firms which did not pay dividends. Significant differences in firm size emerged after At (or total assets) (see table 5.12). Firms which paid dividends have less assets than non-payers. This could be a result of the downsizing which many firms undertook due to the Asian Economic Crisis. The variable Lt/At (total liabilities to total assets) was used as an alternative measurement of size. This variable was significantly smaller for payers than non-payers in 1995, 1997, 1999, 2000, 2001, 2002. The arithmetic means are presented in tables 5.9 and 5.11 and indicate that the characteristics of payers and non-payers differ. ANOVA suggested that profitability, as measured by the Et/At ratio (earnings to assets) was a significant characteristic of firms which paid dividends before and after the crisis. A firm’s Yt/BEt (earnings to book equity) was only important before the crisis. The results for investment opportunities were mixed. However, it was clear that Vt/At was low and a significant characteristic of payers after the crisis. In addition to size, both At (total assets) and Lt/At (total liabilities to total assets) tended to be characteristics of payers after the crisis (smaller At and Lt/At than non-payers). The second ANOVA test divided the total samples into three groups (1) payers (2) former payers and (3) firms that have never paid dividends. It is appropriate to divide the non-payers into two groups (group (2) and (3)) because this is a better classification of the characteristics of firms which do not pay dividends (Fama and French 2001). Table 5.14 presents the values for the ANOVA F-statistics over a range of firm characteristics. The table presents the results of six ratios for each year between 1991 and 2002. Table 5.15 is a summary of the characteristics of payers, former payers and firms which have never paid dividends. 186 DBA Thesis: Malinee Ronapat Table 5.14: ANOVA Test for Payers, Former Payers and Never Paid Firms Year Sum of Squares 1991 1992 1993 1994 1995 dAt/At 2 2.615 F Sig. 6.849 0.001 (I) Multiple Comparisons: Tukey HSD Std. (J) Mean Difference (I-J) Error Sig. 1 2 0.426 0.136 0.006 3 2 0.743 0.22 0.002 Lt/At 0.494 2 0.247 4.402 0.013 1 3 0.209 0.07 0.009 Et/At 0.065 2 0.033 5.317 0.005 1 2 0.071 0.022 0.004 Yt/BEt 1.048 2 0.524 9.319 0 1 2 0.285 0.068 0 dAt/At 5.365 2 2.683 18.644 0 1 3 -0.672 0.117 0 2 3 -0.866 0.155 0 Et/At 0.187 2 0.094 9.026 0 1 2 0.093 0.022 0 dAt/At 7.789 2 3.894 17.753 0 1 3 -0.645 0.109 0 2 3 -0.668 0.143 0 Et/At 0.194 2 0.097 25.546 0 1 2 0.083 0.012 0 3 2 0.065 0.017 0.001 Yt/BEt 138.99 2 69.495 5.981 0.003 1 2 2.23 0.646 0.002 dAt/At 10.814 2 5.407 24.561 0 1 3 -0.672 0.104 0 2 3 -0.862 0.131 0 Vt/At 9.238 2 4.619 4.243 0.015 1 3 -0.625 0.231 0.02 2 3 -0.793 0.291 0.019 Lt/At 0.524 2 0.262 4.367 0.013 2 3 0.195 0.068 0.013 Et/At 0.209 2 0.104 31.705 0 Yt/BEt 0.608 2 0.304 3.556 0.03 dAt/At 1.68 2 0.84 6.297 0.002 Vt/At 1996 5.23 ANOVA Mean df Square 5.915 2 2.958 5.384 0.005 1 2 0.069 0.009 0 3 1 -0.061 0.018 0.003 1 3 0.234 0.094 0.035 1 2 0.154 0.059 0.026 3 2 0.417 0.128 0.004 1 3 -0.665 0.238 0.015 2 3 -0.85 0.26 0.003 Lt/At 0.36 2 0.18 4.053 0.018 1 2 -0.094 0.034 0.017 Et/At 0.305 2 0.152 25.96 0 1 2 0.075 0.011 0 3 1 -0.064 0.019 0.002 Yt/BEt Vt/At 2.174 6.69 2 2 1.087 3.345 11.683 6.239 0 0.002 1 2 0.21 0.045 0 3 2 0.265 0.085 0.005 1 3 -0.506 0.182 0.016 2 3 -0.713 0.203 0.001 187 DBA Thesis: Malinee Ronapat Table 5.14: ANOVA Test for Payers, Former Payers and Never Paid Firms (Continued) Year ANOVA Sum of Squares 1997 Et/At dAt/At Lt/At 1998 At 1999 Et/At Yt/BEt 2000 Mean Square 2 0.827 11.861 F Sig. 0 13.802 2 6.901 28.071 0 (I) (J) Mean Difference (I-J) Std. Error 1 2 0.136 0.032 Sig. 3 1 -0.172 0.058 0.01 1 2 0.22 0.06 0.001 0 3 1 0.657 0.11 0 3 2 0.878 0.118 0 2 -0.196 0.042 0 3.055 2 1.528 12.98 0 1 3 1 0.191 0.076 0.033 1.17833E+11 2 58916744022 5.592 0.004 1 2 39057.12 11685.737 0.003 3 1 -29663.95 23106.115 0.405 2.408 2 1.204 8.908 0 1 2 0.175 0.042 0 300.909 2 150.455 4.548 0.011 3.594 2 1.797 7.635 0.001 Vt/At 10.846 2 5.423 3.21 0.041 Lt/At 36.749 2 18.374 18.878 0 1 3 3.654 1.212 0.008 2 3 2.889 1.148 0.033 1 2 0.186 0.055 0.002 3 2 0.247 0.097 0.03 1 2 -0.373 0.147 0.031 1 2 -0.686 0.112 0 3 1 0.503 0.208 0.042 Et/At 3.186 2 1.593 3.451 0.033 1 2 0.199 0.076 0.025 Vt/At 35.687 2 17.843 3.705 0.026 1 2 -0.665 0.246 0.019 69572911571 2 34786455758 3.171 0.043 1 2 -29499.17 11720.662 0.033 Lt/At 80.099 2 40.049 8.336 0 1 2 -1.002 0.245 0 Vt/At 117.491 2 58.745 3.797 0.023 1 2 -1.164 0.431 0.02 73409148645 2 36786455758 3.474 0.032 1 2 -29609.63 11260.019 0.024 189.731 2 94.865 6.332 0.002 1 2 -1.497 0.424 0.001 At Lt/At 2002 df 1.655 dAt/At At 2001 Multiple Comparisons: Tukey HSD Et/At dAt/At Vt/At At Lt/At 0.438 2 0.219 6.086 0.002 1 2 0.069 0.021 0.003 12.497 2 6.249 4.499 0.012 2 3 -0.52 0.189 0.017 18.363 2 9.181 3.618 0.028 1 2 -0.456 0.175 0.025 1.03291E+11 2 51645588378 4.546 0.011 1 2 -38995.56 11694.419 0.011 57.183 2 28.592 14.182 0 1 2 -0.83 0.156 0 Source: Developed for this research Note: Et/At is the ratio of aggregate earnings before interest to aggregate assets, Yt/BEt is the ratio of aggregate common stock earnings over aggregate book equity, dAt/At is the growth rate in assets, Vt/At is the aggregate market value to the aggregate book value of assets, At is total assets, Lt/At is the ratio of total liabilities to total assets. Statistic is significant at 0.05 level. (1) is payers, (2) is former payers (3) is never paid firms. See appendix B for details. Table 5.15: Summary of the Characteristics of Payers, Former Payers and Never Paid Firms 188 DBA Thesis: Malinee Ronapat Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Characteristics confirmed Payers, Former payers and Never paid Profitability Investment Opportunities Size Et/At Yt/BEt dAt/At Vt/At At Lt/At 1(2), 2(3) 1(3) 1(2) 1(2) 1(2), 2(3) 1(2) 1(3), 2(3) 1(2), 2(3) 1(2) 1(3), 2(3) 1(3), 2(3) 2(3) 1(2), 1(3) 1(3) 1(2), 2(3) 1(3), 2(3) 1(2) 1(2), 1(3) 1(2), 2(3) 1(3), 2(3) 1(2), 1(3) 1(2), 1(3), 2(3) 1(2), 1(3) 1(2), 1(3) 1(2) 1(3), 2(3) 1(2), 2(3) 1(2) 1(2), 1(3) 1(2) 1(2) 1(2) 1(2) 1(2) 1(2) 1(2) 1(2) 2(3) 1(2) 1(2) 1(2) Source: Developed for this research Note: 1(2) is payers (1) to former payers (2). 1(3) is payers (1) to never paid firms (3). 2(3) is former payers (2) and never paid firms (3). Et/At is the ratio of aggregate earnings before interest to aggregate assets, Yt/BEt is the ratio of aggregate common stock earnings over aggregate book equity, dAt/At is the growth rate in assets, Vt/At is the aggregate market value to the aggregate book value of assets, At is the total asset, Lt/At is the ratio of total liabilities to total assets. See appendix B for details. The second stage of the ANOVA testing repeated the results of the first stage. Table 5.15 indicates that payers have significantly higher profitability than former payers and firms that have never paid dividends in most years (measured by both Et/At and Yt/BEt). However, the results from the Vt/At ratio are mixed. The analysis of dAt/At indicates that payers have lower investment opportunities than firms that have never paid dividends, but payers have higher investment opportunities than former payers. In relation to firm size there was minimal variation between the three groups before the crisis. However, the value of payers’ assets was significantly smaller than the asset value of former payers after the crisis. This could result from the unsuccessful elimination of NPLs (Non-Performing Loans) and the restructuring which was undertaken by former payers. The results for Lt/At were mixed in the early years, although payers had less liabilities (lower Lt/At or ratio of total liabilities to total assets) than former payers. The ANOVA results are similar to the descriptive tests and confirm that firms which pay dividends are more profitable (as presented by the Et/At ratio) and have low investment opportunities. These firms are well balanced between equity and debt (as 189 DBA Thesis: Malinee Ronapat presented by Lt/At ratio). Former payers are usually distressed firms with low profitability, low investment opportunities and a large asset size. This may have resulted from the NPLs and low quality assets which are owned by these firms. Firms that never paid dividends have been relatively profitable (lower than payers but higher than former payers) and enjoyed very high investment opportunities. It therefore appears that firms are more successful if they have never paid dividends. 5.5.3 REVIEW OF DESCRIPTIVE STATISTIC RESULTS The evidence from the descriptive statistic is largely consistent with the findings of Fama and French (2001) and suggests the three characteristics: profitability, investment opportunities and size are factors affecting the dividend policies of listed firms. However, the relationships between characteristics and dividend payments were confirmed for only some of the years which were investigated. The higher the level of profitability, the more likely firms will pay dividends. This trend is consistent over time and also relevant to the findings of Fama and French (2001). In relation to investment opportunities, evidence from the pre-crisis period suggests that payers have lower investment opportunities than the firms which do not pay dividends and are larger in size than firms in the other groups. All of these findings are consistent with Fama and French (2001). After the crisis, many listed firms tended to downsize and ceased paying dividends. Most of the non-payers were heavily in debt and owned a high proportion of low quality assets. It appears that the characteristics of the payers changed in the latter part of 1997. The asset size of the payers has stabilized since 1997. Indeed, the asset growth of the payer group has been slightly positive, while the other dividend groups engaged in downsizing and liquidated some of their assets. These firms therefore experienced negative growth in their assets. Furthermore, the percentage of firms which pay dividends has fallen and more firms are now non-payers. Lastly, firms that have never paid dividends have greater investment opportunities than payers and this is consistent over time. 190 DBA Thesis: Malinee Ronapat Firms which pay dividends were large in size and profitable before the crisis (confirmed by descriptive statistic approach). However, they tended to have low investment opportunities (confirmed by descriptive and ANOVA approaches). Former payers are distressed firms with low profitability and investment opportunities, but large in terms of assets sizes after the crisis. Firms that have never paid dividends were more profitable and had better investment opportunities than former payers. The newly listed firms were added to the total sample. The findings suggest that new firms tend to be small, but quickly adopt the characteristics of non-payers, or growth stock and contribute to the decline in the percentage of payers. After the crisis, new lists continued to be small and grew rapidly, but their profitability declined. Thus, new lists that pay a dividend become unrare. Therefore, new lists rarely paid dividends. The decline in the percentage of payers from 1991 is partly due to (1) the characteristics of new lists are increasingly becoming similar to firms that do not pay dividends and (2) the Asian Economic Crisis in 1997 resulted in financial distress for many firms (both new and mature) and they ceased paying dividends. The findings suggest there is an unusual pattern in the payment of cash dividends and the percentage of dividend payers during, and a year after the Asian Economic Crisis (1997-1998). However, the descriptive statistics suggest that this trend is temporary and the market appears to be recovering because: (1) the percentage of payers is increasing and (2) the profitability of firms is increasing. A formal analysis of the characteristics of firms will be discussed in the section 5.5.4. 191 DBA Thesis: Malinee Ronapat 5.5.4 CONFIRMATION FROM LOGIT REGRESSIONS Table 5.16 summarises the results of the logit regression which was undertaken to identify the relative importance of the three characteristics (profitability, investment opportunities and size) which influence a firm’s dividend behaviour. Logit regression is used for explaining why firms pay dividends and for developing a formula for predicting the firms which are likely to pay dividends. Profitability is measured as the ratio of a firm’s earnings before interest to its total assets (Et/At) (Fama and French 2001). The proxies for investment opportunities are a firm’s rate of growth of assets, dAt/At and its market-to-book ratio, Vt/At (Fama and French 2001). The size, SET, is measured by the percentage of listed firms with the same, or smaller market capitalization. This will ‘eliminate the effects of the growth in size through time’ (Fama and French 2001, p. 12). Logit regression is estimated yearly and applies the time-series standard deviations of the annual coefficients. This permits a correlation of the regression residuals across firms and ensures judgements about the average coefficient can be made (Fama and MacBeth 1973; Fama and French 2001, p. 12). The t-statistics table shows the mean divided by the standard error. This is the times-series standard deviation of the regression coefficient divided by the square root of the number of years in the period of the study (Fama and French 2001). From 1991 to 2001, the slopes from the regressions match the roles of profitability, investment opportunities and size as characteristics which affect the dividend decisions of firms between 1990 and 2002. 192 DBA Thesis: Malinee Ronapat Table 5.16: Logit Regression to Explain Which Firms Pay Dividends Average Coefficient t-statistic Int SETt Vt/At dAt/At Et/At Int SETt Vt/At dAt/At Et/At 1991-1996 1997-2002 1991-2002 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 1.38 -0.77 0.30 1.52 2.61 1.89 1.11 0.66 0.48 0.29 -2.25 -0.91 -0.68 -0.66 -0.44 0.87 1.85 1.36 1.69 0.25 -0.02 1.09 0.83 1.36 2.01 2.69 1.24 1.74 1.73 1.70 -0.36 -0.59 -0.47 -0.05 -0.59 -0.07 -0.46 -0.64 -0.34 -0.08 -0.25 -0.73 -1.17 -0.54 -0.77 -0.92 0.18 -0.37 0.30 -1.34 -2.85 -1.75 0.30 -0.18 -0.19 0.26 0.53 0.75 0.00 -0.29 14.54 3.14 8.84 -3.04 14.64 15.06 20.82 24.16 15.62 1.79 1.64 4.31 5.58 -0.06 5.61 4.22 -2.28 0.76 3.25* 9.45* 6.27* -1.77 1.05 -1.20 -3.46* -3.69* -4.87* 3.79* 3.27* 3.46* 1991-1996 1997-2002 1991-2002 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 1.12 -1.22 -0.05 1.62 1.89 1.85 0.83 0.25 0.28 0.24 -2.37 -1.37 -1.49 -1.27 -1.05 0.39 1.63 1.01 0.91 0.05 -0.20 0.42 0.11 1.01 1.96 2.49 0.75 1.32 1.76 1.52 -1.01 0.28 -0.36 0.34 -1.51 -2.93 -2.05 0.32 -0.21 -0.19 0.44 0.52 0.80 0.01 0.08 13.53 2.43 7.98 -2.90 11.43 15.01 19.90 22.39 15.37 1.83 1.57 3.97 5.06 -0.07 2.20 3.59 -3.53 -0.12 1.92 6.77* 4.19* -1.81 1.83 -1.09 3.71* 3.25* 3.27* Source: Developed for this research Note: Logit regressions are estimated for each year from 1991-2002. The dependent variable is 1.0 in year t if the firm is a payer and 0.0 otherwise. The explanatory variables are Et/At for profitability, dAt/At for investment opportunities, Vt/At for investment opportunities and SETt for size. SETt represents the percentage of SET firms with an equal, or rate of lower market capitalization in time t. For the range of years, means of the regression intercepts and independent variables. The t-statistics table is defined as the mean divided by the standard error which provides the coefficient means. (*) the figure is significant at <0.05 level. See appendix C, D, E and F for details. 193 DBA Thesis: Malinee Ronapat During the pre-crisis period (1991-1996), larger firms tended to pay dividends. This is confirmed by the average slope of 3.25 of SETt. Firms with higher profitability are more likely to pay dividends. The slope on Et/At is 3.79. In relation to investment opportunities, the t-statistics suggest that payers tend to have lower investment opportunities than all other firms, due to relatively low growth in their assets (-3.46) and Vt/At (-1.77). The coefficient for each year from 1991 to 1996 confirmed that Thai firms which paid dividends possess the three characteristics which were suggested by Fama and French (2001). This finding is valid especially for Et/At, SET. The ratio Et/At showed a high positive coefficient every year from 1991 to 1996 in addition to SET. The coefficient for Vt/At was negative in most years (table 5.15). In relation to the post-crisis period (1997-2002), the logit regression suggests that large firms with large capitalisation are more likely to pay dividends. This is confirmed by the slope of 9.45 for SETt. The t-statistics suggest that profitable firms tend to pay dividends. In this time period, the investment opportunities of firms is consistent with Fama and French (2001) because the slope for the growth in assets is negative (-3.69) and Vt/At (1.05) is positive. Therefore, firms which pay dividends tend to have low investment opportunities. In relation to the period 1991 to 2002, the slopes from the regressions confirm that firms with high profitability (3.46) and of larger size (6.27) are more likely to pay dividends. This finding is consistent with Fama and French (2001). Furthermore the two measures of investment opportunities are also consistent with Fama and French (2001). These ratios have negative slopes of -4.87 and -1.2 respectively. The result for the growth in assets are -4.87. This indicates that firms with low investment opportunities tend to pay dividends. The Vt/At ratio is also negative (-1.2) and infers 194 DBA Thesis: Malinee Ronapat that firms with low investment opportunities tend to pay dividends. This conclusion is also similar to the evidence which is presented in the literature. The conclusion that firms which pay dividends tend to be more profitable, larger and experience low investment opportunities is consistent with Fama and Babiak (1968), Easterbrook (1984), Jensen (1986), Fama and French (1988 1995, 1997, 1998, 2001), DeAngelo DeAngelo and Skinner (2002), Aivazian, Booth and Cleary (2003). Furthermore, Fama and French (1999) suggested that dividend paying firms are profitable or have higher payout ratios than firms with higher investment opportunities (Fama and French 2001). Firms with high investment opportunities tend to achieve high growth with small value. This is consistent with the characteristics of non-payers (Fama and French 2001). Hence, high growth firms are reluctant to issue equities because the trasaction costs are high (Myers 1984, Fama and French 1999, Fama and French 2001). In short, the findings for Thailand are consistent with the literature (see table 5.9). Figure 5.6 re-states the hypotheses which were developed in chapter 3. Controlling for one investment variable, suggests similar results to the earlier parts. Payers, before, during and after the crisis, have high profitability and size but low investment opportunities (see table 5.16). 195 DBA Thesis: Malinee Ronapat Figure 5.6: Hypotheses Revisiting CHARACTERISTICS OF DIVIDEND PAYERS BEFORE THE ASIAN ECONOMIC CRISIS (Step A) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividend before the Asian Economic Crisis. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividend before the Asian Economic Crisis. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividend before the Asian Economic Crisis. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividend before the Asian Economic Crisis. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividend before the Asian Economic Crisis. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividend before the Asian Economic Crisis. CHARACTERISTICS OF THE DIVIDEND PAYERS DURING AND AFTER THE ASIAN ECONOMIC CRISIS (Step B) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividend after the Asian Economic Crisis. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividend after the Asian Economic Crisis. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividend after the Asian Economic Crisis. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividend after the Asian Economic Crisis. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividend after the Asian Economic Crisis. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividend after the Asian Economic Crisis. CHARACTERISTICS OF THE DIVIDEND PAYERS FOR THE WHOLE PERIOD (Step C) (1) PROFITABILITY Ho : Ha : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividend for the whole period of investigation. There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividend the whole period of investigation. (2) SIZE Ho : Ha : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividend the whole period of investigation. There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividend the whole period of investigation. (3) INVESTMENT OPPORTUNITIES Ho : Ha : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividend the whole period of investigation. There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividend the whole period of investigation. Source: Developed for this research 196 DBA Thesis: Malinee Ronapat 5.5.5 HYPOTHESES TESTING USING LOGIT REGRESSION (AVERAGE COEFFICIENT) Each year during the investigation period (1991 to 2002) has its own coefficient. As stated earlier, the twelve year period (1991 to 2002) is grouped into three steps (A, B, and C). Step A represents the period 1991 to 1996 (before the crisis), step B the period from 1997 to 2002 (during and after the crisis), and step C is while period from 1991 to 2002. The coefficient of each variable, in each year, is averaged to obtain an average coefficient for each step (A, B and C) (Fama and MacBeth 1973; Fama and French 2001). The hypotheses tested by logit regression were also divided into the three steps (A, B, and C). Step A’s null hypothesis is to test each characteristic of the firms and their relation to the probability that firms will pay dividends before the crisis (1991-1996), after the crisis in step B (1997-2002) and for the whole period of investigation in step C (1991-2002). Table 5.16 presents the logit regression for two sets of results (1) controlling two variables for investment opportunities (Et/At, dAt/At, Vt/At and SETt) and (2) controlling one variable for investment opportunities (Et/At, dAt/At and SETt). As stated in chapter 3, the hypothesis from step A (1) states there is no significant relationship between the profitability of the firms and the probability that listed firms will pay dividends before the Asian Economic Crisis (Ho). With the ratio of Et/At (the ratio of aggregate earnings before interest to aggregate assets) as a measure of profitability, the logit regression results indicate that the null hypothesis is rejected (0.012). This infers there is a relationship between a firm’s profitability and its probability to pay dividends before the crisis (table 5.16). Table 5.16 indicates that the value of the t-statistic is 3.79 for Et/At. This was calculated from the average coefficient during 1991 to 1996. The null hypothesis (1) in steps B (1997-2002) and C (1991-2002) states that there is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends in step B (0.022) and C (0.005) (Ho) is also rejected. 197 DBA Thesis: Malinee Ronapat Table 5.16 presents the t-statistics for Et/At (earnings to total assets) for the period 1997 to 2002 (3.27) and the period 1991 to 2002 (3.46). Hence, there is a relationship between the firm’s profitability and the probability that it will pay dividends before the crisis, after the crisis and during the whole period of investigation. Firms with higher profitability are more likely to pay dividends. Therefore, Et/At (earnings to total assets) has significant power for predicting the dividend behaviour of firms. When controlling only one investment variable, the null hypothesis (1) of steps A (0.013), B (0.022) and C (0.007) is also rejected (Ho). The lower part of table 5.16 shows that the value for the t-statistics of Et/At (earnings to total assets) are highly positive for steps A (3.71), B (3.25) and C (3.27). This infers that Et/At (earnings to total assets), as a measure of a firm’s profitability, is a significant variable for predicting the dividend behaviour of the firm in every time period. Therefore, the higher the level of profitability, the more likely a firm will pay dividends. Null hypothesis number 2 from step A (before the crisis), B (during and after the crisis) and C (the whole period of investigation) stated that there is no significant relationship between the size of the firms and the probability that these firms will pay dividends (Ho). SETt (the percentage of SET firms with the same or a lower level of market capitalization) is a variable which represents the size of the firm. The null hypotheses (Ho) in step A (3.25), B (9.45) and C (6.27) are rejected. Table 5.16 presents the value of SET in all three steps. This infers that the firm size is significant for predicting dividend behaviour in all time periods (1991-1996 (0.02), 1997-2002 (0.000) and 19912002 (0.000)). Consequently, larger firms are more likely to pay dividends. By controlling one investment variable, the variable SETt represents the size of the firm. However, null hypothesis number 2 in step A (0.11) is accepted. This infers there is no relationship between the size of the firm and its dividend behaviour after the crisis (Ho). Table 5.16 indicates that the t-statistic for SET is 1.92 for the period 1991 to 1996. For groups B (0.001) and C (0.001), null hypothesis number 2 is rejected. This infers there is a relationship between a firm’s size and its dividend behaviour. The results indicate that the variable SETt in steps B and C has significant power for 198 DBA Thesis: Malinee Ronapat predicting the dividend behaviour of a firm. In addition, firms with high investment opportunities are less likely to pay dividends. Null hypothesis number 3 in steps A (1991-1996) (0.018), B (1997-2002) (0.014) and C (1991-2002) (0.000) states that there is no significant relationship between the investment opportunities of the firms and the probability that they will pay dividends (Ho). The investment opportunities are represented by two variables, Vt/At (the marketto-book ratio) and dAt/At (the growth rate of assets). Table 5.16 indicates that the tstatistics (for dAt/At) are strongly negative for group A (1991-1996), B (1997-2002) and C (1991-2002). Therefore, the null hypotheses are rejected in all groups. This infers that there is a negative relationship between investment opportunities (measured by dAt/At) and the probability that firms will pay dividends in step A (-3.46), B (-3.69) and C (-4.87). Investment opportunities (dAt/At) is a significant variable for predicting the dividend behaviour of the firm. If a firm’s investment opportunities rise, it will be less likely to pay dividends. Table 5.16 indicates that the t-statistics for Vt/At (book value to total assets) in every time period (A (0.137), B (0.340), C (0.253)), are small and the signs of t are randomly positive and negative. This infers that the null hypotheses can not be rejected and there is no significant relationship between investment opportunities (measured by Vt/At) and the dividend behaviour of firms in every period. Therefore, when controlling two investment variables only dAt/At (growth rate of assets) is significant for predicting the dividend behaviour of the firm in steps A, B and C. The variable Vt/At was found to be insignificant. After controlling one investment variable, null hypothesis number 3 is accepted for steps A (0.129), B (0.127) and C (0.300). The values for the t-statistic in table 5.16 are randomly positive and negative and are not sufficiently significant to reject the null hypothesis. This infers that dAt/At, (when used as the only measure of investment opportunities) has no relationship and is not a significant predictor of a firm’s dividends behaviour. 199 DBA Thesis: Malinee Ronapat Table 5.17 summarises the results of hypotheses testing the applying average coefficient method. Table 5.17: Summary of Hypothesis Testing Results (Average Coefficient) Periods Controlled Variables 2 Variables Before the Crisis (Group A) 1991-1996 Characteristics Profitability Size Investment (dAt/At) Accepted Investment (Vt/At) Profitability Size Investment (dAt/At) 1 Variable Profitability Size Investment (dAt/At) Investment (Vt/At) 2 Variables During and After the Crisis (Group B) 1996-2002 Profitability Size Investment (dAt/At) 1 Variable The Whole Period of Investigation (Group C) 1991-2002 Accepted Accepted Rejected Rejected Rejected Accepted Rejected Rejected Accepted Accepted Investment (Vt/At) 1 Variable Rejected Rejected Rejected Rejected Profitability Size Investment (dAt/At) 2 Variables Null Hypothesis Rejected Rejected Rejected Profitability Size Investment (dAt/At) Rejected Rejected Accepted Source: Developed for this research Note: See appendix C, D, E and F for details. 5.5.6 HYPOTHESES TESTING USING LOGIT REGRESSION (GROUP COEFFICIENT) Different from average coefficient, calculating coefficient using firm period (hereafter group coefficient) allows data to be entered simultaneously into the logit regression equation to identify the coefficient for each step. Each year during the investigation period (1991 to 2002) has its own coefficient. As stated earlier, the twelve years (1991 200 DBA Thesis: Malinee Ronapat to 2002) are grouped into three steps (A, B, and C). The coefficient of each variable in each step is observed by running a logit regression. Tables 5.18 and 5.19 present the results of the logit regression. These results were obtained by using the group coefficient controlling two investment variables (Et/At, dAt/At, Vt/At and SETt) and one investment variable (Et/At, Vt/At and SETt) respectively. Where: Et/At is the ratio of aggregate earnings before interest to aggregate assets, dAt/At is the growth rate in assets, Vt/At is the aggregate market value to the aggregate book value of assets, SETt represents the percentage of SET firms with an equal or lower rate of market capitalisation in time t. Table 5.18: Logit Regression Results (Controlling Two Investment Variables) Variables in the Equation Group A (1991-1996) Variables in the Equation Group B (1997-2002) Variables in the Equation Group C (1991-2002) SETt Vt/At dAt/At Et/At Constant SETt Vt/At dAt/At Et/At Constant SETt Vt/At dAt/At Et/At Constant B 0.630 -0.054 -0.577 12.705 0.809 B 1.699 -0.464 0.156 0.059 -0.676 B 0.741 0.005 0.465 1.491 -0.009 S.E. 0.255 0.018 0.195 1.229 0.142 S.E. 0.167 0.082 0.093 0.096 0.115 S.E. 0.119 0.014 0.074 0.212 0.066 Wald 6.079 8.559 8.788 106.957 32.597 Wald 103.767 32.339 2.808 0.381 34.352 Wald 38.657 0.124 39.226 49.547 0.019 df 1 1 1 1 1 df 1 1 1 1 1 df 1 1 1 1 1 Sig. 0.014 0.003 0.003 0.000 0.000 Sig. 0.000 0.000 0.094 0.537 0.000 Sig. 0.000 0.725 0.000 0.000 0.890 Exp(B) 1.877 0.947 0.562 1.000 2.246 Exp(B) 5.467 0.629 1.169 1.061 0.509 Exp(B) 2.098 1.005 1.593 4.443 0.991 Source: Developed for this research Note: The explanatory variables are Et/At for profitability, dAt/At for investment opportunities, Vt/At for investment opportunities and SET for size. SETt represents the percentage of SET firms with equal or lower market capitalization rate in time t. Statistic is significant at 0.05 level. The constant is the autonomous and B is the beta coefficient (Ticehurst and Veal 2000). Table 5.19: Logit Regression Results (Controlling One Investment Variable) 201 DBA Thesis: Malinee Ronapat Variables in the Equation Group A (1991-1996) Variables in the Equation Group B (1997-2002) Variables in the Equation Group C (1991-2002) SETt dAt/At Et/At Constant SETt dAt/At Et/At Constant SETt dAt/At Et/At Constant B 0.473 -0.610 12.454 0.802 B 1.523 0.249 0.139 -1.085 B 0.747 0.459 1.502 -0.004 S.E. 0.248 0.196 1.215 0.142 S.E. 0.159 0.096 0.132 0.094 S.E. 0.118 0.074 0.212 0.065 Wald 3.643 9.685 105.001 32.063 Wald 92.053 6.708 1.114 133.582 Wald 39.983 38.399 50.056 0.005 df 1 1 1 1 df 1 1 1 1 df 1 1 1 1 Sig. 0.056 0.002 0.000 0.000 Sig. 0.000 0.010 0.291 0.000 Sig. 0.000 0.000 0.000 0.945 Exp(B) 1.605 0.544 1.000 2.231 Exp(B) 4.587 1.283 1.150 0.338 Exp(B) 2.110 1.583 4.492 0.996 Source: Developed for this research Note: The explanatory variables are Et/At for profitability, dAt/At for investment opportunities and SET for size. SETt represents the percentage of SET firms with equal or lower market capitalization rate in time t. Statistic is significant at 0.05 level. The following results are presented in table 5.18: Et/At (0.000), dAt/At (0.003), Vt/At (0.003) and SETt (0.014). In step A, the null hypothesis is rejected for each component of the firms’ characteristics (profitability, investment opportunities and size). In other words, variable Et/At (earnings to total assets), dAt/At (growth rate of assets), Vt/At (market value to total assets) and SETt (SETt represents the percent of SET firms with equal or lower market capitalization rate) are significant for predicting the dividend payment behaviour of firms. In step A (1991-1996), large firms with high profitability, and low investment opportunities are more likely to pay dividends. However, the null hypothesis is rejected in step B for Et/At (0.537) and dAt/At (0.094) and it is accepted in step B for Vt/At (0.000) and SETt (0.000). Therefore, during and after the Asian Economic Crisis, only Vt/At and SETt are significant for predicting the dividend behaviour of firms. During and after the Asian Economic crisis (step B), large firms with low investment opportunities (measured by Vt/At or the ratio of market value to total assets) are more likely to pay dividends. In step C (1991-2002), the logit regression suggests that the hypotheses number 1 and 2 should be rejected. In relation to null hypothesis 3, dAt/At (growth rate of assets) should 202 DBA Thesis: Malinee Ronapat be rejected and Vt/At, is accepted (0.725). This infers that, dAt/At, (a measure of investment opportunities) is a significant variable for predicting dividend behaviour in step C (1991-2002) while Vt/At (market value to total assets) is insignificant. Therefore, over the period 1990 to 2002, large firms with high profitability and positive growth in their assets are more likely to pay dividends. Table 5.19 presents the results of the logit regression after controlling one investment variable. The variables include: Et/At (earnings to total assets), dAt/At (growth rate of assets), and SETt (SETt represents the percent of SET firms with equal or lower market capitalization rate). Table 5.17 indicates in step A, null hypothesis number 2 is accepted (0.056) while number 1(0.000) and 3 (0.002) are rejected. This infers that investment opportunities (dAt/At) and profitability (Et/At) are significant variables for predicting the dividend behaviour of a firm and size (SETt) is insignificant between 1991 and 1996. In short, before the Asian Economic Crisis, highly profitable firms with low investment opportunities were more likely to pay dividends. In step B, null hypothesis number 1 is accepted (0.291) because there is no relationship between profitability and the dividend behaviour of firms. Null hypotheses numbers 2 (0.000) and 3 (0.10) are also rejected. This infers that firm size and investment opportunities are significant for predicting the dividend behaviour of a firm. Consequently, in step B (1997-2002), firms which are large and experiencing high growth in their assets are more likely to pay dividends. Lastly, step 3 considers the period 1991 to 2002 and null hypotheses numbers 1 (0.000), 2 (0.000) and 3 (0.000) were rejected. This implies that profitability, size and investment opportunities are significant variables when predicting the dividend behaviour of a firm. This suggests that large firms which experience high profitability and a positive growth rate in their assets are more likely to pay dividends. Table 5.20 shows the summary of the hypotheses testing results applying group coefficient method. 203 DBA Thesis: Malinee Ronapat Table 5.20: Summary of Hypothesis Testing Summary (Group Coefficient) Periods Controlled variables 2 Variables Before the Crisis (Group A) 1991-1996 Characteristics Profitability Size Investment (dAt/At) Rejected Investment (Vt/At) Profitability Size Investment (dAt/At) 1 Variable 2 Variables During and After the Crisis (Group B) 1996-2002 The Whole Period of Investigation (Group C) 1991-2002 Rejected Accepted Profitability Size Investment (dAt/At) Rejected Rejected Rejected Accepted Investment (Vt/At) Rejected Profitability Size Investment (dAt/At) 1 Variable 2 Variables Null Hypothesis Rejected Rejected Rejected Accepted Rejected Profitability Size Investment (dAt/At) Rejected Rejected Rejected Rejected Investment (Vt/At) Accepted 1 Variable Profitability Size Investment (dAt/At) Rejected Rejected Rejected Source: Developed for this research Similar to the average coefficient method, the group coefficient method has identified firms characteristics which influence the dividend behaviour of firms. It is observed that testing null hypotheses applying the group coefficient method results in less null hypotheses being accepted (see table 5.17 and 5.20). However, the rest of this study, in the spirit of Fama and MacBeth (1973) and Fama and French (2001), will apply the average coefficient method to predict the dividend behaviour and present the propensity to pay dividends of listed firms. Hosmer and Lemeshow Test for goodness of fit was undertaken to ensure that this study is consistent with the methodology (average coefficient) of Fama and French. The results of this analysis are presented in table 5.21. 204 DBA Thesis: Malinee Ronapat Table 5.21: Hosmer and Lemeshow Test (Goodness of Fit) Step A Step B Step C Hosmer and Lemeshow Test Chi-square df 46.587 8 C1 49.184 8 C2 92.958 8 C1 99.265 8 C2 130.946 8 C1 130.742 8 C2 Sig. 0.000 0.000 0.000 0.000 0.000 0.000 Source: Developed for this research Note: Step A covers 1991 to 1996. Step B covers 1997 to 2002. Step C covers 1991-2002. C 1 and C 2 controlled two and one investment variable(s) respectively. The statistics are significant at the 0.05 level. The table presents the chi-square of the Hosmer and Lemeshow test and indicates that the data was well fit in every group and step (0.000). This infers that the data which was collected before, during and after the crisis was sufficient for predicting or explaining the dependent variable (payers/non-payers) (Tabachnick and Fidell 2001). Therefore, it is suggested that the logit regression model could be a valuable tool for predicting and identifying firms that are likely to pay dividends. 5.6 THE PROPENSITY TO PAY DIVIDENDS (DESCRIPTIVE STATISTICS EVIDENCE) Fama and French (2001) suggested that changes in firm characteristics may not be the only factor affecting the payment of dividends. The researchers suggested that given the characteristics of firms, they will be less likely, rather than unlikely, to pay dividends (Fama and French 2001). Therefore, it is important to investigate changes in propensity to pay dividends of listed firms since 1991. Three factors are considered to identify the propensity to pay dividends. Yt is a proxy for common stock earnings (representing the profitability), Et represents earnings before interest that exceed investment outlays and dAt is the change in assets (representing investment opportunities and asset growth) (Fama and French 2001). 205 DBA Thesis: Malinee Ronapat A firm’s propensity to pay dividends is expressed by a time series plot in figure 5.7. This figure shows the percentage of dividend payers among (1) firms with positive common stock earnings (Y>0) (2) firms with negative common stock earnings (Y<0) (3) firms with earnings before interest that exceed investment outlays (Et>dAt) and (4) firms with earnings before interest which are lower than investment outlays (Et<dAt) (Fama and French 2001). Firms with earnings before investment, which are higher than the growth in their assets have low investment opportunities and are low growth firms. Likewise, firms with earnings before investment which are lower than changes in their assets are considered growth firms (Fama and French 2001). Figure 5.7 indicates that most firms with positive common stock earnings (83.93 percent) paid dividends in 1991 and 100 percent of firms with negative earnings also paid dividends. In addition, 91.7 percent of firms with Y>0 paid dividends in 1992 although this figure declined after 1992. This figure fell to its minimum level in 1998, when only 38.19 percent of firms with positive earnings paid dividends. Furthermore, in 1999, 13.10 percent of firms with negative earnings paid dividends. Importantly, the percentage of firms with positive earnings that paid dividends almost doubled from 39.19 percent in 1998 to 56.7 percent in 2002. This analysis suggests that firms with positive earnings exhibited a small decline in the propensity to pay dividends before the Asian Economic Crisis (96.95 percent to 91.26 percent). During the crisis, (1997 to 1998), firms with Yt>0 exhibited a substantial decline in the propensity to pay a dividends (89.92 percent in 1997 to 39.19 percent in 1998). It appears that firms with positive earnings were not willing to pay dividends during the crisis. Firms with negative earnings were also less willing to pay dividends during the crisis. However after the crisis, the propensity to pay dividends these groups has increased. The propensity to pay of firms with Yt>0 increased from 39.19 percent in 1998 to 56.7 percent in 2002, while the propensity to pay of firms with Yt<0 increased from 5.76 percent in 1999 to 7.78 percent in 2002. In short, the firms in the sample are more 206 DBA Thesis: Malinee Ronapat willing to pay dividends if they have positive earnings, or negative earnings after the crisis. This finding was also supported by the descriptive statistics. Similarly, before the crisis, firms with earnings before interest which are higher, or lower than changes in their assets (investment) exhibit a slight decline in the propensity to pay dividends from 76.25 percent to 75.43 percent (when Et>dAt) and 88.59 percent to 83.25 percent (when Et<dAt). During the crisis, firms with Et>dAt and Et<dAt exhibited a significant decline in the propensity to pay a dividends from 69.01 percent to 28.53 percent and 69.66 percent to 24.36 percent respectively. After the crisis, both types of firms exhibit an increase in the propensity to pay dividends. The propensity to pay of firms with earnings higher than investment increased from 30.2 percent in 1999 to 47.52 percent in 2002. Firms with lower earnings than investment, payers of this group (Et<dAt) accounted for 26.62 percent in 1998 and this increased to 40.72 percent in 2002 (figure 5.7). These findings suggest that the Asian Economic Crisis also influenced the propensity to pay dividends of listed firms. 207 DBA Thesis: Malinee Ronapat Figure 5.7: Percentage of Payers Among Firms with (1) Positive Earnings (2) Negative Earnings (3) Earnings above Investment, and (4) Earnings Below Investment Percentage of Payers Among Firms with (i) positive earnings, (ii) negative earnings, (iii) earnings above investment, and (iv) earnings below investment Y>0 Y<0 E>dA E<dA 120.00 100.00 Percent 80.00 60.00 40.00 20.00 0.00 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Year Source: Developed for this research In summary, the results suggest that irrespective of whether firms have high or low profitability and high or low investment opportunities, they demonstrated a lower propensity to pay dividends before the crisis. The propensity to pay dividends of these types of firms, declined significantly during the crisis, although this has gradually recovered since the downturn. However, the percentage of payers with positive earnings in recent years (2002) has been lower than the figure which was observed before the crisis. This indicates that firms with high profitability and investment opportunities now have a lower propensity to pay dividends. 208 DBA Thesis: Malinee Ronapat 5.7 CHANGING CHARACTERISTICS AND PROPENSITY TO PAY DIVIDENDS (LOGIT REGRESSIONS) This section focuses on the quantitative analysis which is used to confirm changes in the characteristics of firms and the propensity to pay dividends. The analysis will estimate the probability that firms paid dividends between 1991 and 1996 (pre-crisis), with given characteristics (profitability, investment opportunities and size). The next step will be to apply these probabilities (those observed during 1991-1996) to the sample firms with their observed characteristics from 1997 to 2002 (post-crisis). This will estimate the expected percentage of dividend payers. Any variation in the expected percentage of payers after 1996 will result from changes in characteristics of the sample firms (Fama and French 2001). The actual percentage of payers is the observed number of payers, to the total number of firms in the sample from 1991 to 2002. Any differences between the expected percentage of payers (calculated from the base probabilities in 1991-1996) and actual percentage of payers will represent a change in the propensity to pay dividends. Positive variation between the expected and actual percentage of payers will imply a decline in the propensity to pay dividends, while negative variation will imply an increase in the propensity to pay dividends. Expected is the expected percentage of payers. Actual is the actual percentage of firms that pay dividends in time t. Expected – Actual is the change in propensity to pay dividends over time. Decline in the expected percentage of payers represents changes in the characteristics of the payers since the base period. 5.7.1 REGRESSION ESTIMATES Table 5.22 presents the expected percentage of dividend payers which was obtained by applying the average coefficient to the proxies of the characteristics of the sample firms 209 DBA Thesis: Malinee Ronapat from 1991 to 2002. Two sets of results are presented so that the methodology is consistent with Fama and French (2001). Firstly, proxy for profitability is Et/At while the two proxies for investment opportunities dAt/At and Vt/At. Second, one proxy biased for profitability but only one controlled proxy is used for investment opportunities. While the other proxies for profitability, investment opportunities and size are constant over time, Vt/At is variable (Fama and French 2001, p. 15). Fama and French (2001) stated that the elimination of Vt/At is due to the unstable and unclear tendency of the Vt/At trend (see table 5.8) which results from three factors: (1) increasing profitability of assets in place (2) more profitable or more abundant expected investments, or (3) lower discount rates for expected cash flows (Fama and French 2001, p. 15) Therefore, Vt/At results in an over-statement of the level of investment opportunities and the level of understating of the propensity to pay dividends. 210 DBA Thesis: Malinee Ronapat Table 5.22: Logit Regression Estimates of the Effect of Changing Firm Characteristics and Decline in the Propensity to Pay Dividends 19911996 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Firms Payers 302 225 255 289 322 345 378 394 389 385 369 377 388 256 192 230 245 275 290 306 281 110 115 130 153 179 Vt/At and dAt/At Expected Actual Expected - Actual 84.7 85.3 90.2 84.8 82.8 84.1 81.0 71.3 28.3 29.9 35.2 40.6 46.1 88.8 98.8 90.7 97.6 98.3 97.9 62.3 86.3 81.0 85.3 88.8 91.7 3.5 8.6 5.9 14.8 14.2 16.9 -9.0 58.0 51.1 50.1 48.2 45.6 dAt/At Actual Expected Expected - Actual 85.3 90.2 84.8 82.8 84.1 81.0 71.3 28.3 29.9 35.2 40.6 46.1 99.6 99.2 93.0 98.5 98.5 97.6 61.8 86.9 80.8 86.4 88.8 93.0 14.3 9.0 8.2 15.7 14.4 16.6 -9.5 58.6 50.9 51.2 48.2 46.9 Source: Developed for this research Note: To explain why firms pay dividends, all firms for 1991-1996 are used to estimate the logit regression. The explanatory variables are Et/At (for profitability), dAt/At (for investment opportunities), Vt/At (for investment opportunities) and SET (for size). SETt is the percentage of SET firms with the same, or a lower rate of market capitalization. The firm column presents the number (or average number) of firms in the sample in time t. Payers is the number (or average number) of sample firms that paid dividends in time t. The actual percentage is the percentage of payers to the number of sample listed firms. Expected is obtained from applying the average logit regression coefficient for 1991-1996 to the values of the explanatory variables mentioned above for each firm in every year from 1991 to 2002. This summing over firms and divided by number of firms then multiply by 100. The changes in the percentage of the expected column represents the effect of changing characteristics of the listed firms over time. Expected – Actual measures the effect of the propensity to pay dividends of the listed firms. Two sets of results are presented, one with Vt/At and dAt/At as a measurement of investment opportunities. Another only uses dAt/At as a measurement of investment. See appendix G and H for details. This study conducts regression by using Et/At, dAt/At, Vt/At and SETt first as the factors for explaining the expected percentage of payers in each year from 1991-2002. The base years (1991-1997) are applied to each of the twelve years (1991-2002) to estimate the expected percentage of payers. Any change in the expected percentage of payers in any year is due to changes in characteristics of the sample. Before the Asian Economic Crisis (the base year period 1991-1996), the expected percentage of payers is above 90 percent for most years. This implies that firm characteristics have changed little since 211 DBA Thesis: Malinee Ronapat the base years. When the expected percentage of payers is subtracted from the actual percentage of payers, the propensity to pay dividends of the samples in the pre-crisis period is high. However, the figure tends to increase slightly over time from a variation of 3.5 percent in 1991 and 16.9 percent in 1996. During the Asian Economic Crisis (1997), most of the samples exhibit a decline largely in the characteristics from the base years (62.3 percent expected as payers). The number of actual payers rose to 71.3 percent in the market. Therefore, in 1997, the propensity to pay dividends of the samples increased (differences -9.0 percent). This may be due to (1) an expectation by the firms that the crisis will not last long or (2) the result of policies by firms aimed at stopping investors from panicking and raising the level of confidence. However, the Asian Economic Crisis was long lasting and severally affected the Thai stock market. In 1998, the characteristics of most of the sample firms tended to be similar to the characteristics of payers in the base years (86.3 percent). However, firms were not willing to pay dividends out of their earnings. This resulted in a strong divergence between the expected percentage of payers and the actual percentage of payers (59 percent). Therefore, the sample of listed firms had the lowest propensity to pay in 1998. After the crisis (1999-200), the number of expected papers increased from 81 percent in 1999 to 91.7 percent in 2002. This implies that the firm characteristics which were present in the base years returned. However, the propensity to pay dividends was large in 1998 (59 percent). Firms were not willing to pay dividends even though they had the ability to pay (given these characteristics). After 1998, the difference between the expected and the actual percentage of payers fell from 59 percent in 1998 to 45.6 percent in 2002. This indicates that a recovery is under way and the propensity to pay dividends of the sample firms is increasing. This raises the expectation that (1) the base year characteristics may return and (2) the propensity to pay dividends of listed firms may rise. Fama and French (2001) omitted the Vt/At ratio because they anticipated that it would reduce the percentage of expected payers. This would reduce the contribution to the decline in the percentage of dividend payers (Fama and French 2001). After discussing the Vt/At ratio, this study found that the expected percentage of payers increased. Fama 212 DBA Thesis: Malinee Ronapat and French (2001) concluded that ‘with or without Vt/At, the regression approach bares the tracks of a potentially elusive phenomenon’ (Fama and French 2001, p. 17). This was supported by the second set of findings of Fama and French (2001) and by this study. Table 5.22 indicates there is no significant difference in the findings of this analysis if two variables are controlled, or if one variable is controlled. Before the crisis, there was minimal change in the characteristics of firms. Therefore, only a slightly lower propensity to pay dividends was recorded. During the crisis, more firms were willing to pay dividends because firms characteristics differed from those observed during the base years. Therefore, during the crisis sample firms tended to have a higher propensity to pay dividends. After the crisis, the characteristics of firms were similar to the base years and the propensity to pay dividends rose slightly from a variation of 50.9 percent in 1999 to 46.9 percent in 2002. The evolution of Expected Percentage column (see table 5.22) indicates that the impact of characteristics of firms on the percentage of payers was below 90 percent in only 1991. From 1992 to 1996, the Expected Percent column increased to over 90 percent and has been above this level ever since. In 1997, the characteristics of payers which are similar to the base period fell strongly to 62.3 percent. However, this figure has increased and the characteristics of firms were similar to the base years after 1998. Recently (2002), the Expected Percentage stood at 91.7 percent. This figure suggests that the characteristics of the payers which were observed during the base period have returned. The logit regression of Fama and French (2001) indicated that changes in firm characteristics were maintained over time. The results of this analysis are not consistent with this result. This study indicates the characteristics of payers which existed established during the base years have largely been maintained throughout the period of investigation. These characteristics changed only a short period during the Asian Economic Crisis. Logit regression suggests that changes in firm characteristics and the propensity to pay dividends has resulted in the partial disappearance of dividends from Thailand’s capital 213 DBA Thesis: Malinee Ronapat market, before, during and after the crisis. However, the evidence suggests that before the crisis, there was slight change in the characteristics of firms and the propensity to pay dividends. During the crisis, large changes in firm characteristics were observed and the propensity to pay dividends remained relatively high. After the crisis, propensity to pay the dividends has slowly risen, although it is below the figures observed before the crisis. This outcome appears to result from changes in the characteristics of firms. 5.8 CONCLUSION From a peak of 90.6 in 1992, the percentage of dividend payers in the Stock Exchange of Thailand currently stands at 46.4 percent. The decline is partly due to the fact that many newly listed firms and mature firms are adopting the characteristics of firms which do not pay dividends. Changes in the characteristics of firms are also an important factor leading to the observed decline in the payment of dividends. The methodology which has been adopted for this study suggests that the lower propensity, or reduced willingness to pay dividends among listed firms is common. The next chapter will summarise findings of this study and discuss its implications for future research. 214 DBA Thesis: Malinee Ronapat C HAPTER 6.1 6 CONCLUSIONS AND IMPLICATIONS INTRODUCTION The previous chapter presented the results of the analysis of the characteristics of listed firms, the propensity to pay dividends and their changes over time in Thailand. This chapter is a conclusion to this study. The objectives of this chapter are to summarise (1) the characteristics of firms and propensity to pay dividends of firms which pay dividends (2) and indicate how the findings of this study can be implemented by investors, analysts, regulators and (3) to suggest areas where further research is needed. The chapter is presented in six sections. Section 6.1 discusses the objectives and structure of the chapter 6. Section 6.2 addresses the research hypotheses, questions and problem. The implications of this study and its contribution to the body of knowledge are presented in sections 6.3 and 6.4. Section 6.5 discusses the limitations of the study and finally, section 6.6 provides several recommendations for further research. 215 DBA Thesis: Malinee Ronapat Figure 6.1: The Structure of Chapter Six 6.1 Introduction 6.2 Conclusion 6.2.1 Research hypotheses and 6.2.2 Research problem questions 6.3 Implications 6.3.1 Implications for theory 6.3.2 Implications for policy 6.3.3 Implications for investment practices 6.4 Contributions of the study 6.5 Limitations of the study 6.6 Implications for further research Source: Developed for this research 216 DBA Thesis: Malinee Ronapat 6.2 CONCLUSIONS This section summarises the findings of the analysis of data which was conducted in chapter 5 and addresses the research hypotheses, questions and the research problem. 6.2.1 CONCLUSIONS – RESEARCH HYPOTHESES AND RESEARCH QUESTIONS Research Questions and Hypotheses Several research questions and hypotheses were developed to answer the research problem which was developed in chapter 3. Two sets of research questions were also developed. The first set of research questions related to the literature review, while the second set related to the variables which were used for developing the formula which predicts the dividend behaviour of firm. Both sets of research questions and hypotheses are presented below. 6.2.1.1 CONCLUSIONS RELATING TO THE CHAPTERS Research question # 1 is restated below: What is the background and performance of the Stock Exchange of Thailand and its listed firms? This research question was answered in chapter 2. The history of SET has been short although it provides a vital role in supporting the economic growth of the nation. However, the performance of SET has varied through time and swings according to several internal and external factors. The period which the study focuses upon (19902002) has been associated with major fluctuations in SET’s index. Indeed the index rose to the highest point at 1682.9 in 1993 and fell to 269.19, its lowest level in its history, in 2000. The background and performance of SET is a key factor which 217 DBA Thesis: Malinee Ronapat explains why the payment of dividends disappeared and reappeared in Thailand during the period of investigation. Research question # 2 is asked: How have researchers investigated dividends and their patterns throughout the world? This research question was answered in chapter 3. The nature of dividend policies and theories were discussed in chapter 3 together with the relationships between dividends and investment, firm characteristics, and the propensity to pay dividends of firms. Fama and French (2001) raised awareness of the issues in the world’s capital markets with their study on disappearing dividends and its relation to the characteristics of firms and the propensity to pay dividends. A number of researchers and analysts around the world have shown interest in this issue and have conducted studies aiming to support, or reject the suggestions of Fama and French. Research question # 3 stated that: What is the appropriate methodology for collecting and analysing data? Research question # 3 is answered in chapter 4. The data was obtained from the Stock Exchange of Thailand and was divided into three periods of time (1) before the crisis (1991-1996), (2) after the crisis (1997-2002) and (3) the whole period of investigation (1991-2002). Descriptive statistics have been used by researchers and analysts to initially understand the phenomenon. ANOVA was also used to confirm whether the means of a variety of dividend groups were statistically different to each another. Logit regression analysis was also used to obtain a formula to predict the percentage of payers in the capital market for each year between 1990 and 2002. This method of regression is commonly used in the area of finance when the dependent variable has two possible outcomes, occurrence and non-occurrence (payers and non-payers). 218 DBA Thesis: Malinee Ronapat Research question # 4 asked: Does the data from Thailand support a model? Research questions # 4.1-4.3 asked: What are the characteristics of the dividend payers? Do these characteristics change over time? How has the propensity to pay dividends of firms changed over time? The above research questions were answered in chapter 5. The data collected from the Stock Exchange of Thailand was used to calculate the values for six ratios which were used to represent three characteristics of firms. Et/At is the ratio of aggregate earnings before interest to aggregate assets (profitability), Yt/BEt is the ratio of aggregate common stock earnings over aggregate book equity (profitability), dAt/At is the growth rate in assets (investment opportunities), Vt/At is the aggregate market value to the aggregate book value of assets (investment opportunities), At is the total asset (size), Lt/At is the ratio of total liabilities to total asset (size). When applying logit regression to the data, the formula predicted the percentage of dividend payers more accurately for the period before the crisis, while it was less accurate after the crisis. Logit regression suggested there were changes in the characteristics of the payers and the propensity to pay dividends of these payers given the characteristics which were calculated from the predicted value. The changes in characteristics were minimal before and after the crisis but very large during the crisis in 1997. The characteristics of firms which were observed during the base years (19911996) appeared to re-occur after the crisis. An important factor was that the propensity 219 DBA Thesis: Malinee Ronapat to pay dividends of firms is lower after the crisis than before the crisis. This finding is consistent with the literature. Research question # 5 stated the following: How could a model to predict payers be implemented? The answer to this research question is discussed in section 6.3. The model to predict the percentage of payers given their characteristics, could be implemented in a capital market such as SET by regulators, analysts and investors. 6.2.1.2 CONCLUSIONS RELATED TO TESTING THE MODEL This section is a conclusion on the relationship between firm characteristics and the probability that firms will pay dividends to predict the dividend behaviour of listed firms. Research question # 4, as stated earlier, has several sub-questions (research questions #4.1-4.3) that lead to the hypotheses which were developed. To answer research question 4, and its subquestions, these hypotheses were tested and analysed. These hypotheses were answered in chapter 5 by using logit regression in step, A (1991 to 1996), B (1997 to 2002) and C (1991-2002). Every dividend group was analysed twice, firstly with one investment variable and secondly with two investment variables. The same methodology was applied by Fama and French (2001). These researchers used logit regression with average coefficient and group coefficient techniques. Three firm characteristics were analysed (1) profitability, (2) investment opportunities and (3) size. Variable Et/At (the ratio of aggregate earnings before interest to aggregate assets for profitability) as a proxy for profitability, dAt/At (the growth rate of assets for investment opportunities) as a proxy for investment opportunities, Vt/At (the ratio of aggregate market value to the aggregate book value of assets for investment opportunities) as a proxy for investment opportunities and SET (the percent of sample 220 DBA Thesis: Malinee Ronapat firms with lower or equal market capitalisation rate for size) as a proxy for size. All of the variables were statistically significant for predicting payers, when applying logit regression in every period (controlling two then one investment variables). However, the results were mixed when different steps and methods are applied (see Chapter 5). Research question # 4.1.1 stated the following: Does the dividend payment behaviour depend on the profitability of the listed firms? Research question # 4.1.1 was tested with the research hypothesis # 1 which is: Ho : There is no significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends. Ha : There is a significant relationship between the profitability of the firms and the probability that the listed firms will pay dividends. Table 6.1 and 6.2 present the answers to research problem #4.1.1 in three different period of times. Table 6.1: Results of Hypotheses Testing Following Fama and French (2001) Method of Averaging Coefficient (Profitability) Periods Before the Crisis (Step A) 1991-1996 Controlled Variables Characteristics 2 Variables Profitability 1 Variable During and After the Crisis (Group B) 1997-2002 2 Variables 2 Variables Rejected Yes Rejected Yes Rejected Yes Profitability Rejected Yes Rejected Yes Profitability Rejected Yea Profitability 1 Variable Answer to Question # 4.1.1 Profitability Profitability 1 Variable The Whole Period of Investigation (Group C) 1991-2002 Null Hypothesis Source: Developed for this research Note: ‘Yes’ means that dividend payment behaviour depends on the profitability of the listed firms 221 DBA Thesis: Malinee Ronapat Table 6.2: Results of Hypotheses Testing Using Group Coefficient (Profitability) Periods Before the Crisis (Step A) 1991-1996 Controlled Variables Characteristics 2 Variables Profitability 1 Variable During and After the Crisis (Stpe B) 1997-2002 2 Variables 2 Variables Rejected Yes Rejected Yes Rejected Yes Profitability Accepted No Rejected Yes Profitability Rejected Yes Profitability 1 Variable Answer to Question #4.1.1 Profitability Profitability 1 Variable The Whole Period of Investigation (Step C) 1991-2002 Null Hypothesis Source: Developed for this research Note: ‘Yes’ means that dividend payment behaviour depends on the profitability of the listed firms The analysis found that profitability (Et/At: the ratio of aggregate earnings before interest to aggregate assets) is a significant factor for predicting the dividend behaviour in all time periods, with two, or one controlled investment variable. The finding was consistent with Fama and French (2001). When applying the group coefficient technique, Et/At was not a significant factor for predicting the dividend behaviour from 1997 to 2002 controlling one investment variable. Research question # 4.1.2 stated the following: Does the dividend payment behaviour depend on the investment opportunities of the listed firms? Research hypothesis # 2 was generated from research question 4.1.2: Ho : There is no significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends. Ha : There is a significant relationship between the investment opportunities of the firms and the probability that the listed firms will pay dividends. Two measurements were used for controlling investment opportunities: dAt/At (the growth rate of assets for investment opportunities) and Vt/At (the ratio of aggregate 222 DBA Thesis: Malinee Ronapat market value to the aggregate book value of assets). Tables 6.3 and 6.4 present the answers to the null hypothesis and research question 4.1.2. Table 6.3: Results of Hypotheses Testing Following Fama and French (2001) Method of Averaging Coefficient (Investment Opportunities) Periods Before the Crisis (Step A) 1991-1996 Controlled variables 2 Variables Null Hypothesis Answers to question #4.1.2 Investment (dAt/At) Rejected Yes Investment (Vt/At) Accepted No Accepted No Investment (dAt/At) Rejected Yes Investment (Vt/At) Accepted No Accepted No Investment (dAt/At) Rejected Yes Investment (Vt/At) Accepted No Accepted No Characteristics Investment (dAt/At) 1 Variable During and After the Crisis (Step B) 1997-2002 2 Variables Investment (dAt/At) 1 Variable The Whole Period of Investigation (Step C) 1991-2002 2 Variables Investment (dAt/At) 1 Variable Source: Developed for this research Note: ‘Yes’ means that dividend payment behaviour depends on the investment opportunities of the listed firms. Table 6.4: Results of Hypotheses Testing Using Group Coefficient Technique (Investment Opportunities) Periods Before the Crisis (Step A) 1991-1996 Controlled Variables 2 Variables Characteristics Null Hypothesis Answers to Question #4.1.2 Investment (dAt/At) Rejected Yes Investment (Vt/At) Rejected Yes Rejected Yes 1 Variable During and After the Crisis (Step B) 1997-2002 2 Variables Investment Investment (dAt/At) Accepted No Investment (Vt/At) Rejected Yes Rejected Yes Investment (dAt/At) Rejected Yes Investment (Vt/At) Accepted No Rejected Yes 1 Variable The Whole Period of Investigation (Step C) 1991-2002 2 Variables 1 Variable Investment Investment Source: Developed for this research Note: ‘Yes’ means that dividend payment behaviour depends on the investment opportunities of the listed firms. 223 DBA Thesis: Malinee Ronapat The results suggest that by controlling two investment variables, dAt/At is a significant variable for predicting the dividend behaviour of firms in every period. However, Vt/At. was found to be insignificant. When controlling one investment variable, the results were mixed for Vt/At, especially when using the average coefficient technique. It appears that Vt/At was not related to the dividend behaviour of firms in any time period which was analysed. However, the analysis with logit regression which used a group coefficient suggested that there is a relationship between the dividend behaviour of firms and dAt/At (growth rate in assets) in every time period. Table 6.5 presents the answer to research question # 4.1.3. Research question # 4.1.3 stated that: Does the dividend payment behaviour depend on the size of the listed firms? Research hypothesis # 3 was developed to answer research question #4.1.3: Ho : There is no significant relationship between the size of the firms and the probability that the listed firms will pay dividends. Ha : There is a significant relationship between the size of the firms and the probability that the listed firms will pay dividends. Size is measured in term of SET or the percentage of firms with identical or less market capitalisation. Size was found to be a significant variable for predicting the dividend behaviour of firms in most time periods. Table 6.5 and 6.6 summarise the answers to research questions number 4.1.3. 224 DBA Thesis: Malinee Ronapat Table 6.5: Results of hypotheses testing following Fama and French (2001) method of averaging coefficient (size) Periods Before the Crisis (Step A) 1991-1996 Controlled Variables 2 Variables Characteristics Null Hypothesis Answers to Question #4.1.3 Size Rejected Yes Accepted No Rejected Yes Rejected Yes Rejected Yes Rejected Yes 1 Variable During and After the Crisis (Step B) 1997-2002 2 Variables Size Size 1 Variable The Whole Period of Investigation (Step C) 1991-2002 2 Variables Size Size 1 Variable Size Source: Developed for this research Note: ‘Yes’ means that dividend payment behaviour depends on the size of the listed firms. Table 6.6: Results of hypotheses testing using group coefficient method (size) Periods Before the Crisis (Step A) 1991-1996 Controlled Variables 2 Variables Characteristics Null Hypothesis Answer to Question #4.1.3 Size Rejected Yes Accepted No Rejected Yes Rejected Yes Rejected Yes Rejected Yes 1 Variable During and After the Crisis (Step B) 1997-2002 2 Variables Size Size 1 Variable The Whole Period of Investigation (Step C) 1991-2002 2 Variables Size Size 1 Variable Size Source: Developed for this research Note: ‘Yes’ means that dividend payment behaviour depends on the size of the listed firms. The logit regression, which applied an average coefficient and the regression with a group coefficient both suggested that the dividend behaviour of firms depends on firm size (as measured by the percent of firms with same or lesser market capitalisation) in every time period, with the exception of step A (before the crisis) controlling one investment variable and the null hypothesis is accepted. Although the results are mixed, every variable in the step that controlled two investment variables were significant for predicting the dividend behaviour of listed 225 DBA Thesis: Malinee Ronapat firms in every time period. However, the tests for goodness of fit suggested that every model which was developed from the samples is significant for predicting the percentage of payers in the market (chapter 5). The logit regression model which was developed identified the following x There was slight change (or decline) in the characteristics of the dividend payers even before the Asian Economic Crisis in 1997. x The characteristics which were present during the base years changed significantly (during the crisis) in 1997 but resumed during the period 19982002. x The characteristics of the dividend payers during the base years (1991 to 1996) and whole period of investigation (1991 to 2002), were slightly different. x The predicted percentage of payers is, in almost every year, higher than the actual percentage of payers in the market. x The decline in the propensity to pay dividends of the payers is evidenced even before the crisis. x The divergence between the actual percentage of payers are the predicted percentage of payers increased for every year after 1997 (the actual percentage was always lower than the expected percentage). x Declining in the propensity to pay dividend of payers was large after the crisis. x The divergence between the actual and the expected percentage of payers has tended to decline in recent years and indicates that the market is recovering. x The empirical findings of this study are consistent with the literature. x The model could be used for predicting the percentage of dividend payers in the market. The reason why listed firms in Thailand continue to pay dividends or resume paying dividends after a major economic downturn is still a puzzle. Several researchers have studied this issue although there is a consensus as to why some firms pay dividends and some do not and why some investors prefer dividends and some do not. Trends such as the disappearance of dividends in the short-term (before crisis), the long-term (19902002) and the reappearance of dividend payments of firms (1998 till current) could be 226 DBA Thesis: Malinee Ronapat explained by the general advantages and disadvantages of dividends as discussed below. Lintner (1956) observed that managers are reluctant to omit dividends. If firms are reluctant to omit dividends, especially, during bad times, the benefit of dividends increase, because investors do not need to sell stocks to receive returns from their investment. Therefore, reliability in the payment of dividends will make investors feel that their assets are secure, while capital gains are likely to fluctuate. Dividends could therefore be viewed as a cushion during bad times, or when stock prices fall (Linter 1956, Gwilym, Morgan and Thomas 2000). Dividends are useful when the market has many high-tech stocks which do not pay dividends but experience high growth rates. Dividends also provide investors with some control over the management of firms as they can demand that dividends should be paid. As stated earlier, cash dividends are real money and can provide investors with an indication of a firm’s performance. Fama and French (2001) suggested that firms with positive earnings are more likely to pay dividends, while young firms are not. Therefore, the payment of dividends does not disclose a firm’s poor performance in the last accounting period, although it appears sometimes that earnings do. Researchers have also suggested that dividend payments provide information to investors on the current and future performance of the firm, external financing, as well as the true value of firms. However, forgoing dividends should actually provide investors with higher returns in the long-run when compared with other types of investment (Bhattacharya 1979, 1980; Asquith and Mullins 1983; John and Williams 1985; Miller and Rock 1985; Richardson, Sefcik and Thompson 1986; Healey and Palepu 1988). In Thailand, the SEC (2002b) has surveyed investors understanding of investment, their behaviours and the information they need to undertake informed decisions. SEC found that 26 percent of the sample investors prefer and are interested in firms which pay dividends, while only 16 percent prefer firms that provide large capital gains. The 227 DBA Thesis: Malinee Ronapat research also stated that, when making investment decisions, Thai investors perceive that dividend history and financial statements are more important than the investment opportunities of the firms (SEC 2002b). Knowing that dividend payment received is taxed at the rate of 5.71 percent in Thailand but capital gain is tax-free (see section 5.4 in Chapter 5), Thai investors seem to be fond of dividend payments. This investment culture of Thai investors could be a result of reappearing dividend pattern after the Asian Economic Crisis in 1997. Whether this culture is a country factor and how it affects the Thailand’s capital market, is an issue for future research. However, deciding whether to invest in dividend-paying stocks or mutual funds, many factors such as taxation, risk and the need for income will be considered. Therefore, the phenomenon of disappearing dividends could be related to the above issues. Several explanations for the disappearance of dividends could include the following. In many countries, including Thailand, capital gains are usually taxed less than dividends. Investors can decide to pay tax on dividends or sell their shares and pay tax on capital gains. From the perspective of the firm, managers can be more flexible when they report their earnings if they do not pay dividends. As dividend payment is a reduction of retained earnings, managers of non-payer firms are more likely to have higher retained earnings in their balance sheets. Furthermore, issuing stocks are costly, while borrowing is less expensive as the interest from borrowing is tax deductible (Banker and Wulgler 2003b). The declining propensity to pay dividends of the listed firms could result from six factors according to Baker and Wugler (2003b). 1. Agency: Dividends are viewed as a tool for good corporate governance (Easterbrook 1984) to keep managers close to capital markets. Dividends could be disappearing due to a new corporate governance mechanism which help reduce the agency problems between manager and investors. Therefore the need to control agency costs through dividends is diminish (Fama and French 2001, Banker and Wulgler 2002, 2003a, 2003b). 228 DBA Thesis: Malinee Ronapat 2. Information Asymmetry: Dividends are a signal of the profitability of the firm, and could be used as a tool for predicting the performance of the firm and its stocks (Ross 1977; Battacharya 1979; John and Williams 1985; Miller and Rock 1985). However, as information transmission rates are relatively fast and accurate nowadays, the gap between managers and investors is narrowed and prediction models (such as Dividend Discount Model) may be used less often (Baker and Wulgler 2002, 2003a, 2003b). The role of dividends in transmitting information is reduced. 3. Stock Options: Managers could hold options which are not dividend protected and this could create personal incentives for not paying dividends. Further research has suggested that options correlate with lower dividend payments (Lambert, Lanen and Larcker 1989; Jolls 1998; Weisbenner 1999; Fama and French 2001; Baker and Wulgler 2002, 2003a, 2003b). 4. Clientele Equilibrium: Baker and Wurgler (2003b) explained that the decline in the payment of dividends could result from changes in the demand for dividends from the firms by investors. 5. Catering: Baker and Wurgler (2003b) suggested a new theory on the disappearance of dividends. They found that dividend payments depend on the demand from investors. Therefore, managers will pay dividends if their share prices are perceived to be at a premium and omit dividends if their share prices are discounted (Baker and Wurgler 2002). 6. Slow Learning about Taxes: Baker and Wurgler (2003b) stated that managers may have been slow to understand the tax disadvantages of paying dividends. After fully understanding these disadvantages, their propensity to pay dividend declined (Grinblatt and Titman 1998). Several of these causes and effects could be used to explain the pattern of disappearing and reappearing dividends on the Thai stock market of Thailand. However, this study concentrates on the characteristics of firms, changes in these characteristics and the propensity to pay dividends of listed firms over time. 229 DBA Thesis: Malinee Ronapat CONCLUSIONS – THE RESEARCH PROBLEM Research Problem The research problem is restated as follows: ‘How do the characteristics of publicly traded firms and their propensity to pay dividends influence the payment of dividends and reason for their disappearance from the stock market of Thailand? The payment of dividends disappeared during the crisis and re-appeared after the Asian Economic Crisis. In addition, changes in the characteristics of payers is evident before and after the crisis. It is more important, however, that listed firms have shown a lower propensity to pay dividends before and after the crisis. The percentage of dividend payers in the market declined irrespective of changes in the characteristic of firms. Figure 6.2 shows the relationships between the findings of the logit regression, ANOVA and the descriptive statistical techniques. The figure below shows that firms started to have lower propensity to pay dividends and changed characteristics before the crisis and the Asian Economic Crisis triggered firms to have even lower propensity to pay dividends. These results in the decline in percentage of payers continue paying dividends, non-payers to start paying dividends. The percentage of never paid firms increased, as well as the new lists that tend not paying dividends. All of these result in the decline in the percentage of dividend payers in the market. 230 DBA Thesis: Malinee Ronapat Figure 6.2: Relationships of Logit Regression, ANOVA and Descriptive Statistics Findings FIRMS HAVE LOWER PROPENSITY TO PAY DIVIDENDS % PAYERS CONTINUE TO PAY DECLINED RESULTS IN FIRMS CHANGED CHARACTERISTICS ASIAN ECONOMIC CRISIS TRIGGERED % NON-PAYER START TO PAY DECLINED RESULTS IN % PAYER DECLINED % NEVER PAID FIRMS INCREASED LOWER PROPENSITY CHANGES TO CHARACTERISTICS % NEW LISTS THAT DO NOT PAY HAS INCREASED Source: Developed for this research 6.3 IMPLICATIONS This section discusses the implications of the findings of this study for theory, policy and investment practice. The contribution of this study to the body of knowledge is also discussed. 6.3.1 IMPLICATIONS FOR THEORY As this research is concerned with the disappearance of dividends from the stock market of Thailand, it will be compared with the existing literature and will provide additional support for several of the findings of the Fama and French (2001) study. It 231 DBA Thesis: Malinee Ronapat will also be relevant to scholars who are interested in the pattern of dividend payment in developed and developing capital markets. In particular, the findings of this study provide insight for emerging stock markets such as Thailand. The findings of this research have implications for countries with similar market conditions, especially in developing economies such as those of Asian countries. Therefore, this study will help direct further research. It could also be used as a teaching tool and add to the knowledge on finance and investment in the business schools at both undergraduate and postgraduate levels. The methodology, sample size, analysis and implications could be used for discussion in classes and students could use this thesis as a guideline for developing their own research papers. 6.3.2 IMPLICATIONS FOR POLICY Since this research concentrates on the dividend payment of listed firms in Thailand, its findings have implications for the Securities Exchange Commission of Thailand (SEC) and the Stock Exchange of Thailand (SET). Several of these possible implications are discussed below. The research also provides a tool to supervise and monitor the characteristics of listed firms and their dividend behaviour. It will help maintain the safety and soundness of the financial system and capital market in Thailand. The findings of this research could be used as a preliminary investigation of the changes in the characteristics of listed firms. It could help predict the expected number of payers in the market and monitor the performance of new and existing listed firms. 232 DBA Thesis: Malinee Ronapat Listed firms could be grouped into three dividend groups as specified by Fama and French (2001) (payers/non-payers, payer/former payers/never paid) according to their dividend behaviour. SEC and SET could provide this information to the public to enhance the quality of information for investors. This will help increase the level of understanding of investments in different types of securities with various risk-return trade-offs. The SEC and SET should issue CDs and annual reports to enhance the flow of information to investors. Therefore, investors will be able to make decisions using accurate information. The SEC and SET should impose a strict regulation, or a code of conduct, stating that listed firms must provide at least 5 to 10 years of historical financial statements and a copy of the company’s annual report to investors free of charge. This study could provide a tool for the early detection of an increase, or decrease in the number of payers and non-payers in the market, an upward or downward trend in the payment of dividends and changes in the propensity to pay dividend in the next period. This research could be an ideal report for investors on the pattern of dividend payments and for the SEC or SET. Furthermore, the analysis of listed firms characteristics is based on information from the financial statements of the listed firms. It may therefore increase public awareness of the information which is currently provided by SET and SEC. This will help enhance the protection and education of investors. 6.3.3 IMPLICATION FOR INVESTMENT PRACTICE The information provided in the research could also be used as a guide on investment practice for the different types of stock being traded in the market (e.g. value and growth stocks, payers and non-payers). It helps provide information on the dividend behaviour of listed firms with the given characteristics and provides a warning system for the firms that may or may change their dividends behaviour. This study also highlights changes in firm characteristics and the propensity to pay dividends. It can be 233 DBA Thesis: Malinee Ronapat used as a tool for investors in the stock selection process by providing insight on the expected returns. Furthermore, the research could also aid investors decisions by suggesting stocks for a hedging strategy or helping to diversify their portfolios. Investors could shift their funds to different dividend groups (from payers to nonpayers or to former-payers or firms that have never paid dividend) to enhance the performance of securities in their diversified portfolio and magnify investors’ tax preferences. This would aid technical and fundamental analysis in the stock market, thus, increasing efficiency. Importantly, this research could increase confidence and encourage investors to invest in the capital market in Thailand. 6.4 CONTRIBUTIONS OF THE STUDY This research has contributed to the level of knowledge on the phenomenon of disappearing dividends in the following ways. The logit regression model which was developed to predict the percentage of payers in the market in a particular year, is important and a significant indicator of changes in firm characteristics and the propensity to pay dividends of the listed firms. The descriptive statistical findings of this research also contribute to the literature on investment and financial management. The year-to-year analysis of the data provides a clearer picture of the characteristics, the propensity to pay dividends and the pattern of dividend payments than a methodology which only focuses on a range of years. A range of years may contain mixed and misleading signals of the phenomenon. The findings of this study contribute to the level of knowledge of the phenomenon of disappearing dividends and the dividend payments pattern in an emerging capital market such as Thailand. Minimal research has focused on the disappearance of dividends because the issue is relatively new. Furthermore, most of the relevant studies 234 DBA Thesis: Malinee Ronapat have obtained data from developed stock markets such as the US and UK. This research expands and contributes to the literature by concentrating on Thailand’s stock market. The findings of this research support the view that emerging markets (Thailand) often follow the same pattern as developed markets (US and UK). The results of this research indicate that the phenomenon of disappearing dividends is present in Thailand and it is related to firm characteristics, changes in these characteristics and the propensity to pay dividends. Furthermore, it is interesting to note that the propensity to pay has also declined in Thailand. The review of literature noted that it has also occurred in the US although it has not been empirically explained. The research also helps confirm the market efficiency and the implications for technical and fundamental analysis in an emerging capital market. Market participants will be able to use the findings of this study to more accurately predict expected cash inflows and the value of stocks. With this information, market participants will be able to make more informed investment decisions. 6.5 LIMITATIONS OF THE STUDY Doctoral research is constrained by resources, time, funding, scope and focus. Consequently, this research and its findings maybe influenced by the following limitations. The period of investigation covers the period 1990 to 2002. However, the data was obtained from the SET database was only complete from 1997 onwards. The data search was conducted to the highest possible level, however, the data from the 19901996 period is incomplete. Consequently, there could be some sample bias. 235 DBA Thesis: Malinee Ronapat The characteristics of the listed firms which were suggested by Fama and French (2001) may be limited and may overlook other characteristics which may be important. Each ratio which represents a proxy for a firm’s characteristics has its own limitations in terms of expressing the relationship between the characteristics of the firm and its dividend behaviour. This problem occurs even when firm characteristics are being analysed in developed markets. This study has analysed an emerging market, and some of the homogenous characteristics may not be presented in the study. Lastly, listed firms may have their own accounting policies, and these accounts may represent different characteristics of firms. 6.6 SUGGESTIONS FOR FURTHER RESEARCH It is suggested that more research is needed on the characteristics of the different dividend groups. In particular, the issue of whether the characteristics which were suggested by Fama and French (2001) are adequate needs further consideration. Furthermore, expanding the size of the sample of listed firms and the length of time frame should be considered because the findings of this study indicate an upward trend in the percentage of payers in recent years in the Thailand’s capital market. In addition, more than three dividend groups could be assessed in further research. Different patterns of dividends payments may also be identified if the time frame and size of samples are expands. Factors such as inflation rates, interest rates, business cycles, GDP, GNP could be incorporated in the analysis to enhance the predictive power of the logit regression. The fact that many firms resumed paying dividends after the crisis in Thailand indicates that investors appreciate value stocks and firms still want to pay dividends. This implies further research is needed to investigate the performance of value and growth stocks in Thailand. This is to provide additional information for investors and expand the 236 DBA Thesis: Malinee Ronapat literature of the dividend behaviour of firms. Behaviourial finance could also be used to explain the behaviour of the Thai investors, why they prefer value stocks or growth stocks, and is it culturally different. Finally, though descriptive statistics and logit regression are recommended as an appropriate methodology in the literature (Fama and French 2001), researchers could expand the research effort by applying different statistical tools to verify the characteristics which were suggested by Fama and French (2001), or analyse the propensity to pay dividends of listed firms over time. 237 DBA Thesis: Malinee Ronapat References ADB 2000a, Asian Recovery Report 2000a, Asia Recovery Information Center 2000a, Asian Development Bank, Manila. ADB 2000b, Asian Recovery Report 2000b, Asia Recovery Information Center 2000b, Asian Development Bank, Manila. ADB 2000c, Asian Recovery Report 2000c, Asia Recovery Information Center 2000c, Asian Development Bank, Manila. Aivazian, V., Booth, L. & Cleary, S. 2003, ‘Do Emerging Market Firms Follow Different Dividend Policies From U.S. Firms?’, Journal of Financial Research, vol. 26, no. 3, Fall, pp. 371-387. Ambarish, R., John, K. & Williams, J. 1987, ‘Efficient Signaling with Dividends and Investments’, Journal of Finance, vol. 42, no. 2, June, pp. 321-343. Asquith, P. & Mullins, D.W. Jr. 1983, ‘The Impact of Initiating Dividend Payments on Shareholders' Wealth’, Journal of Business, vol. 56, no. 1, February, pp. 77-96. Auerbach, A.J., 1979a, ‘Share Valuation and Corporate Equity Policy’, Journal of Public Economics, vol. 2, no. 2, June, pp. 291-305. Auerbach, A.J., 1979b, ‘Wealth Maximization and the Cost of Capital’, Quarterly Journal of Economics, vol. 93, no. 3, August, pp. 433-446. Baker, H.K. & Farrelly, G.E. 1988, ‘Dividend Achievers: A Behavioral Perspective’, Akron Business and Economic Review, vol. 19, no. 1, Spring, pp. 79-92. Baker, H.K., Farrelly, G.E. & Edelman, R.B. 1985, ‘A Survey of Management Views on Dividend Policy’, Financial Management, vol. 14, no. 3, Autumn, pp. 78-84. Baker, M. & Wurgler, J. 2002, ‘Why are Dividends Disappearing’, Unpublished Working Paper, Stern School of Business, New York University, December. Baker, M. & Wurgler, J. 2003a, ‘A Catering Theory of Dividends’, National Bureau of Economic Research Ind., NBER Working Papers, no. 9542, March, http://pages.stern.nyu.edu/~jwurgler/papers/catering.pdf (accessed 14th, October 2003). Baker, M. & Wurgler, J. 2003b, ‘Appearing and Disappearing Dividends: The Link to Catering Incentives’, National Bureau of Economic Research Ind., NBER Working Paper, no. 9995, September, http://pages.stern.nyu.edu/~jwurgler/papers/appdisdiv.pdf (accessed 14th October 2003). 238 DBA Thesis: Malinee Ronapat Bar-Yosef, S. & Huffman, L. 1986, ‘The Information Content of Dividends: A Signalling Approach’, Journal of Financial and Quantitative Analysis, vol. 21, no. 1, March, pp. 4758. Bar-Yosef, S. & Kolodny, R. 1976, ‘Dividend Policy and Capital Market Theory’, Review of Economics and Statistics, vol. 58, no. 2, May, pp. 181-190. Bar-Yosef, S. & Kolodny, R. 1978, ‘Dividend Policy and Capital Market Theory: A Reply’, Review of Economics and Statistics, vol. 60, no. 3, August, pp. 477-478. Benito, A. & Young, G. 2001, ‘Hard Times or Great Expectations?: Dividend Omissions and Dividend Cuts by UK Firms’, Working Papers no. 147, Bank of England, London. Berle, A.A. & Means, G.C. 1932, The Modern Corporation and Private Property, MacMillan Company, New York. Bhattacharya, S. 1979, ‘Imperfect Information, Dividend Policy, and "The Bird in the Hand" Fallacy’, Bell Journal of Economics, vol. 10, no. 1, Spring, pp. 259-270. Bhattacharya, S. 1980, ‘Nondissipative Signaling Structures and Dividend Policy’, Quarterly Journal of Economics, vol. 95, no. 1, August, pp. 1-24. Black, F. 1976, ‘The Dividend Puzzle’, Journal of Portfolio Management, vol. 2, no. 2, Winter, pp. 5-8. Black, R.T. 1999, Doing Quantitative Research in the Social Sciences, An Integrated Approach to Research Design, Measurement and Statistics, SAGE Publication, London. BOI 2004, ‘Country Review’, http://www.boi.go.th/english/thailand/index.html, Accessed 16th February 2004. BOT 1990, Economic Review (1990), Publication Department, Bank of Thailand, Bangkok. BOT 1991, Economic Review (1991), Publication Department, Bank of Thailand, Bangkok. BOT 1992, Economic Review (1992), Publication Department, Bank of Thailand, Bangkok. BOT 1993, Economic Review (1993), Publication Department, Bank of Thailand, Bangkok. BOT 1994, Economic Review (1994), Publication Department, Bank of Thailand, Bangkok. BOT 1995, Economic Review (1995), Publication Department, Bank of Thailand, Bangkok. 239 DBA Thesis: Malinee Ronapat BOT 1996, Economic Review (1996), Publication Department, Bank of Thailand, Bangkok. BOT 1997, Economic Review (1997), Publication Department, Bank of Thailand, Bangkok. BOT 1998, Economic Review (1998), Publication Department, Bank of Thailand, Bangkok. BOT 1999, Economic Review (1999), Publication Department, Bank of Thailand, Bangkok. BOT 2000, Economic Review (2000), Publication Department, Bank of Thailand, Bangkok. BOT 2001, Economic Review (2001), Publication Department, Bank of Thailand, Bangkok. BOT 2002, Economic Review (2002), Publication Department, Bank of Thailand, Bangkok. Brealey, R.A. & Myers, S.C. 1991, Principles of Corporate Finance, 4th edition, McGraw Hill, New York. Brealey, R.A. & Myers, S.C. 2000, Principles of Corporate Finance, 6th edition, Irwin McGraw-Hill, London. Brennan, M.J. 1970, ‘Taxes, Market Valuation, and Corporation Financial Policy’, National Tax Journal, vol. 23, no. 3, December, pp. 417-427. Bringham, E.F., Gapenski, L.C. & Ehrhardt, M.C. 1999, Financial Management, Theory and Practice, 9th edition, Dryden Press, Fort Worth, Texas. Bryman, A. & Cramer, D. 1999, Quantitative Data Analysis with SPSS Release 8 for Windows: A Guide for Social Scientists, Routledge, London. Bryman, A. & Cramer, D. 2001, Quantitative Data Analysis with SPSS Release 10 for Windows: A Guide for Social Scientists, Routledge, London. Case, K.E. & Fair, R.C. 1996, Principles of Economics, 4th edition, Prentice-Hall International. Cavana, R.Y., Delahaye, B.L. & Sekeran, U. 2001, Applied Business Research, Qualitative and Quantitative Methods, John Wiley and Sons, Sydney. Cholviroj, B. 1989, Taxation, Tana Publishing, Bangkok. 240 DBA Thesis: Malinee Ronapat Christensen, T. & Jordan, W. 2001, Fundamentals of Corporate Finance, 2nd edition, McGraw-Hill, Roseville, New South Wales. Chua, S.J., Sim, W.Y. & Tham, S.S. 1999, ‘The Asian Financial Crisis – A Critical Study’, Applied Research Project, Nanyang Technological University, pp. 36-47. Clover, V.T. & Balsley, H.L. 1984, Business Research Methods, 3rd edition, Wiley & Sons, New York. Darling, P.G. 1957, ‘The Influence of Expectations and Liquidity on Dividend Policy’, Journal of Political Economy, vol. 65, no. 2, June, pp. 209-224. DeAngelo, H. & DeAngelo, L. 1990, ‘Dividend Policy and Financial Distress: An Empirical Examination of Troubled NYSE Firms’, Journal of Finance, vol. 45, no. 5, December, pp. 1415-1431. DeAngelo, H., DeAngelo, L. & Skinner, D.J. 1992, ‘Dividends and Losses’, Journal of Finance, vol. 57, no. 5, December, pp. 1837-1863. DeAngelo, H., DeAngelo, L. & Skinner, D.J. 2002, ‘Are Dividends Disappearing? Dividend Concentration and the Consolidation of Earnings’, Centre in Law, Economics and Organization Research Papers, Working Paper no.02-09, University of Southern California. DeAngelo, H. & Masulis, R.W. 1980. ‘Leverage and Dividend Irrelevancy Under Corporate and Personal Taxation’, Journal of Finance, vol. 8, no. 1, March, pp. 3-29. Easterbrook, F.H. 1984, ‘Two Agency-Cost Explanations of Dividends, The American Economic Review, vol. 74, no. 4, September, pp. 650-659. Emory, C.W. 1995, Business Research Methods, 5th edition, Irwin, Homewood, Illinois. Emory, C.W. & Cooper, D.R. 1991, Business Research Methods, 4th edition, Irwin, Homewood, Illinois. Fama, E.F. 1970, ‘Efficient Capital Markets: A Review of Theory and Empirical Work’, Journal of Finance, vol. 25, no. 2, March, pp. 383-423. Fama, E.F. 1976, Foundations of Finance: Portfolio Decisions and Securities Prices, Basic Books, New York. Fama, E.F. 1998, ‘Market Efficiency, Long-Term Returns, and Behavioral Finance’, Journal of Financial Economics, vol. 49, no. 3, September, pp. 283-306. Fama, E.F. & Babiak, H. 1968, ‘Dividend Policy: An Empirical Analysis,’ Journal of the American Statistical Association, vol. 63, no. 324, December, pp. 1132-1161. Fama, E.F. & French, K.R. 1988, ‘Dividend Yields and Expected Stock Returns’, Journal of Financial Economics, vol. 22, no. 1, October, pp. 3-25. 241 DBA Thesis: Malinee Ronapat Fama, E.F. & French, K.R. 1995, ‘Size and Book-to-Market Factors in Earnings and Returns’, Journal of Finance, vol. 50, no. 1, March, pp. 131-155. Fama, E.F. & French, K.R. 1997, ‘Dividends, Debt, Investment and Earnings’, Working Paper, Graduate School of Management, University of Chicago, Chicago. Fama, E.F. & French, K.R. 1998, ‘Value versus Growth: The International Evidence’, Journal of Finance, vol. 53, no. 6, December, pp. 1975–1999. Fama, E.F. & French, K.R. 1999, ‘Testing Tradeoff and Pecking Order Predictions about Dividends and Debt’, Working Paper, Sloan School of Management, MIT. Fama, E.F. & French, K.R. 2000, ‘Forecasting Profitability and Earnings’, Journal of Business, vol. 72, no. 2, April, pp. 161-175. Fama, E.F. & French, K.R. 2001, ‘Disappearing Dividends: Changing Firms Characteristics or Lower Propensity to Pay?’, Journal of Financial Economics, vol. 60, no. 1, April, pp. 3-44. Fama, E.F & French, K.R. 2002, ‘The Equity Premium’, Journal of Finance, vol. 57, no. 2, April, pp. 637-659. Farrar, D.E. & Selwyn, L.L. 1967, ‘Taxes, Corporate Financial Policy and Return to Investors, National Tax Journal, vol. 20, no. 3, May, pp. 444-462. Fazzari, S., Hubbard, R.G. & Petersen, B.C. 1988, ‘Financing Constraints and Corporate Investment’, Brookings Papers on Economic Activity, pp. 1141-1195. Fama, E.F. & MacBeth, J.D. 1973, ‘Risk, Return and Equilibrium: Empirical Tests’, Journal of Political Economic, vol. 81, no. 3, May, pp. 607-636. Fama, E.F & Miller, M.H. 1972, The Theory of Finance, Dryden Press, Hinsdale, Illinois. Feldstein, M.S. & Green, J. 1983, ‘Why Do Companies Pay Dividends?’, American Economic Review, vol. 73, no. 1, March, pp. 17-30. Ferris, P.S., Sen, N. & Yui, H.P., 2002, ‘International Dividend Policy: Legal Influences, Firm Attributes and the Desire to Pay’, Research Proposal, Nanyang Technological University, Singapore. Firth, M. 1977, The Valuation Of Shares and The Efficient-Markets Theory, 1st edition, Macmillan Press, London. Fisher E.D. & Jordan, J.R. 1991, Security Analysis and Portfolio Management, 5th edition, Englewood Cliffs, New Jersey. Frankfurter, G.M. & Gong, J. 1992, ‘Empirical Tests of a Dividend Signaling Equilibrium Model’, Unpublished Working Paper, Bucknell University. 242 DBA Thesis: Malinee Ronapat Frankfurter, G.M. & Gong, J. 1993, ‘Time-Series Cross-Sectional Tests of Dividend Policy Determinants’, Unpublished Working Paper, Bucknell University. Frankfurter, G.M. & Lane, W.R. 1992, ‘The Rationality of Dividends’, International Review of Financial Analysis, vol. 1, pp. 115-129. Frankfurter, G.M. & Wood, B.G. 2001, ‘Dividend Policy Theories and Their Empirical Tests’, Unpublished Working Paper, Bucknell University. Fung, W.K.H. & Theobald, M.F. 1984, ‘Dividends and Debt Under Alternative Tax Systems’, Journal of Financial and Quantitative Analysis, vol. 19, no. 1, February, pp. 5972. Gitman, L.J. 2000, Principles of Managerial Finance, 9th edition, Addison Wesley Publisher, Melbourne. Gitman L.J., Juchau, R. & Flanagan, J. 2002, Principles of Managerial Finance, 3rd edition, Addison Wesley Publisher, Melbourne. Goldstein, M. 1998, ‘The Asian Financial Crisis: Causes, Crues, and Systematic Implication’, Institute for International Economics, Washington, D.C., June, pp. 1-22. Gordon, M.J. 1962, The Investment, Financing and Valuation of the Corporation, Richard D. Irwin, Homewood, Illinois. Gordon, M.J. 1963, ‘Optimal Investment and Financing Policy’, Journal of Finance, vol.18, no. 2, May, pp. 264-72. Grinblatt, M. & Titman, S. 1998, Financial Markets and Corporate Strategy, Irwin/McGraw-Hill, Boston, Massachusetts. Gwilym, O.A., Morgan, G. & Thomas, S. 2000, ‘Dividend Stability, Dividend Yield and Stock Returns: UK Evidence, Journal of Business Finance and Accounting, vol. 27, no. 3 and 4, April/May, pp. 261-281. Hakansson, N.H. 1982, ‘To Pay or Not to Pay Dividend’, Journal of Finance, vol. 37, May, pp. 415-428; reprinted in summary form in the C.F.A. Digest, 1983; vol. 13, Summer, pp. 16-17. Harkins, E.P. & Walsh, F.J. Jr. 1971, Dividend Policies and Practices, The Conference Board, Inc., New York. Haugen, R. 1995, The New Finance: The Case against Efficient Markets, Prentice Hall, Englewood Cliffs, New Jersey. 243 DBA Thesis: Malinee Ronapat Healey, P.M. & Palepu, K.G. 1988, ‘Earnings Information Conveyed by Dividend Initiations and Omissions’, Journal of Financial Economics, vol. 21, no. 2, September, pp. 149-275. Higgins, R.C. 1972, ‘The Corporate Dividend-Saving Decision’, Journal of Financial and Quantitative Analysis, vol. 7, no. 6, December, pp. 1527-1541. Huck, S.W., Cormier, W.H., & Bounds, W.G. 1974, Reading Statistics and Research, Harper & Row publishers, New Jersey. Hurley, W.J. & Johnson, D.L. 1994, ‘A Realistic Dividend Valuation Model’, Financial Analysts Journals, vol. 50, no. 1, July-August, pp. 50-54. Hurley, W.J. & Johnson, D.L. 1998, ‘Generalized Markov Dividend Discount Models’, Journal of Portfolio Management, vol. 25, no. 1, February, pp. 27-31. James H. L., Dodd, P. & Kimpton, M.H. 1985, The Stock Market: Theories and Evidence, 2nd edition, Irwin Professional Publication, February, Homewood, Illinois. James, L.D., Fama, E.F. & French, K.R. 2000, ‘Characteristics, Covariances and Average Returns 1929 to 1997’, Journal of Finance, vol. 55, no. 1, February, pp. 389-406. Jensen, M.C. 1978, ‘Some Anomalous Evidence Regarding Market Efficiency’, Journal of Financial Economics, vol. 6, no. 2, June/September, pp. 95-101. Jensen, M.C. 1986, ‘Agency Costs of Free Cashflow, Corporate Finance and Takeovers’, American Economic Review, vol. 76, no. 2, May, pp. 323-329. John, K. & Williams, J. 1985, ‘Dividends, Dilution, and Taxes: A Signalling Equilibrium’, Journal of Finance, vol. 40, no. 4, September, pp. 1053-1070. Jolls, C. 1998, The Role of Incentive Compensation in Explaining the Stock-Repurchase Puzzle, Harvard University Working Paper, Harvard University. Kahneman, D. & Tversky, A. 1982, ‘The Psychology of Preferences’, Scientific American, vol. 246, no. 14, December, pp. 167-173. Kale, J.R. & Noe, T.H. 1990, ‘Dividends, Uncertainty, and Underwriting Costs Under Asymmetric Information’, Journal of Finance, vol. 13, no. 2, Winter, pp. 265-277. Kindleberger, C.P. 1984, A Financial History of Western Europe, George Allen & Unwin, London. Kumar, P. 1988, ‘Shareholder-Manager Conflict and the Information Content of Dividends’, Review of Financial Studies, vol. 1, no. 2, June, pp. 111-136. 244 DBA Thesis: Malinee Ronapat Lambert, R.A., Lanen, W.N. & Larcker, D.F. 1989, ‘Executive Stock Option Plans and Corporate Dividend Policy, Journal of Financial and Quantitative Analysis, vol. 24, no. 3, August, pp. 409-425. Laub, P.M. 1976, ‘On the Informational Content of Dividends, Journal of Business, vol. 49, no. 1, February, pp. 73-80. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 2000, ‘Agency problems and Dividend Policies Around the World’, Journal of Finance, vol. 55, no. 2, June, pp. 859874. Lakonishok, J., Shleifer, A. & Vishny, R.W. 1994, ‘Contrarian Investment, Extrapolation, and Risk’, Journal of Finance, vol. 49, no. 5, December, pp. 1541–1578. Lim, C.Y. 2000, ‘From Recession to Recovery in East Asian: a Non-IMF and Non-world Bank Explanation’, Accounting and Business Review, vol. 7, no. 2, July, pp. 1-10. Lintner, J. 1956, ‘Distribution of Incomes of Corporations Among Dividends, Retained Earnings, and Taxes’, American Economic Review, vol. 46, no. 2, May, pp. 97-113. Lintner, J. 1962, ‘Dividends, Earnings, Leverage, Stock Prices, and the Supply of Capital to Corporations’, Review of Economics and Statistics, vol. 44, no. 2, August, pp. 243-69. Madura, J. 2000, International Financial Management, 6th edition, South-Western College Publishing, Cincinnati, Ohio. Makhija, A.K. & Thompson, H.E. 1986, ‘Some Aspects of Equilibrium for a CrossSection of Firms Signaling Profitability with Dividends: A Note’, Journal of Finance, vol. 41, no. 2, May, pp. 249-253. March, T.A. & Merton, R.C. 1987, ‘Dividend Behavior for the Aggregate Stock Market’, Journal of Business, vol. 60, no. 1, January, pp. 1-40. Mason, R.D., Lind, D.A. & Marchal W.G. 1999, Statistical Techniques in Business and Economics, 9th edition, Irwin McGraw-Hill, Chicago. Masulis, R.W. & Trueman, B. 1988, ‘Corporate Investment and Dividend Decisions Under Differential Personal Taxation’, Journal of Financial and Quantitative Analysis, vol. 23, no. 4, September, pp. 369-386. Melicher, R.W. & Norton, E.A. 2000, Finance: Introduction to Institutions, Investments, and Management, 10th edition, South-Western College Publishing, Thomson Learning, Cincinnati, Ohio. Miller, M.H. 1977, ‘Debt and Taxes’, Journal of Finance, vol. 32, no. 2, May, pp. 261275. Miller, M.H. 1986, ‘Behavior Rationality in Finance: The Case of Dividends’, Journal of Business, vol. 59, no. 4, October, pp. 451-468. 245 DBA Thesis: Malinee Ronapat Miller, M.H. & Modigliani, F. 1961, ‘Dividend Policy, Growth, and the Valuation of Shares’, Journal of Business, vol. 34, no. 4, October, pp. 411-413. Miller, M.H. & Rock K, 1985, ‘Dividend Policy Under Asymmetric Information’, Journal of Finance, vol. 40, no. 4, September, pp. 1,031-1,051. Miller, M.H. & Scholes, M.S. 1978, ‘Dividends and Taxes’, Journal of Financial Economics, vol. 6, no. 4, December, pp. 333-364. Miller, M.H. & Scholes, M.S. 1982, ‘Dividends and Taxes: Some Empirical Evidence’, Journal of Political Economy, vol. 90, no. 6, December, pp. 1,118-1,141. Modigliani, F. 1982, ‘Debt, Dividend Policy, Inflation and Market Valuation, Journal of Finance, vol. 37, no. 2, May, pp. 255-273. Moore, D.S. & McLabe, G.P. 1998, Introduction to the Practice of Statistics, 2nd edition, W.H. Freeman, New York. Morgan, I.G. 1980, ‘Dividend and Stock Price Behavior in Canada’, Journal of Business Administration, vol. 12, no. 1, February, pp. 91-107. Morgan, I.G. 1982, ‘Dividends and Capital Asset Prices’, Journal of Finance, vol. 37, no. 5, December, pp. 1071-1086. Myers, S.C. 1984, ‘The Capital Structure Puzzle’, Journal of Finance, vol. 39, no. 3, July, pp. 575-592. Ofer, A. R. & Thakor, A.V. 1987, ‘A Theory of Stock Price Responses to Alternative Corporate Cash Disbursement Methods: Stock Repurchases and Dividends’, Journal of Finance, vol. 42, no. 2, June, pp. 365-394 Page, C & Meyer, D. 2000, Applied Research Design for Business and Management, Irwin McGraw-Hill, London. Penman, S.H. & Sougiannis T. 1998, ‘A Comparison of Dividend, Cash Flow, and Earnings Approaches to Equity Valuation’, Contemporary Accounting Research, vol. 15, no. 3, Fall, pp. 343-383. Perry, C. 1995, ‘A Structured Approach to Presenting PhD Theses: Notes for Candidates and Their Supervisors’, Research Workshop – Handout, pp. 45-85, Southern Cross University. Perry, C. 1998, ‘A Structured Approached for Presenting Theses’, Australasian Marketing Journal, vol. 6, no. 1, pp. 63-86. Petty, J.W. et al. 2000, Financial Management, 2nd edition, Prentice Hall, Sydney, New South Wales. 246 DBA Thesis: Malinee Ronapat Pringvanich, N. 2002, Thailand Overview, December, Bangkok. Radelet, S. & Schs, J. 1998, ‘The Onset of the East Asian Financial Crisis’, Paper Presented to the National Bureau of Economic Research Currency Crises Conference, February 6-7 1998. Rahman S., Yui H.P. & Foon, H.S. 2003, ‘Comparative Speed of Recovery of the Asian Economic Crisis Driven Countries’, International Trade Counselling Centre, The Seventh International Conference on Global Business and Economic Development, pp. 495-511. Rappaport, A. 1986 ‘The Affordable Dividend Approach to Equity Valuation’, Financial Analysts Journal, vol. 42, no. 4, July/August, pp. 52-59. Reilly F.K. & Brown K.C. 2000, ‘Investment Analysis and Portfolio Management’, 6th edition, Harcourt College Publishers, Fort Worth, Texas. Reilly F.K. & Norton, E.A. 1999, Investment, 5th edition, Dryden Press, Fort Worth, Texas. Richardson, G., Sefcik, S.E. & Thompson, R. 1986, ‘A Test of Dividend Irrelevance Using Volume Reaction to a Change in Dividend Policy, Journal of Financial Economics, vol. 17, no. 2, April, pp. 313-333. Rodriguez, R.J. 1992, ‘Quality Dispersion and the Feasibility of Dividends as Signals’, Journal of Financial Research, vol. 15, no. 4, September, pp. 307-315. Roll, R., 1977, ‘A Critique of the Asset Price Theory's Test; Part I: On Past and Potential Testability of Theory’, Journal of Financial Economics, vol. 4, no. 2, March, pp. 129-176. Roll, R. 1983, ‘Vas ist das?’, Journal of Portfolio Management, vol. 9, no.2, Winter, pp. 18-28. Saengkaew, V. 1996, Statistic for Business Management, 6th edition, Assumption University, Sarm Dee Press, Bangkok. SCU, 1995, DBA Manual, Southern Cross University. SCU, 2000, DBA Manual, Southern Cross University. SCU, 2002, DBA Manual, Southern Cross University. Sekeran, M. 2000, Research Methods for Business. SEC 2001, Annual Report 2001, Investment Management Supervision Department, Investment Advice Supervision Department and Broker-Dealer Supervision Department, The Stock Exchange Commission of Thailand, Bangkok. 247 DBA Thesis: Malinee Ronapat SEC 2002a, Annual Report 2002, Investment Management Supervision Department, Investment Advice Supervision Department and Broker-Dealer Supervision Department, The Stock Exchange Commission of Thailand, Bangkok. SEC 2002b, ‘Investors Behaviour and Understanding about Investment’, Strategic Management and Development, The Stock Exchange Commission of Thailand, Bangkok. SEC 2003, Annual Report 2003, Investment Management Supervision Department, Investment Advice Supervision Department and Broker-Dealer Supervision Department, The Stock Exchange Commission of Thailand, Bangkok. SET 1990, Fact Book 1990, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1991, Fact Book 1991, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1992, Fact Book 1992, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1993, Fact Book 1993, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1994, Fact Book 1994, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1995, Fact Book 1995, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1996, Fact Book 1996, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1997, Fact Book 1997, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1998, Fact Book 1998, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 1999, Fact Book 1999, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 2000a, Annual Report, Investor Relations and Communications Department, The Stock Exchange of Thailand, Bangkok. SET 2000b, Fact Book 2000, Publication Department, The Stock Exchange of Thailand, Bangkok. 248 DBA Thesis: Malinee Ronapat SET 2001a, Annual Report, Investor Relations and Communications Department, The Stock Exchange of Thailand, Bangkok. SET 2001b, Fact Book 2001, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 2002a, Annual Report, Investor Relations and Communications Department, The Stock Exchange of Thailand, Bangkok. SET 2002b, Fact Book 2002, Publication Department, The Stock Exchange of Thailand, Bangkok. SET 2002c, The Stock Exchange in Thailand 2002, Investor Relations and Communications Department, The Stock Exchange of Thailand, Bangkok. SET 2003, Fact Book 2003, Publication Department, The Stock Exchange of Thailand, Bangkok. Shefrin, H.M. & Statman, M. 1984, ‘Explaining Investor Preference for Cash Dividends’, Journal of Financial Economics, vol. 13, no. 2, May, pp. 253-282. Shiller, R.J. 1984, ‘Stock Prices and Social Dynamics’, Brookings Papers on Economic Activity, no. 2, October, pp. 457-510. Smith, A. 1937, The Wealth of Nations, Random House, New York. Sutcliffe, C. 1993, Stock Index Futures: Theories and International Evidence, Chapman and Hall Series in Accounting and Finance, Chapman and Hall, London. Sutcliffe, C. 1997, Stock Index Futures: Theories and International Evidence, Second edition, International Thomson Business Press, London. Tabachnick, B.G. & Fidell, L. 2001, Using Mutivariate Analysis, 4th edition, McGrawHill, New York. Talmor, E., 1981, ‘Asymmetric Information, Signaling, and Optimal Corporate Financial Decisions’, Journal of Financial and Quantitative Analysis, vol. 16, no. 4, August, pp. 413-435. Thaler, R.H. & Shefrin, H.M. 1981, ‘An Economic Theory of Self Control’, Journal of Political Economy, vol. 89, no. 1, April, pp. 392-410. Ticehurst, G.W. & Veal, A.J. 2000, Business Research Methods: A Managerial Approach, Longman, French Forest, New South Wales. Tobin, J. 1958, ‘Liquidity Preference as Behaviour Towards Risk’, Review of Economic Studies, vol. 25, no. 67, February, pp. 65-86. 249 DBA Thesis: Malinee Ronapat TRIS 2001, ‘Moody’s and S&P’s Credit Rating’, TRISVIEW, no. 2, January. TRIS 2002, ‘Sovereign Credit Rating of Thailand’, TRISVIEW, no. 2, December. Turnovsky, S.J. 1967, ‘The Allocation of Corporate Profits Between Dividends and Retained Earnings’, Review of Economics and Statistics, vol. 49, no. 1, January, pp. 583589. Watts, R. 1973, ‘The Information Content of Dividends’, Journal of Business, vol. 46, no. 2, Spring, pp. 191-211. Watts, R. 1976a, ‘Comments on The Informational Content of Dividends’, Journal of Business, vol. 49, no. 1, January, pp. 81-85. Watts, R. 1976b, ‘Comments on "The Impact of Dividend and Earnings Announcements": A Reconciliation’, Journal of Business, vol. 49, no. 1, January, pp. 97-106. Weisbenner, S.J. 1999, Corporate Share Repurchase in the Mid-1990s: What Role Do Stock Option Play?, Federal Reserve Board Working Paper. Wetherilt, V. & Weeken, O. 2002, ‘Equity Valuation Measures: What Can They Tell Us?’, Quarterly Bulletin, Bank of England Financial Stability Review, Monetary Instruments and Markets Division, Winter, pp. 391-403. Yao, Y. 1997, ‘A Trinomial Dividend Valuation Model’, Journal of Portfolio Management, pp. 99-103. Yardeni, E. 2003, ‘Stock Valuation Models 4.1’, Prudential Financial Research, January. Zikmund, W.G. 2000, Business Research Methods, 6th edition, Dryden Press, Fort Worth, Texas. 250 DBA Thesis: Malinee Ronapat Appendices Appendix A: Descriptive statistics and ANOVA comparing payers and non-payers 251 Appendix B: ANOVA comparing payers, former payers, never paid firms 264 Appendix C: Logit Regression results controlling two investment variables 277 Appendix D: Logit Regression results controlling one investment variable 280 Appendix E: t-statistic results (two investment variables) 283 Appendix F: t-statistic results for (one investment variable) 285 Appendix G: Expected percentage of payers controlling two investment variables 287 (crosstabs) Appendix H: Expected percentage of payers controlling one investment variables (crosstabs) 292 251 DBA Thesis: Malinee Ronapat APPENDIX A: DESCRIPTIVE STATISTICS AND ANOVA: COMPARING PAYERS AND NON-PAYERS 252 DBA Thesis: Malinee Ronapat Table A1: Descriptive statistics: comparing payers and non-payers (1991) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum Upper Bound .00 32 .1220 .06134 .01084 .0999 .1442 .02 .30 1.00 192 .1096 .06955 .00502 .0997 .1195 -.42 .43 Total 224 .1113 .06845 .00457 .1023 .1203 -.42 .43 .00 32 .1661 .10292 .01819 .1290 .2032 .00 .48 1.00 192 .1663 .18261 .01318 .1403 .1923 -.49 2.00 Total 224 .1662 .17330 .01158 .1434 .1891 -.49 2.00 .00 32 .1247 1.22796 .21707 -.3181 .5674 -6.18 1.00 1.00 192 .3185 .47411 .03422 .2511 .3860 -2.76 1.00 Total 224 .2908 .63778 .04261 .2069 .3748 -6.18 1.00 .00 32 6.4437 12.92814 2.28539 1.7826 11.1048 .46 63.74 1.00 192 4.0363 7.22362 .52132 3.0080 5.0646 .28 69.34 Total 224 4.3802 8.28492 .55356 3.2894 5.4711 .28 69.34 .00 32 3435.8506 6880.07394 1216.23674 955.3194 5916.3818 58.67 32611.85 1.00 192 14257.5330 58657.01194 4233.20520 5907.6966 22607.3693 137.67 595801.1 Total 224 12711.5783 54478.47848 3639.99645 5538.3867 19884.7700 58.67 595801.1 .00 32 .4366 .24846 .04392 .3470 .5262 .04 .95 1.00 192 .5007 .24053 .01736 .4665 .5349 .01 1.15 Total 224 .4915 .24215 .01618 .4597 .5234 .01 1.15 Source: Developed for this research Table A2: ANOVA: comparing payers and non-payers (1991) Sum of Squares Et/At Yt/BEt dAt/At Vt/At At Lt/At Between Groups df Mean Square .004 1 .004 Within Groups 1.040 222 .005 Total 1.045 223 Between Groups .000 1 .000 Within Groups 6.698 222 .030 Total 6.698 223 Between Groups 1.031 1 1.031 Within Groups 89.678 222 .404 Total 90.709 223 Between Groups 158.965 1 158.965 Within Groups 15147.748 222 68.233 Total 15306.712 223 Between Groups 3212127325.688 1 3212127325.688 Within Groups 658630602430.805 222 2966804515.454 Total 661842729756.492 223 Between Groups .113 1 .113 Within Groups 12.964 222 .058 Total 13.076 223 F Sig. .913 .340 .000 .995 2.552 .112 2.330 .128 1.083 .299 1.928 .166 Source: Developed for this research 253 DBA Thesis: Malinee Ronapat Table A3: Descriptive statistics: comparing payers and non-payers (1992) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum Upper Bound .00 24 .0584 .11604 .02369 .0094 .1074 -.21 .33 1.00 230 .1059 .07370 .00486 .0963 .1154 -.01 .71 Total 254 .1014 .07958 .00499 .0916 .1112 -.21 .71 .00 24 -.0301 .49116 .10026 -.2375 .1773 -1.81 1.01 1.00 230 .1626 .19602 .01293 .1371 .1880 -.29 2.52 Total -1.81 2.52 254 .1444 .24475 .01536 .1141 .1746 .00 24 .4898 .47189 .09632 .2905 .6890 -.25 1.00 1.00 230 .2866 .39351 .02595 .2355 .3377 -2.58 1.00 Total 254 .3058 .40491 .02541 .2558 .3558 -2.58 1.00 .00 24 2.0842 .80515 .16435 1.7442 2.4241 .99 4.05 1.00 230 1.8991 1.26107 .08315 1.7352 2.0629 .52 17.03 Total 254 1.9165 1.22528 .07688 1.7651 2.0680 .52 17.03 .00 24 5586.6921 17914.72551 3656.82803 -1978.0331 13151.4172 285.50 89174.37 1.00 230 13086.9055 60483.38340 3988.15620 5228.7331 20945.0779 59.83 666008.7 Total 254 12378.2239 57837.88692 3629.07179 5231.1851 19525.2628 59.83 666008.7 .00 24 .5655 .21255 .04339 .4758 .6553 .04 .91 1.00 230 .5031 .21815 .01438 .4747 .5314 .04 1.09 Total 254 .5090 .21798 .01368 .4821 .5359 .04 1.09 Source: Developed for this research Table A4: ANOVA comparing: payers and non-payers (1992) Sum of Squares Et/At Yt/BEt dAt/At Vt/At Between Groups .049 Within Groups 1.553 252 .006 Total 1.602 253 .807 1 .807 Within Groups 14.348 252 .057 Total 15.155 253 Between Groups Between Groups .897 1 .897 Within Groups 40.583 252 .161 Total 41.480 253 .745 1 .745 379.087 252 1.504 Between Groups Total Lt/At Mean Square 1 Within Groups At df .049 379.831 253 1222510516.344 1 1222510516.344 Within Groups 845118443932.337 252 3353644618.779 Total 846340954448.681 253 Between Groups Between Groups .085 1 .085 Within Groups 11.937 252 .047 Total 12.022 253 F Sig. 7.941 .005 14.172 .000 5.570 .019 .495 .482 .365 .547 1.789 .182 Source: Developed for this research 254 DBA Thesis: Malinee Ronapat Table A5: Descriptive statistics: comparing payers and non-payers (1993) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum Upper Bound .00 43 .0373 .10815 .01649 .0040 .0706 -.30 .35 1.00 244 .0992 .10145 .00649 .0864 .1120 -.13 1.12 Total 287 .0900 .10465 .00618 .0778 .1021 -.30 1.12 .00 43 .1428 .54845 .08364 -.0260 .3116 -.62 2.71 1.00 244 .1285 .20362 .01304 .1028 .1541 -1.77 1.27 Total -1.77 2.71 287 .1306 .28183 .01664 .0979 .1633 .00 43 .4805 .44028 .06714 .3450 .6160 -.11 1.00 1.00 244 .1930 .49242 .03152 .1309 .2551 -6.10 1.00 Total 287 .2360 .49503 .02922 .1785 .2935 -6.10 1.00 .00 43 2.4644 2.01057 .30661 1.8456 3.0831 .77 10.74 1.00 244 2.4169 2.73211 .17491 2.0723 2.7614 .74 32.56 Total 287 2.4240 2.63364 .15546 2.1180 2.7300 .74 32.56 .00 43 6185.4342 17854.48186 2722.78296 690.6357 11680.2327 359.78 112231.5 1.00 244 14693.3015 69420.14214 4444.16920 5939.2908 23447.3122 61.42 782870.4 Total 287 13418.6036 64425.64905 3802.92582 5933.3304 20903.8769 61.42 782870.4 .00 43 .5441 .26420 .04029 .4628 .6254 .09 1.15 1.00 244 .5515 .49332 .03158 .4893 .6137 .00 5.65 Total 287 .5504 .46587 .02750 .4963 .6045 .00 5.65 Source: Developed for this research Table A6: ANOVA: comparing payers and non-payers (1993) Sum of Squares Et/At Yt/BEt dAt/At Vt/At Between Groups .140 Within Groups 2.992 285 .010 Total 3.132 286 .008 1 .008 Within Groups 22.709 285 .080 Total 22.716 286 Between Groups Between Groups 3.022 1 3.022 Within Groups 67.063 285 .235 Total 70.084 286 .082 1 .082 1983.630 285 6.960 Between Groups Total Lt/At Mean Square 1 Within Groups At df .140 1983.713 286 2646170367.499 1 2646170367.499 Within Groups 1184443806742.514 285 4155943181.553 Total 1187089977110.013 286 Between Groups Between Groups .002 1 .002 Within Groups 62.070 285 .218 Total 62.072 286 F Sig. 13.359 .000 .094 .759 12.842 .000 .012 .913 .637 .426 .009 .924 Source: Developed for this research 255 DBA Thesis: Malinee Ronapat Table A7: Descriptive statistics: comparing payers and non-payers (1994) N Mean Std. Deviation 95% Confidence Interval for Mean Std. Error Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum Upper Bound .00 52 .0374 .08448 .01172 .0138 .0609 -.27 .22 1.00 273 .0894 .05828 .00353 .0825 .0964 -.14 .48 Total 325 .0811 .06588 .00365 .0739 .0883 -.27 .48 .00 52 -1.1973 8.67080 1.20242 -3.6112 1.2167 -62.37 2.31 1.00 273 .1269 .12339 .00747 .1122 .1416 -.61 1.15 Total 325 -.0850 3.47613 .19282 -.4643 .2943 -62.37 2.31 .00 52 .2693 .48244 .06690 .1350 .4036 -1.57 1.00 1.00 273 .2321 .50844 .03077 .1715 .2927 -5.84 1.00 Total 325 .2381 .50383 .02795 .1831 .2931 -5.84 1.00 .00 52 1.6024 .87804 .12176 1.3580 1.8469 .73 6.28 1.00 273 1.6913 1.08964 .06595 1.5615 1.8211 .35 10.04 Total 325 1.6771 1.05791 .05868 1.5616 1.7925 .35 10.04 .00 52 11566.3871 41476.29233 5751.72687 19.3136 23113.4606 336.90 286428.5 1.00 273 15581.7186 76425.69599 4625.49354 6475.3991 24688.0380 64.43 898373.5 Total 325 14939.2655 71947.33881 3990.92030 7087.8771 22790.6540 64.43 898373.5 .00 52 .5625 .21827 .03027 .5017 .6232 .07 1.10 1.00 273 .4930 .25252 .01528 .4629 .5231 .00 2.96 Total 325 .5041 .24836 .01378 .4770 .5312 .00 2.96 Source: Developed for this research Table A8: ANOVA: comparing payers and non-payers (1994) Sum of Squares Et/At Between Groups Within Groups Total Yt/BEt dAt/At Vt/At .118 1.288 323 .004 1.406 324 1 76.584 Within Groups 3838.464 323 11.884 Total 3915.048 324 Between Groups .060 1 .060 Within Groups 82.186 323 .254 Total 82.247 324 .345 1 .345 362.266 323 1.122 Between Groups Total Lt/At Mean Square 1 76.584 Between Groups Within Groups At df .118 362.611 324 704247691.139 1 704247691.139 Within Groups 1676455690198.802 323 5190265294.733 Total 1677159937889.941 324 Between Groups Between Groups .211 1 .211 Within Groups 19.775 323 .061 Total 19.986 324 F Sig. 29.695 .000 6.444 .012 .238 .626 .308 .580 .136 .713 3.442 .064 Source: Developed for this research 256 DBA Thesis: Malinee Ronapat Table A9: Descriptive statistics: comparing payers and non-payers (1995) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum Upper Bound .00 54 .0208 .08060 .01097 -.0012 .0428 -.22 .21 1.00 290 .0884 .05193 .00305 .0824 .0944 -.07 .37 Total 344 .0778 .06232 .00336 .0712 .0844 -.22 .37 .00 54 .0331 .68460 .09316 -.1538 .2200 -2.27 3.90 1.00 290 .1192 .12572 .00738 .1047 .1338 -.72 1.03 Total -2.27 3.90 344 .1057 .29448 .01588 .0745 .1369 .00 54 .1280 .32943 .04483 .0381 .2179 -.60 1.00 1.00 290 .2045 .37735 .02216 .1609 .2481 -4.04 1.00 Total 344 .1925 .37084 .01999 .1531 .2318 -4.04 1.00 .00 54 1.3616 .63364 .08623 1.1886 1.5345 .60 3.77 1.00 290 1.3891 .77130 .04529 1.3000 1.4783 .29 6.20 Total 344 1.3848 .75059 .04047 1.3052 1.4644 .29 6.20 .00 54 5925.6809 11092.62927 1509.51564 2897.9760 8953.3859 495.38 69338.75 1.00 290 19275.3505 88614.49312 5203.62221 9033.5478 29517.1532 301.84 1035447 Total 344 17179.7628 81602.30361 4399.70135 8525.9714 25833.5543 301.84 1035447 .00 54 .6028 .24187 .03291 .5368 .6688 .00 1.45 1.00 290 .5151 .20429 .01200 .4915 .5387 .01 .94 Total 344 .5289 .21266 .01147 .5063 .5514 .00 1.45 Source: Developed for this research Table A10: ANOVA: comparing payers and non-payers (1995) Sum of Squares Et/At Yt/BEt dAt/At Vt/At At Lt/At Between Groups df Mean Square .208 1 .208 Within Groups 1.124 342 .003 Total 1.332 343 Between Groups .338 1 .338 Within Groups 29.407 342 .086 Total 29.745 343 Between Groups .266 1 .266 Within Groups 46.903 342 .137 Total 47.169 343 .035 1 .035 Within Groups 193.205 342 .565 Total 193.240 343 Between Groups Between Groups 8112866862.121 1 8112866862.121 Within Groups 2275902165295.430 342 6654684693.846 Total 2284015032157.551 343 Between Groups .350 1 .350 Within Groups 15.161 342 .044 Total 15.511 343 F Sig. 63.380 .000 3.929 .048 1.940 .165 .061 .805 1.219 .270 7.895 .005 Source: Developed for this research 257 DBA Thesis: Malinee Ronapat Table A11: Descriptive statistics: comparing payers and non-payers (1996) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum Upper Bound .00 72 .0032 .12794 .01508 -.0269 .0333 -.73 .25 1.00 306 .0753 .05837 .00334 .0688 .0819 -.11 .52 Total 378 .0616 .08151 .00419 .0534 .0698 -.73 .52 .00 72 -.0627 .58979 .06951 -.2013 .0759 -2.80 2.63 1.00 306 .0844 .19082 .01091 .0629 .1059 -2.05 1.34 Total -2.80 2.63 378 .0564 .31355 .01613 .0247 .0881 .00 72 .1050 .39639 .04671 .0119 .1981 -1.36 1.00 1.00 306 .1316 .93697 .05356 .0262 .2370 -15.12 1.00 Total 378 .1265 .86020 .04424 .0395 .2135 -15.12 1.00 .00 72 1.1447 .83235 .09809 .9491 1.3402 .42 6.82 1.00 306 1.1837 .72083 .04121 1.1027 1.2648 .25 6.51 Total 378 1.1763 .74234 .03818 1.1012 1.2514 .25 6.82 .00 72 6129.9290 12463.49563 1468.83705 3201.1511 9058.7069 248.14 85913.50 1.00 306 21576.2876 97755.05651 5588.28463 10579.8156 32572.7596 294.15 1155109 Total 378 18634.1240 88301.55432 4541.74188 9703.8042 27564.4439 248.14 1155109 .00 72 .5944 .22530 .02655 .5415 .6473 .11 1.44 1.00 306 .5534 .38156 .02181 .5105 .5963 .00 5.75 Total 378 .5612 .35721 .01837 .5251 .5973 .00 5.75 Source: Developed for this research Table A12: ANOVA: comparing payers and non-payers (1996) Sum of Squares Et/At Yt/BEt dAt/At Vt/At Between Groups .303 Within Groups 2.201 376 .006 Total 2.505 377 Between Groups 1.261 1 1.261 Within Groups 35.803 376 .095 Total 37.064 377 Between Groups .041 1 .041 Within Groups 278.918 376 .742 Total 278.959 377 .089 1 .089 207.665 376 .552 Between Groups Total Lt/At Mean Square 1 Within Groups At df .303 207.754 377 13906388138.733 1 13906388138.733 Within Groups 2925624626874.058 376 7780916560.835 Total 2939531015012.791 377 Between Groups Between Groups .098 1 .098 Within Groups 48.008 376 .128 Total 48.106 377 F Sig. 51.805 .000 13.242 .000 .056 .814 .161 .688 1.787 .182 .767 .382 Source: Developed for this research 258 DBA Thesis: Malinee Ronapat Table A13: Descriptive statistics: comparing payers and non-payers (1997) N Mean Std. Deviation 95% Confidence Interval for Mean Std. Error Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum Upper Bound .00 112 -.1964 .34283 .03239 -.2605 -.1322 -2.08 .38 1.00 281 -.0536 .22506 .01343 -.0801 -.0272 -1.60 .51 Total 393 -.0943 .27133 .01369 -.1212 -.0674 -2.08 .51 .00 112 -3.9705 31.00026 2.92925 -9.7750 1.8340 -325.37 13.21 1.00 281 -2.4684 33.22680 1.98214 -6.3702 1.4334 -550.44 10.03 Total 393 -2.8965 32.57562 1.64322 -6.1271 .3342 -550.44 13.21 .00 112 -.0466 .71001 .06709 -.1796 .0863 -4.33 1.00 1.00 281 .0014 .43729 .02609 -.0500 .0527 -2.98 1.00 Total 393 -.0123 .52897 .02668 -.0647 .0402 -4.33 1.00 .00 112 1.0166 .44518 .04207 .9333 1.1000 .40 3.49 1.00 281 1.0482 .57407 .03425 .9808 1.1156 .08 5.09 Total 393 1.0392 .54011 .02724 .9856 1.0928 .08 5.09 .00 112 9047.9238 21683.09435 2048.85983 4987.9712 13107.8763 377.12 190277.4 1.00 281 25384.9138 122294.36702 7295.47030 11023.9811 39745.8464 248.85 1408619 Total 393 20729.0795 104261.44910 5259.29439 10389.1273 31069.0316 248.85 1408619 .00 112 .8211 .39637 .03745 .7469 .8953 .18 3.46 1.00 281 .6258 .31882 .01902 .5883 .6632 .00 2.03 Total 393 .6814 .35339 .01783 .6464 .7165 .00 3.46 Source: Developed for this research Table A14: ANOVA: comparing payers and non-payers (1997) Sum of Squares Et/At Yt/BEt dAt/At Vt/At At Lt/At Between Groups df Mean Square 1.631 1 1.631 Within Groups 27.228 391 .070 Total 28.860 392 Between Groups 180.685 1 180.685 Within Groups 415798.385 391 1063.423 Total 415979.070 392 .185 1 .185 Within Groups 109.499 391 .280 Total 109.684 392 Between Groups Between Groups .080 1 .080 Within Groups 114.273 391 .292 Total 114.353 392 Between Groups 21373511506.616 1 21373511506.616 Within Groups 4239842797717.739 391 10843587717.948 Total 4261216309224.355 392 Between Groups 3.055 1 3.055 Within Groups 45.901 391 .117 Total 48.956 392 F Sig. 23.423 .000 .170 .680 .659 .417 .273 .602 1.971 .161 26.023 .000 Source: Developed for this research 259 DBA Thesis: Malinee Ronapat Table A15: Descriptive statistics: comparing payers and non-payers (1998) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At .00 278 .0195 .70414 .04223 1.00 110 .0956 .09681 Total 388 .0411 .59892 .00 278 -9.8999 1.00 110 Total .00 1.00 Minimum -.0636 .1026 -2.75 .00923 .0774 .1139 -.15 .44 .03041 -.0187 .1009 -2.75 10.23 155.80777 9.34473 -28.2956 8.4958 -2597.13 25.05 .1054 .39044 .03723 .0316 .1792 -.77 3.43 388 -7.0633 131.89511 6.69596 -20.2283 6.1017 -2597.13 25.05 278 -1.2208 15.44686 .92644 -3.0445 .6030 -253.88 1.00 110 -.0236 .20114 .01918 -.0616 .0144 -1.12 1.00 Total 388 -.8814 13.08006 .66404 -2.1869 .4242 -253.88 1.00 .00 278 2.2227 15.10458 .90591 .4393 4.0060 .22 239.23 1.00 110 1.0109 .59571 .05680 .8984 1.1235 .20 3.87 Total 388 1.8791 12.79449 .64954 .6021 3.1562 .20 239.23 .00 278 9186.2069 23507.93125 1409.91202 6410.7034 11961.7105 20.38 208724.5 1.00 110 47848.2083 189230.91295 18042.45962 12088.6399 83607.7767 281.56 1266949 Total 388 20147.0836 103853.32174 5272.35360 9781.0418 30513.1254 20.38 1266949 .00 278 1.7203 12.65716 .75913 .2259 3.2146 .01 207.51 1.00 110 .3895 .24863 .02371 .3426 .4365 .01 .95 Total 388 1.3430 10.72594 .54453 .2724 2.4136 .01 207.51 Source: Developed for this research Table A16: ANOVA: comparing payers and non-payers (1998) Sum of Squares Et/At Between Groups Within Groups Total Yt/BEt dAt/At Vt/At Mean Square 1 .457 138.362 386 .358 138.819 387 1 7889.733 Within Groups 6724485.843 386 17420.948 Total 6732375.575 387 Between Groups 112.966 1 112.966 Within Groups 66098.111 386 171.239 Total 66211.077 387 115.725 1 115.725 63235.767 386 163.823 Between Groups Total Lt/At df .457 7889.733 Between Groups Within Groups At Maximum Upper Bound 63351.492 387 117807901049.125 1 117807901049.125 Within Groups 4056185411677.855 386 10508252361.860 Total 4173993312726.979 387 Between Groups Between Groups 139.564 1 139.564 Within Groups 44383.159 386 114.982 Total 44522.723 387 F Sig. 1.275 .260 .453 .501 .660 .417 .706 .401 11.211 .001 1.214 .271 Source: Developed for this research 260 10.23 DBA Thesis: Malinee Ronapat Table A17: Descriptive statistics: comparing payers and non-payers (1999) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At .00 270 1.00 Total -.0840 .43517 .02648 115 .0884 .07908 385 -.0325 .37518 .00 270 -.9563 1.00 115 Total .00 1.00 Minimum Maximum -3.55 1.60 Upper Bound -.1362 -.0319 .00737 .0738 .1030 -.16 .36 .01912 -.0701 .0051 -3.55 1.60 6.90977 .42052 -1.7842 -.1284 -88.22 4.87 .1087 .12965 .01209 .0848 .1327 -.40 .43 385 -.6382 5.80427 .29581 -1.2198 -.0565 -88.22 4.87 270 -.1056 .56979 .03468 -.1739 -.0373 -4.27 1.00 115 .0543 .18969 .01769 .0193 .0894 -.33 1.00 Total 385 -.0578 .49345 .02515 -.1073 -.0084 -4.27 1.00 .00 270 1.3639 1.47699 .08989 1.1869 1.5409 .07 17.50 1.00 115 1.0019 .71886 .06703 .8691 1.1347 .27 6.98 Total 385 1.2558 1.30733 .06663 1.1248 1.3868 .07 17.50 .00 270 29084.0368 119161.81966 7251.95740 14806.2237 43361.8499 219.57 1182878 1.00 115 3034.9757 3855.20365 359.49959 2322.8098 3747.1417 288.50 21983.67 Total 385 21303.1484 100468.89164 5120.36885 11235.6791 31370.6177 219.57 1182878 .00 270 1.0047 1.17016 .07121 .8645 1.1449 .05 10.87 1.00 115 .3374 .19475 .01816 .3015 .3734 .01 .79 Total 385 .8054 1.03149 .05257 .7020 .9087 .01 10.87 Source: Developed for this research Table A18: ANOVA: comparing payers and non-payers (1999) Sum of Squares Et/At Yt/BEt dAt/At Vt/At Between Groups 2.397 Within Groups 51.655 383 .135 Total 54.052 384 Between Groups 91.477 1 91.477 Within Groups 12845.297 383 33.539 Total 12936.774 384 Between Groups 2.062 1 2.062 Within Groups 91.437 383 .239 Total 93.499 384 Between Groups 10.566 1 10.566 645.736 383 1.686 Total Lt/At Mean Square 1 Within Groups At df 2.397 656.302 384 54724905688.052 1 54724905688.052 Within Groups 3821370398100.955 383 9977468402.352 Total 3876095303789.008 384 Between Groups Between Groups 35.904 1 35.904 Within Groups 372.657 383 .973 Total 408.562 384 F Sig. 17.776 .000 2.728 .099 8.638 .003 6.267 .013 5.485 .020 36.901 .000 Source: Developed for this research 261 DBA Thesis: Malinee Ronapat Table A19: Descriptive statistics comparing payers and non-payers (2000) N Mean Std. Deviation 95% Confidence Interval for Mean Std. Error Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At .00 238 1.00 Total -.1070 .84208 .05458 130 .0850 .06391 368 -.0392 .68396 .00 238 -.7258 1.00 130 Total .00 1.00 Minimum Maximum Upper Bound -.2145 .0005 -10.38 1.09 .00560 .0739 .0961 -.16 .30 .03565 -.1093 .0309 -10.38 1.09 9.11072 .59056 -1.8892 .4376 -115.08 23.63 .1092 .09381 .00823 .0929 .1255 -.20 .36 368 -.4308 7.33251 .38223 -1.1825 .3208 -115.08 23.63 238 -.3209 2.06700 .13398 -.5849 -.0570 -28.92 1.00 130 .0314 .29794 .02613 -.0203 .0831 -1.86 1.00 Total 368 -.1964 1.67890 .08752 -.3685 -.0243 -28.92 1.00 .00 238 1.4839 2.69624 .17477 1.1396 1.8282 .22 35.69 1.00 130 .8666 .54561 .04785 .7719 .9613 .13 5.55 Total 368 1.2658 2.21055 .11523 1.0392 1.4924 .13 35.69 .00 238 32136.7395 129928.04388 8421.98418 15545.2284 48728.2505 187.39 1236328 1.00 130 3893.1621 6539.65238 573.56546 2758.3488 5027.9754 302.36 51960.20 Total 368 22159.3888 105353.32673 5491.92180 11359.8049 32958.9726 187.39 1236328 .00 238 1.2658 2.71992 .17631 .9184 1.6131 .00 35.51 1.00 130 .3164 .18941 .01661 .2835 .3492 .00 .73 Total 368 .9304 2.23530 .11652 .7012 1.1595 .00 35.51 Source: Developed for this research Table A20: ANOVA comparing payers and non-payers (2000) Sum of Squares Et/At Yt/BEt dAt/At Vt/At Between Groups 3.099 Within Groups 168.584 366 .461 Total 171.683 367 58.614 1 58.614 Within Groups 19673.390 366 53.752 Total 19732.005 367 Between Groups Between Groups 10.436 1 10.436 Within Groups 1024.034 366 2.798 Total 1034.470 367 32.035 1 32.035 1761.329 366 4.812 Between Groups Total Lt/At Mean Square 1 Within Groups At df 3.099 1793.364 367 67067466394.794 1 67067466394.794 Within Groups 4006384241002.584 366 10946405030.062 Total 4073451707397.378 367 Between Groups Between Groups 75.784 1 75.784 Within Groups 1757.952 366 4.803 Total 1833.736 367 F Sig. 6.728 .010 1.090 .297 3.730 .054 6.657 .010 6.127 .014 15.778 .000 Source: Developed for this research 262 DBA Thesis: Malinee Ronapat Table A21: Descriptive statistics comparing payers and non-payers (2001) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum 7.28 Upper Bound .00 223 .0238 .54671 .03661 -.0484 .0959 -1.41 1.00 153 -.0728 2.03734 .16471 -.3982 .2527 -25.09 .41 Total 376 -.0155 1.36442 .07036 -.1539 .1228 -25.09 7.28 .00 223 -4.4534 56.32197 3.77160 -11.8861 2.9793 -823.12 3.88 1.00 153 .1200 .10318 .00834 .1035 .1365 -.13 .52 Total 376 -2.5924 43.39339 2.23784 -6.9927 1.8079 -823.12 3.88 .00 223 -2.1588 31.32864 2.09792 -6.2932 1.9756 -467.83 1.00 1.00 153 -.0202 1.12787 .09118 -.2004 .1599 -13.61 1.00 Total 376 -1.2886 24.13837 1.24484 -3.7363 1.1592 -467.83 1.00 .00 223 1.9945 5.08704 .34065 1.3232 2.6659 .26 55.65 1.00 153 .9897 .58141 .04700 .8968 1.0826 .18 5.31 Total 376 1.5856 3.96246 .20435 1.1838 1.9875 .18 55.65 .00 223 31918.3958 133049.88853 8909.67953 14360.0244 49476.7671 72.52 1248748 1.00 153 5289.2451 11474.18865 927.63317 3456.5259 7121.9643 265.60 100742.0 Total 376 21082.5978 103463.52221 5335.72375 10590.9099 31574.2856 72.52 1248748 .00 223 1.6500 5.02831 .33672 .9865 2.3136 .01 55.19 1.00 153 .3258 .20244 .01637 .2934 .3581 .00 .88 Total 376 1.1112 3.92544 .20244 .7131 1.5092 .00 55.19 Source: Developed for this research Table A22: ANOVA comparing payers and non-payers (2001) Sum of Squares Et/At Yt/BEt dAt/At Vt/At At Lt/At Between Groups df Mean Square .846 1 .846 Within Groups 697.266 374 1.864 Total 698.112 375 Between Groups 1897.999 1 1897.999 Within Groups 704222.008 374 1882.947 Total 706120.007 375 Between Groups 415.004 1 415.004 Within Groups 218082.793 374 583.109 Total 218497.797 375 91.620 1 91.620 Within Groups 5796.295 374 15.498 Total 5887.915 375 Between Groups Between Groups 64346225910.489 1 64346225910.489 Within Groups 3949916434699.211 374 10561273889.570 Total 4014262660609.700 375 Between Groups 159.137 1 159.137 Within Groups 5619.254 374 15.025 Total 5778.392 375 F Sig. .454 .501 1.008 .316 .712 .399 5.912 .016 6.093 .014 10.592 .001 Source: Developed for this research 263 DBA Thesis: Malinee Ronapat Table A23: Descriptive statistics comparing payers and non-payers (2002) N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Lower Bound Et/At Yt/BEt dAt/At Vt/At At Lt/At Minimum Maximum -1.52 2.09 Upper Bound .00 207 .0234 .25154 .01748 -.0111 .0579 1.00 179 .0908 .06516 .00487 .0812 .1004 -.09 .39 Total 386 .0547 .19223 .00978 .0354 .0739 -1.52 2.09 .00 207 .0161 1.03165 .07170 -.1253 .1574 -6.14 8.87 1.00 179 .1271 .10380 .00776 .1118 .1424 -.14 .58 Total 386 .0676 .75995 .03868 -.0085 .1436 -6.14 8.87 .00 207 -.0964 1.56272 .10862 -.3106 .1177 -19.32 1.00 1.00 179 .0473 .47023 .03515 -.0221 .1167 -4.81 1.00 Total 386 -.0298 1.18914 .06053 -.1488 .0892 -19.32 1.00 .00 207 1.5445 2.10332 .14619 1.2562 1.8327 .30 23.35 1.00 179 1.1108 .58533 .04375 1.0245 1.1971 .28 5.13 Total 386 1.3434 1.60387 .08163 1.1829 1.5039 .28 23.35 .00 207 35133.9306 143651.74933 9984.48678 15449.0490 54818.8123 93.47 1245098 1.00 179 8912.4088 27779.52726 2076.33935 4815.0004 13009.8172 261.19 234465.8 Total 386 22974.2094 107562.65648 5474.79507 12209.9694 33738.4493 93.47 1245098 .00 207 1.0454 1.93700 .13463 .7799 1.3108 .00 23.26 1.00 179 .3296 .20163 .01507 .2999 .3594 .00 .84 Total 386 .7134 1.46767 .07470 .5666 .8603 .00 23.26 Source: Developed for this research Table A24: ANOVA comparing payers and non-payers (2002) Sum of Squares Et/At Yt/BEt dAt/At Vt/At Between Groups .437 Within Groups 13.790 384 .036 Total 14.227 385 1.183 1 1.183 Within Groups 221.165 384 .576 Total 222.348 385 Between Groups Between Groups 1.983 1 1.983 Within Groups 542.430 384 1.413 Total 544.412 385 18.053 1 18.053 972.315 384 2.532 Between Groups Total Lt/At Mean Square 1 Within Groups At df .437 990.368 385 66001204221.708 1 66001204221.708 Within Groups 4388342947562.820 384 11427976425.945 Total 4454344151784.528 385 Between Groups Between Groups 49.175 1 49.175 Within Groups 780.140 384 2.032 Total 829.315 385 F Sig. 12.163 .001 2.055 .153 1.404 .237 7.130 .008 5.775 .017 24.205 .000 Source: Developed for this research 264 DBA Thesis: Malinee Ronapat APPENDIX B: ANOVA COMPARING PAYERS, FORMER PAYERS, NEVER PAID FIRMS 265 DBA Thesis: Malinee Ronapat Table B1: ANOVA comparing payers, former payers and never paid firms (1991) Dependent Variable (I) PAYER3 (J) PAYER3 Et/At 1.00 2.00 3.00 -.0165 .01502 .517 Lower Bound -.0519 Upper Bound .0190 -.0042 .02026 .976 -.0520 .0436 2.00 1.00 .0165 .01502 .517 -.0190 .0519 3.00 .0122 .02425 .869 -.0450 .0695 1.00 2.00 .0042 .02026 .976 -.0436 .0520 -.0122 .02425 .869 -.0695 .0450 2.00 -.0097 .03809 .965 -.0996 .0801 3.00 .0203 .05137 .918 -.1010 .1415 1.00 3.00 1.00 .0097 .0300 .03809 .06150 .965 .877 -.0801 -.1151 .0996 .1751 3.00 Yt/BEt 1.00 2.00 3.00 dAt/At Vt/At 95% Confidence Interval Sig. -.0203 .05137 .918 -.1415 .1010 -.0300 .06150 .877 -.1751 .1151 1.00 2.00 .4264(*) .13627 .006 .1049 .7479 -.3170 .18382 .198 -.7506 .1167 2.00 3.00 1.00 -.4264(*) .13627 .006 -.7479 -.1049 3.00 -.2242 -.7434(*) .22005 .002 -1.2625 3.00 1.00 .3170 .18382 .198 -.1167 .7506 1.00 2.00 2.00 3.00 .7434(*) -2.9025 .22005 1.80915 .002 .246 .2242 -7.1707 1.2625 1.3658 -.0196 2.44035 1.000 -5.7771 5.7378 2.00 1.00 2.9025 1.80915 .246 -1.3658 7.1707 3.00 2.8828 2.92139 .586 -4.0095 9.7751 1.00 2.00 .0196 2.44035 1.000 -5.7378 5.7771 -2.8828 2.92139 .586 -9.7751 4.0095 2.00 9922.2383 11908.01231 .683 -18171.8597 38016.3363 3.00 12865.6516 16062.61734 .703 -25030.2402 50761.5434 1.00 3.00 1.00 -9922.2383 2943.4133 11908.01231 19228.86596 .683 .987 -38016.3363 -42422.4825 18171.8597 48309.3090 -12865.6516 16062.61734 .703 -50761.5434 25030.2402 2.00 -2943.4133 19228.86596 .987 -48309.3090 42422.4825 1.00 2.00 .0110 .05226 .976 -.1123 .1343 3.00 1.00 .2091(*) .07049 .009 .0428 .3755 2.00 -.0110 .05226 .976 -.1343 .1123 3.00 .1981 .08439 .051 -.0010 .3972 1.00 -.2091(*) .07049 .009 -.3755 -.0428 -.1981 .08439 .051 -.3972 .0010 1.00 2.00 3.00 Lt/At Std. Error 2.00 3.00 At Mean Difference (I-J) 3.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B2: ANOVA comparing payers, former payers and never paid firms (1992) 266 DBA Thesis: Malinee Ronapat Dependent Variable (I) PAYER3 (J) PAYER3 Et/At 1.00 2.00 3.00 .0713(*) .02231 .004 Lower Bound .0187 Upper Bound .1239 .0192 .02415 .705 -.0377 .0762 2.00 1.00 -.0713(*) .02231 .004 -.1239 -.0187 3.00 -.0521 .03206 .237 -.1277 .0235 1.00 2.00 -.0192 .02415 .705 -.0762 .0377 .0521 .03206 .237 -.0235 .1277 2.00 .4442 3.00 Yt/BEt 1.00 2.00 3.00 dAt/At Vt/At 95% Confidence Interval Sig. .2848(*) .06759 .000 .1255 3.00 .0838 .07317 .487 -.0887 .2563 1.00 3.00 1.00 -.2848(*) -.2010 .06759 .09712 .000 .098 -.4442 -.4300 -.1255 .0280 -.0838 .07317 .487 -.2563 .0887 2.00 .4300 .2010 .09712 .098 -.0280 2.00 .1937 .10814 .174 -.0612 .4487 -.6722(*) .11707 .000 -.9482 -.3962 2.00 3.00 1.00 -.1937 .10814 .174 -.4487 .0612 3.00 -.8660(*) .15540 .000 -1.2323 -.4996 3.00 1.00 .6722(*) .11707 .000 .3962 .9482 1.00 2.00 2.00 3.00 .8660(*) .0520 .15540 .34961 .000 .988 .4996 -.7722 1.2323 .8763 -.4654 .37850 .437 -1.3577 .4270 2.00 1.00 -.0520 .34961 .988 -.8763 .7722 3.00 -.5174 .50241 .559 -1.7019 .6671 1.00 2.00 .4654 .37850 .437 -.4270 1.3577 .5174 .50241 .559 -.6671 1.7019 2.00 4927.3893 16540.21785 .952 -34069.5252 43924.3039 3.00 10540.8237 17906.96329 .826 -31678.4702 52760.1175 1.00 3.00 1.00 -4927.3893 5613.4343 16540.21785 23769.02047 .952 .970 -43924.3039 -50426.8469 34069.5252 61653.7156 -10540.8237 17906.96329 .826 -52760.1175 31678.4702 1.00 2.00 3.00 Lt/At Std. Error 1.00 3.00 At Mean Difference (I-J) 2.00 -5613.4343 23769.02047 .970 -61653.7156 50426.8469 1.00 2.00 -.0822 .06214 .384 -.2287 .0643 -.0391 .06728 .830 -.1977 .1195 2.00 3.00 1.00 .0822 .06214 .384 -.0643 .2287 3.00 .0431 .08930 .880 -.1674 .2536 .0391 .06728 .830 -.1195 .1977 -.0431 .08930 .880 -.2536 .1674 3.00 1.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B3: ANOVA comparing payers, former payers and never paid firms (1993) 267 DBA Thesis: Malinee Ronapat Dependent Variable (I) PAYER3 (J) PAYER3 Et/At 1.00 2.00 3.00 .0928(*) .02221 .000 Lower Bound .0404 Upper Bound .1451 .0265 .02369 .503 -.0293 .0823 2.00 1.00 -.0928(*) .02221 .000 -.1451 -.0404 3.00 -.0663 .03113 .086 -.1396 .0071 1.00 2.00 -.0265 .02369 .503 -.0823 .0293 .0663 .03113 .086 -.0071 .1396 2.00 -.0781 .06140 .412 -.2228 .0666 3.00 .0590 .06548 .640 -.0953 .2133 1.00 3.00 1.00 .0781 .1371 .06140 .08607 .412 .250 -.0666 -.0657 .2228 .3399 -.0590 .06548 .640 -.2133 .0953 2.00 .0657 3.00 Yt/BEt 1.00 2.00 3.00 dAt/At Vt/At 95% Confidence Interval Sig. -.1371 .08607 .250 -.3399 2.00 .0230 .10216 .973 -.2177 .2637 -.6446(*) .10893 .000 -.9012 -.3879 2.00 3.00 1.00 -.0230 .10216 .973 -.2637 .2177 3.00 -.6675(*) .14319 .000 -1.0049 -.3301 3.00 1.00 .6446(*) .10893 .000 .3879 .9012 1.00 2.00 2.00 3.00 .6675(*) .7060 .14319 .57236 .000 .434 .3301 -.6425 1.0049 2.0545 -.9140 .61033 .294 -2.3520 .5240 2.00 1.00 -.7060 .57236 .434 -2.0545 .6425 3.00 -1.6201 .80229 .109 -3.5103 .2702 1.00 2.00 .9140 .61033 .294 -.5240 2.3520 1.6201 .80229 .109 -.2702 3.5103 2.00 7488.3241 14085.92710 .856 -25699.0841 40675.7323 3.00 9680.3420 15020.37202 .796 -25708.6824 45069.3664 1.00 3.00 1.00 -7488.3241 2192.0179 14085.92710 19744.40037 .856 .993 -40675.7323 -44327.1405 25699.0841 48711.1763 -9680.3420 15020.37202 .796 -45069.3664 25708.6824 1.00 2.00 3.00 Lt/At Std. Error 1.00 3.00 At Mean Difference (I-J) 2.00 -2192.0179 19744.40037 .993 -48711.1763 44327.1405 1.00 2.00 -.0732 .10171 .752 -.3129 .1664 .1001 .10845 .626 -.1554 .3557 2.00 3.00 1.00 .0732 .10171 .752 -.1664 .3129 3.00 .1733 .14256 .445 -.1625 .5092 1.00 -.1001 .10845 .626 -.3557 .1554 -.1733 .14256 .445 -.5092 .1625 3.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B4: ANOVA comparing payers, former payers and never paid firms (1994) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 268 DBA Thesis: Malinee Ronapat Variable Et/At 1.00 2.00 3.00 .0831(*) .01168 .000 Lower Bound .0556 .0179 .01366 .391 -.0143 .0500 2.00 1.00 -.0831(*) .01168 .000 -.1106 -.0556 3.00 -.0652(*) .01718 .001 -.1056 -.0247 1.00 2.00 -.0179 .01366 .391 -.0500 .0143 .0652(*) .01718 .001 .0247 .1056 2.00 3.00 Yt/BEt 1.00 2.00 3.00 dAt/At Vt/At .64580 .002 .7093 3.7504 .0759 .75524 .994 -1.7023 1.8541 1.00 3.00 1.00 -2.2299(*) -2.1540 .64580 .95023 .002 .062 -3.7504 -4.3913 -.7093 .0834 -.0759 .75524 .994 -1.8541 1.7023 2.00 4.3913 2.1540 .95023 .062 -.0834 2.00 .1899 .08889 .084 -.0194 .3992 -.6721(*) .10396 .000 -.9168 -.4273 2.00 3.00 1.00 -.1899 .08889 .084 -.3992 .0194 3.00 -.8620(*) .13080 .000 -1.1700 -.5540 3.00 1.00 .6721(*) .10396 .000 .4273 .9168 1.00 2.00 2.00 3.00 .8620(*) .1677 .13080 .19769 .000 .673 .5540 -.2978 1.1700 .6332 -.6248(*) .23119 .020 -1.1692 -.0805 2.00 1.00 -.1677 .19769 .673 -.6332 .2978 3.00 -.7925(*) .29088 .019 -1.4774 -.1077 1.00 2.00 .6248(*) .23119 .020 .0805 1.1692 .7925(*) .29088 .019 .1077 1.4774 2.00 1.00 14544.8699 13583.15096 .533 -17436.6886 46526.4284 3.00 8831.2903 15885.01192 .843 -28570.0048 46232.5855 1.00 3.00 1.00 -14544.8699 -5713.5796 13583.15096 19986.31626 .533 .956 -46526.4284 -52771.4047 17436.6886 41344.2455 -8831.2903 15885.01192 .843 -46232.5855 28570.0048 2.00 5713.5796 19986.31626 .956 -41344.2455 52771.4047 1.00 2.00 -.1092 .04641 .050 -.2185 .0001 .0856 .05428 .257 -.0422 .2135 2.00 3.00 1.00 .1092 .04641 .050 -.0001 .2185 3.00 .1949(*) .06829 .013 .0341 .3557 -.0856 .05428 .257 -.2135 .0422 -.1949(*) .06829 .013 -.3557 -.0341 2.00 3.00 Lt/At 2.2299(*) 3.00 1.00 3.00 At Upper Bound .1106 3.00 1.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B5: ANOVA comparing payers, former payers and never paid firms (1995) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 269 DBA Thesis: Malinee Ronapat Variable Et/At Yt/BEt .0692(*) .00929 .000 .0607(*) .01846 .003 .0172 .1041 2.00 1.00 -.0692(*) .00929 .000 -.0911 -.0474 3.00 -.0086 .02011 .905 -.0559 .0388 3.00 1.00 2.00 -.0607(*) .01846 .003 -.1041 -.0172 .0086 .02011 .905 -.0388 .0559 1.00 2.00 .0524 .04729 .509 -.0589 .1638 3.00 .2345(*) .09402 .035 .0131 .4558 1.00 3.00 1.00 -.0524 .1820 .04729 .10240 .509 .179 -.1638 -.0590 .0589 .4231 -.2345(*) .09402 .035 -.4558 -.0131 2.00 -.1820 .10240 .179 -.4231 .0590 1.00 2.00 .1536(*) .05909 .026 .0145 .2927 -.2630 .11747 .066 -.5395 .0135 2.00 3.00 1.00 -.1536(*) .05909 .026 -.2927 -.0145 3.00 -.4166(*) .12795 .004 -.7178 -.1154 3.00 1.00 .2630 .11747 .066 -.0135 .5395 1.00 2.00 2.00 3.00 .4166(*) .1849 .12795 .11991 .004 .273 .1154 -.0974 .7178 .4672 -.6646(*) .23839 .015 -1.2258 -.1035 2.00 1.00 -.1849 .11991 .273 -.4672 .0974 3.00 -.8495(*) .25965 .003 -1.4607 -.2383 1.00 2.00 .6646(*) .23839 .015 .1035 1.2258 .8495(*) .25965 .003 .2383 1.4607 2.00 3.00 Vt/At 3.00 At 1.00 16523.7653 13210.49180 .424 -14573.6465 47621.1771 3.00 -616.3515 26262.29757 1.000 -62437.6292 61204.9262 1.00 3.00 1.00 -16523.7653 -17140.1168 13210.49180 28604.97886 .424 .821 -47621.1771 -84476.0510 14573.6465 50195.8174 616.3515 26262.29757 1.000 -61204.9262 62437.6292 2.00 17140.1168 28604.97886 .821 -50195.8174 84476.0510 1.00 2.00 -.0942(*) .03410 .017 -.1745 -.0140 -.0588 .06780 .661 -.2184 .1008 2.00 3.00 1.00 .0942(*) .03410 .017 .0140 .1745 3.00 .0354 .07384 .881 -.1384 .2092 .0588 .06780 .661 -.1008 .2184 -.0354 .07384 .881 -.2092 .1384 2.00 3.00 Lt/At Upper Bound .0911 2.00 3.00 2.00 dAt/At Lower Bound .0474 1.00 3.00 1.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B6: ANOVA comparing payers, former payers and never paid firms (1996) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 270 DBA Thesis: Malinee Ronapat Variable Et/At Yt/BEt .0745(*) .01122 .000 .0645(*) .01909 .002 .0196 .1094 2.00 1.00 -.0745(*) .01122 .000 -.1009 -.0481 3.00 -.0100 .02126 .885 -.0600 .0400 3.00 1.00 2.00 -.0645(*) .01909 .002 -.1094 -.0196 .0100 .02126 .885 -.0400 .0600 1.00 2.00 .3148 3.00 Vt/At .04467 .000 .1046 -.0555 .07601 .746 -.2343 .1234 1.00 3.00 1.00 -.2097(*) -.2652(*) .04467 .08464 .000 .005 -.3148 -.4643 -.1046 -.0660 .0555 .07601 .746 -.1234 .2343 2.00 .2652(*) .08464 .005 .0660 .4643 2.00 .1569 .12541 .424 -.1382 .4520 -.3949 .21337 .155 -.8970 .1072 2.00 3.00 1.00 -.1569 .12541 .424 -.4520 .1382 3.00 -.5518 .23762 .054 -1.1109 .0073 3.00 1.00 .3949 .21337 .155 -.1072 .8970 1.00 2.00 2.00 3.00 .5518 .2074 .23762 .10724 .054 .131 -.0073 -.0449 1.1109 .4598 -.5056(*) .18246 .016 -.9349 -.0762 2.00 1.00 -.2074 .10724 .131 -.4598 .0449 3.00 -.7130(*) .20319 .001 -1.1911 -.2349 1.00 2.00 .5056(*) .18246 .016 .0762 .9349 .7130(*) .20319 .001 .2349 1.1911 2.00 1.00 18742.6503 12930.55809 .317 -11683.6920 49168.9926 3.00 4781.8852 21999.95609 .974 -46985.2690 56549.0394 1.00 3.00 1.00 -18742.6503 -13960.7651 12930.55809 24500.00481 .317 .836 -49168.9926 -71610.6768 11683.6920 43689.1466 -4781.8852 21999.95609 .974 -56549.0394 46985.2690 2.00 13960.7651 24500.00481 .836 -43689.1466 71610.6768 1.00 2.00 -.0557 .05238 .538 -.1789 .0676 .0065 .08911 .997 -.2032 .2162 2.00 3.00 1.00 .0557 .05238 .538 -.0676 .1789 3.00 .0622 .09924 .806 -.1714 .2957 1.00 -.0065 .08911 .997 -.2162 .2032 -.0622 .09924 .806 -.2957 .1714 2.00 3.00 Lt/At .2097(*) 3.00 1.00 3.00 At Upper Bound .1009 2.00 3.00 2.00 dAt/At Lower Bound .0481 1.00 3.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B7: ANOVA comparing payers, former payers and never paid firms (1997) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 271 DBA Thesis: Malinee Ronapat Variable Et/At Yt/BEt .1355(*) .03199 .000 .1721(*) .05847 .010 .0345 .3096 2.00 1.00 -.1355(*) .03199 .000 -.2108 -.0603 3.00 .0365 .06282 .830 -.1113 .1843 3.00 1.00 2.00 -.1721(*) .05847 .010 -.3096 -.0345 -.0365 .06282 .830 -.1843 .1113 1.00 2.00 2.3217 3.95331 .827 -6.9792 11.6226 3.00 -1.8509 7.22612 .964 -18.8518 15.1499 1.00 3.00 1.00 -2.3217 -4.1726 3.95331 7.76291 .827 .853 -11.6226 -22.4364 6.9792 14.0911 1.8509 7.22612 .964 -15.1499 18.8518 2.00 -14.0911 22.4364 3.00 Vt/At 4.1726 7.76291 .853 1.00 2.00 .2204(*) .06005 .001 .0791 .3617 -.6573(*) .10977 .000 -.9155 -.3990 2.00 3.00 1.00 -.2204(*) .06005 .001 -.3617 -.0791 3.00 -.8777(*) .11793 .000 -1.1551 -.6002 3.00 1.00 .6573(*) .10977 .000 .3990 .9155 1.00 2.00 2.00 3.00 .8777(*) .0246 .11793 .06556 .000 .925 .6002 -.1296 1.1551 .1788 .0601 .11983 .871 -.2218 .3420 2.00 1.00 -.0246 .06556 .925 -.1788 .1296 3.00 .0355 .12873 .959 -.2673 .3384 1.00 2.00 -.0601 .11983 .871 -.3420 .2218 -.0355 .12873 .959 -.3384 .2673 2.00 3.00 At 1.00 18465.3199 12625.50595 .310 -11238.6592 48169.2990 3.00 7630.1860 23077.72876 .942 -46664.6973 61925.0694 1.00 3.00 1.00 -18465.3199 -10835.1338 12625.50595 24792.06301 .310 .900 -48169.2990 -69163.3247 11238.6592 47493.0571 -7630.1860 23077.72876 .942 -61925.0694 46664.6973 2.00 10835.1338 24792.06301 .900 -47493.0571 69163.3247 1.00 2.00 -.1963(*) .04155 .000 -.2940 -.0985 -.1913(*) .07595 .033 -.3700 -.0126 2.00 3.00 1.00 .1963(*) .04155 .000 .0985 .2940 3.00 .0050 .08159 .998 -.1870 .1969 1.00 .1913(*) .07595 .033 .0126 .3700 -.0050 .08159 .998 -.1969 .1870 2.00 3.00 Lt/At Upper Bound .2108 2.00 3.00 2.00 dAt/At Lower Bound .0603 1.00 3.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B8: ANOVA comparing payers, former payers and never paid firms (1998) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 272 DBA Thesis: Malinee Ronapat Variable Et/At 1.00 2.00 3.00 .0544 .06830 .705 Lower Bound -.1063 Upper Bound .2151 .0617 .13506 .891 -.2561 .3795 2.00 1.00 -.0544 .06830 .705 -.2151 .1063 3.00 .0073 .12814 .998 -.2942 .3087 1.00 2.00 -.0617 .13506 .891 -.3795 .2561 -.0073 .12814 .998 -.3087 .2942 2.00 3.00 Yt/BEt 1.00 10.8456 15.04379 .751 -24.5496 46.2408 3.00 .8539 29.74597 1.000 -69.1328 70.8406 1.00 3.00 1.00 -10.8456 -9.9917 15.04379 28.22295 .751 .933 -46.2408 -76.3950 24.5496 56.4116 -.8539 29.74597 1.000 -70.8406 69.1328 2.00 9.9917 28.22295 .933 -56.4116 76.3950 1.00 2.00 1.3075 1.49124 .655 -2.2011 4.8161 -.1511 2.94862 .999 -7.0887 6.7864 2.00 3.00 1.00 -1.3075 1.49124 .655 -4.8161 2.2011 3.00 -1.4587 2.79765 .861 -8.0410 5.1237 3.00 1.00 .1511 2.94862 .999 -6.7864 7.0887 1.00 2.00 2.00 3.00 1.4587 -1.3278 2.79765 1.45860 .861 .634 -5.1237 -4.7596 8.0410 2.1040 .0474 2.88409 1.000 -6.7384 6.8331 2.00 1.00 1.3278 1.45860 .634 -2.1040 4.7596 3.00 1.3751 2.73642 .870 -5.0631 7.8134 1.00 2.00 -.0474 2.88409 1.000 -6.8331 6.7384 -1.3751 2.73642 .870 -7.8134 5.0631 2.00 66551.4820 2.00 3.00 dAt/At Vt/At 3.00 At 1.00 39057.1241(*) 11685.73697 .003 11562.7662 3.00 29663.9558 23106.11544 .405 -24700.4209 84028.3325 1.00 3.00 1.00 -39057.1241(*) -9393.1683 11685.73697 21923.05977 .003 .904 -66551.4820 -60974.0358 -11562.7662 42187.6992 -29663.9558 23106.11544 .405 -84028.3325 24700.4209 2.00 9393.1683 21923.05977 .904 -42187.6992 60974.0358 1.00 2.00 -1.4221 1.22204 .476 -4.2973 1.4532 -.3850 2.41634 .986 -6.0702 5.3002 2.00 3.00 1.00 1.4221 1.22204 .476 -1.4532 4.2973 3.00 1.0371 2.29262 .893 -4.3570 6.4312 .3850 2.41634 .986 -5.3002 6.0702 -1.0371 2.29262 .893 -6.4312 4.3570 2.00 3.00 Lt/At 3.00 1.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B9: ANOVA comparing payers, former payers and never paid firms (1999) Dependent Variable (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 273 DBA Thesis: Malinee Ronapat Et/At 1.00 2.00 3.00 .1746(*) .04164 .000 Lower Bound .0766 .1537 .07748 .118 -.0287 .3360 2.00 1.00 -.1746(*) .04164 .000 -.2726 -.0766 3.00 -.0209 .07340 .956 -.1936 .1518 1.00 2.00 -.1537 .07748 .118 -.3360 .0287 .0209 .07340 .956 -.1518 .1936 2.00 .7654 .65140 .469 -.7672 2.2981 3.00 3.6542(*) 1.21202 .008 .8025 6.5060 1.00 3.00 1.00 -.7654 2.8888(*) .65140 1.14806 .469 .033 -2.2981 .1875 .7672 5.5901 -3.6542(*) 1.21202 .008 -6.5060 -.8025 2.00 -2.8888(*) 1.14806 .033 -5.5901 -.1875 1.00 2.00 .1855(*) .05495 .002 .0562 .3148 -.0615 .10224 .819 -.3021 .1790 2.00 3.00 1.00 -.1855(*) .05495 .002 -.3148 -.0562 3.00 -.2470(*) .09684 .030 -.4749 -.0192 3.00 1.00 .0615 .10224 .819 -.1790 .3021 1.00 2.00 2.00 3.00 .2470(*) -.3729(*) .09684 .14722 .030 .031 .0192 -.7193 .4749 -.0265 -.2673 .27393 .593 -.9118 .3773 2.00 1.00 .3729(*) .14722 .031 .0265 .7193 3.00 .1056 .25948 .913 -.5049 .7162 1.00 2.00 .2673 .27393 .593 -.3773 .9118 -.1056 .25948 .913 -.7162 .5049 2.00 -26395.4496 11327.64108 .053 -53048.0880 257.1888 3.00 -23055.2743 21076.67301 .518 -72646.2651 26535.7165 1.00 3.00 1.00 26395.4496 3340.1753 11327.64108 19964.44118 .053 .985 -257.1888 -43633.8617 53048.0880 50314.2124 23055.2743 21076.67301 .518 -26535.7165 72646.2651 2.00 -3340.1753 19964.44118 .985 -50314.2124 43633.8617 1.00 2.00 -.6862(*) .11174 .000 -.9492 -.4233 -.5028(*) .20791 .042 -.9920 -.0136 2.00 3.00 1.00 .6862(*) .11174 .000 .4233 .9492 3.00 .1835 .19694 .621 -.2799 .6468 1.00 .5028(*) .20791 .042 .0136 .9920 -.1835 .19694 .621 -.6468 .2799 3.00 Yt/BEt 1.00 2.00 3.00 dAt/At Vt/At 3.00 At 1.00 2.00 3.00 Lt/At 3.00 2.00 * The mean difference is significant at the .05 level. Upper Bound .2726 Source: Developed for this research Table B10: ANOVA comparing payers, former payers and never paid firms (2000) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 274 DBA Thesis: Malinee Ronapat Variable Et/At 1.00 2.00 3.00 .1994(*) .07603 .025 Lower Bound .0205 .1425 .13580 .546 -.1771 .4621 2.00 1.00 -.1994(*) .07603 .025 -.3783 -.0205 3.00 -.0569 .13085 .901 -.3649 .2510 1.00 2.00 -.1425 .13580 .546 -.4621 .1771 .0569 .13085 .901 -.2510 .3649 2.00 3.00 Yt/BEt 1.00 2.00 3.00 dAt/At Vt/At .82153 .535 -1.0567 2.8099 .5571 1.46731 .924 -2.8960 4.0101 1.00 3.00 1.00 -.8766 -.3195 .82153 1.41379 .535 .972 -2.8099 -3.6466 1.0567 3.0076 -.5571 1.46731 .924 -4.0101 2.8960 2.00 .3195 1.41379 .972 -3.0076 3.6466 2.00 .4208 .18676 .064 -.0187 .8603 -.1050 .33357 .947 -.8900 .6800 2.00 3.00 1.00 -.4208 .18676 .064 -.8603 .0187 3.00 -.5258 .32140 .232 -1.2822 .2305 3.00 1.00 .1050 .33357 .947 -.6800 .8900 1.00 2.00 2.00 3.00 .5258 -.6652(*) .32140 .24557 .232 .019 -.2305 -1.2431 1.2822 -.0873 -.2972 .43862 .777 -1.3294 .7350 2.00 1.00 .6652(*) .24557 .019 .0873 1.2431 3.00 .3680 .42262 .659 -.6265 1.3626 1.00 2.00 .2972 .43862 .777 -.7350 1.3294 -.3680 .42262 .659 -1.3626 .6265 2.00 1.00 -29499.1726(*) 11720.66169 .033 -57081.5743 -1916.7709 3.00 -19859.4418 20934.10720 .610 -69123.9774 29405.0938 1.00 3.00 1.00 29499.1726(*) 9639.7308 11720.66169 20170.49182 .033 .882 1916.7709 -37827.7777 57081.5743 57107.2393 19859.4418 20934.10720 .610 -29405.0938 69123.9774 2.00 -9639.7308 20170.49182 .882 -57107.2393 37827.7777 1.00 2.00 -1.0015(*) .24529 .000 -1.5788 -.4243 -.6015 .43811 .356 -1.6325 .4295 2.00 3.00 1.00 1.0015(*) .24529 .000 .4243 1.5788 3.00 .4000 .42213 .610 -.5934 1.3934 .6015 .43811 .356 -.4295 1.6325 -.4000 .42213 .610 -1.3934 .5934 2.00 3.00 Lt/At .8766 3.00 1.00 3.00 At Upper Bound .3783 3.00 1.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B11: ANOVA comparing payers, former payers and never paid firms (2001) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 275 DBA Thesis: Malinee Ronapat Variable Et/At 1.00 2.00 3.00 -.1045 .14977 .765 Lower Bound -.4569 Upper Bound .2479 -.0600 .24279 .967 -.6313 .5113 2.00 1.00 .1045 .14977 .765 -.2479 .4569 3.00 .0445 .23863 .981 -.5170 .6060 1.00 2.00 .0600 .24279 .967 -.5113 .6313 -.0445 .23863 .981 -.6060 .5170 2.00 3.00 Yt/BEt 1.00 5.5240 4.75679 .477 -5.6692 16.7173 3.00 .2246 7.71115 1.000 -17.9206 18.3698 1.00 3.00 1.00 -5.5240 -5.2994 4.75679 7.57902 .477 .764 -16.7173 -23.1337 5.6692 12.5348 -.2246 7.71115 1.000 -18.3698 17.9206 2.00 5.2994 7.57902 .764 -12.5348 23.1337 1.00 2.00 2.6670 2.64710 .573 -3.5619 8.8959 -.2791 4.29117 .998 -10.3767 9.8185 2.00 3.00 1.00 -2.6670 2.64710 .573 -8.8959 3.5619 3.00 -2.9461 4.21764 .764 -12.8707 6.9784 3.00 1.00 .2791 4.29117 .998 -9.8185 10.3767 1.00 2.00 2.00 3.00 2.9461 -1.1641(*) 4.21764 .43087 .764 .020 -6.9784 -2.1780 12.8707 -.1502 -.2763 .69848 .917 -1.9199 1.3673 2.00 1.00 1.1641(*) .43087 .020 .1502 2.1780 3.00 .8878 .68651 .400 -.7277 2.5032 1.00 2.00 .2763 .69848 .917 -1.3673 1.9199 -.8878 .68651 .400 -2.5032 .7277 2.00 2.00 3.00 dAt/At Vt/At 3.00 At 1.00 -29609.6303(*) 11260.01861 .024 -56105.6645 -3113.5961 3.00 -12993.4564 18253.39590 .757 -55945.6591 29958.7463 1.00 3.00 1.00 29609.6303(*) 16616.1739 11260.01861 17940.62716 .024 .624 3113.5961 -25600.0504 56105.6645 58832.3982 12993.4564 18253.39590 .757 -29958.7463 55945.6591 2.00 -16616.1739 17940.62716 .624 -58832.3982 25600.0504 1.00 2.00 -1.4975(*) .42403 .001 -2.4952 -.4997 -.5320 .68739 .719 -2.1495 1.0855 2.00 3.00 1.00 1.4975(*) .42403 .001 .4997 2.4952 3.00 .9654 .67561 .327 -.6244 2.5552 .5320 .68739 .719 -1.0855 2.1495 -.9654 .67561 .327 -2.5552 .6244 2.00 3.00 Lt/At 3.00 1.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research Table B12: ANOVA comparing payers, former payers and never paid firms (2002) Dependent (I) PAYER3 (J) PAYER3 Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval 276 DBA Thesis: Malinee Ronapat Variable Et/At 1.00 2.00 3.00 .0690(*) .02082 .003 Lower Bound .0200 .0629 .02989 .090 -.0074 .1332 2.00 1.00 -.0690(*) .02082 .003 -.1180 -.0200 3.00 -.0061 .03041 .978 -.0776 .0654 1.00 2.00 -.0629 .02989 .090 -.1332 .0074 .0061 .03041 .978 -.0654 .0776 2.00 3.00 Yt/BEt 1.00 2.00 3.00 dAt/At Vt/At .08329 .567 -.1114 .2806 .1898 .11960 .253 -.0916 .4711 1.00 3.00 1.00 -.0846 .1051 .08329 .12166 .567 .663 -.2806 -.1811 .1114 .3914 -.1898 .11960 .253 -.4711 .0916 2.00 -.1051 .12166 .663 -.3914 .1811 2.00 .2743 .12930 .087 -.0300 .5785 -.2454 .18565 .384 -.6822 .1914 2.00 3.00 1.00 -.2743 .12930 .087 -.5785 .0300 3.00 -.5197(*) .18886 .017 -.9640 -.0753 3.00 1.00 .2454 .18565 .384 -.1914 .6822 1.00 2.00 2.00 3.00 .5197(*) -.4561(*) .18886 .17479 .017 .025 .0753 -.8673 .9640 -.0448 -.3668 .25096 .311 -.9573 .2237 2.00 1.00 .4561(*) .17479 .025 .0448 .8673 3.00 .0893 .25530 .935 -.5114 .6899 1.00 2.00 .3668 .25096 .311 -.2237 .9573 -.0893 .25530 .935 -.6899 .5114 2.00 1.00 -33995.5603(*) 11694.41946 .011 -61510.9047 -6480.2158 3.00 -3048.9074 16790.96286 .982 -42555.7105 36457.8957 1.00 3.00 1.00 33995.5603(*) 30946.6529 11694.41946 17081.08486 .011 .167 6480.2158 -9242.7670 61510.9047 71136.0727 3048.9074 16790.96286 .982 -36457.8957 42555.7105 2.00 -30946.6529 17081.08486 .167 -71136.0727 9242.7670 1.00 2.00 -.8297(*) .15579 .000 -1.1962 -.4631 -.3762 .22368 .214 -.9025 .1501 2.00 3.00 1.00 .8297(*) .15579 .000 .4631 1.1962 3.00 .4535 .22754 .115 -.0819 .9889 .3762 .22368 .214 -.1501 .9025 -.4535 .22754 .115 -.9889 .0819 2.00 3.00 Lt/At .0846 3.00 1.00 3.00 At Upper Bound .1180 3.00 1.00 2.00 * The mean difference is significant at the .05 level. Source: Developed for this research 277 DBA Thesis: Malinee Ronapat APPENDIX C: LOGIT REGRESSION: CONTROLLING TWO INVESTMENT VARIABLES Table C1: Logit Regression: controlling two investment variables from 1991 to 2002 Logit Regression 1991 278 DBA Thesis: Malinee Ronapat Variables in the Equation B S.E. Wald df Sig. SET 1.69166191 0.7746301 4.76912164 1 0.0289746 PROFIT1 -3.0421817 3.00551323 1.02454964 1 0.31144225 INVEST1 0.30065078 0.25692757 1.36931467 1 0.24192967 INVEST2 -0.04984 0.0217351 5.25815479 1 0.02184417 Constant 1.52435137 0.50680503 9.04666144 1 0.00263175 a Variable(s) entered on step 1: SET, PROFIT1, INVEST1, INVEST2. Logit Regression 1992 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 0.25025485 0.83675582 0.08944736 1 0.7648809 PROFIT 14.63854 4.07803407 12.8852791 1 0.00033118 INVEST1 -1.3425887 0.6029451 4.95827302 1 0.02596613 INVEST2 -0.5921966 0.22108188 7.17506076 1 0.00739239 Constant 2.61120472 0.54400168 23.0399406 1 1.5867E-06 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 1993 Variables in the Equation B S.E. Wald df Sig. Step 1 SET -0.0227352 0.76644939 0.0008799 1 0.97633578 PROFIT 15.0560417 3.09832066 23.6139637 1 1.1773E-06 INVEST1 -2.8461384 0.55088734 26.6923268 1 2.3857E-07 INVEST2 -0.0725844 0.12960446 0.31365077 1 0.57544852 Constant 1.8909287 0.40828848 21.4494444 1 3.6328E-06 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 1994 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 1.09242823 0.66000241 2.73964815 1 0.09788632 PROFIT 20.8225252 3.64845666 32.5722989 1 1.1484E-08 INVEST1 -1.7541047 0.5009255 12.2620971 1 0.00046225 INVEST2 -0.4561003 0.19003079 5.76066644 1 0.01638885 Constant 1.11172585 0.37305171 8.88090724 1 0.00288168 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 1995 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 0.83479723 0.69185978 1.45588087 1 0.22758666 PROFIT 24.1569544 3.92324996 37.9133711 1 7.3957E-10 INVEST1 0.30063641 0.39517304 0.57877338 1 0.44679352 INVEST2 -0.6422795 0.26976171 5.66875274 1 0.01726973 Constant 0.66260573 0.38439776 2.97131488 1 0.08475294 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 1996 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 1.36203933 0.58650508 5.39306611 1 0.02021692 PROFIT 15.6212727 2.87734407 29.4746892 1 5.6652E-08 INVEST1 -0.1821251 0.38740366 0.22101011 1 0.63827134 INVEST2 -0.3380858 0.21839834 2.39637727 1 0.1216166 Constant 0.48208197 0.30412863 2.51262204 1 0.11293788 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 1997 Variables in the Equation Step 1 Exp(B) 5.428494873 0.047730643 1.350737555 0.951381603 4.592164004 Exp(B) 1.284352689 2277388.687 0.261168705 0.553110961 13.6154438 Exp(B) 0.977521272 3457449.26 0.058068126 0.92998728 6.62551893 Exp(B) 2.981505072 1104352741 0.173062112 0.633750283 3.039599764 Exp(B) 2.304346759 30990743365 1.350718145 0.526091831 1.939840452 Exp(B) 3.904147041 6084607.207 0.833497066 0.7131341 1.619442522 279 DBA Thesis: Malinee Ronapat B S.E. Wald df Sig. SET 2.01057411 0.48575242 17.1320849 1 3.4868E-05 PROFIT 1.78983085 0.49581226 13.0313512 1 0.00030632 INVEST1 -0.1917793 0.23938561 0.64181115 1 0.42305571 INVEST2 -0.0762184 0.28028488 0.07394693 1 0.78567442 Constant 0.29174874 0.28714911 1.03229308 1 0.3096207 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 1998 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 2.68757859 0.52590526 26.1160395 1 3.215E-07 PROFIT 1.63713102 0.86163954 3.61007008 1 0.0574307 INVEST1 0.26203625 0.42856851 0.37383696 1 0.54092019 INVEST2 -0.2521342 0.26247227 0.92277658 1 0.33674699 Constant -2.2458309 0.35074476 40.9988539 1 1.5232E-10 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 1999 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 1.24191534 0.50070847 6.15196859 1 0.01312659 PROFIT 4.30982748 1.11488618 14.9437195 1 0.00011077 INVEST1 0.52716789 0.447829 1.38571356 1 0.23913031 INVEST2 -0.7253766 0.26091623 7.72902863 1 0.005434 Constant -0.9079683 0.32811589 7.65749816 1 0.00565369 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 2000 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 1.73732713 0.48178466 13.0034102 1 0.00031092 PROFIT 5.5765423 1.32574359 17.693385 1 2.5953E-05 INVEST1 0.75236727 0.34248438 4.82589923 1 0.02803524 INVEST2 -1.1656548 0.3536621 10.8633274 1 0.00098087 Constant -0.6780377 0.35922275 3.56270578 1 0.05909134 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 2001 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 1.72590489 0.4093318 17.7779934 1 2.4824E-05 PROFIT -0.0555633 0.09701492 0.3280191 1 0.56682785 INVEST1 -0.0046421 0.02280901 0.04142055 1 0.83872831 INVEST2 -0.5444767 0.21879821 6.19257247 1 0.01282876 Constant -0.656846 0.30524045 4.63066009 1 0.03140548 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Logit Regression 2002 Variables in the Equation B S.E. Wald df Sig. Step 1 SET 1.70082859 0.42250429 16.2053664 1 5.6833E-05 PROFIT 5.61125629 1.33189148 17.7493532 1 2.52E-05 INVEST1 -0.2860076 0.17600596 2.6405864 1 0.10416509 INVEST2 -0.7700316 0.2105792 13.3716877 1 0.00025545 Constant -0.4395467 0.28131783 2.44126852 1 0.11818062 a Variable(s) entered on step 1: SET, PROFIT, INVEST1, INVEST2. Step 1 Exp(B) 7.467603327 5.988439457 0.825489055 0.926613824 1.338766591 Exp(B) 14.69604772 5.140400618 1.299573648 0.777140461 0.105839563 Exp(B) 3.462238484 74.42764751 1.694127546 0.484142192 0.403342844 Exp(B) 5.682135472 264.156651 2.122017457 0.311718487 0.507612127 Exp(B) 5.617602064 0.945952137 0.99536866 0.580145289 0.518484077 Exp(B) 5.47848493 273.4876026 0.751256907 0.462998415 0.644328417 Source: Developed for this research Note: PROFIT represents Et/At ratio. INVEST1 represents dAt/At ratio and INVEST2 represents Vt/At. 280 DBA Thesis: Malinee Ronapat APPENDIX D: LOGIT REGRESSION: CONTROLLING ONE INVESTMENT VARIABLE Table D1: Logit Regression: controlling one investment variable Logit regression 1991 281 DBA Thesis: Malinee Ronapat Variables in the Equation B S.E. Wald SET 0.909227133 0.67056316 1.838507451 PROFIT -2.903282463 2.927846101 0.983291063 INVEST1 0.341714711 0.24907159 1.882257189 Constant 1.617157642 0.501465958 10.39972373 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 1992 Variables in the Equation B S.E. Wald Step 1 SET 0.051479911 0.808487973 0.004054417 PROFIT 11.43072322 3.710932953 9.488142579 INVEST1 -1.510840948 0.584779098 6.675039064 Constant 1.893909231 0.507278446 13.93880371 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 1993 Variables in the Equation B S.E. Wald Step 1 SET -0.195398328 0.69583429 0.078855145 PROFIT 15.01121885 3.101116518 23.43126549 INVEST1 -2.926270038 0.534767723 29.9432149 Constant 1.846103657 0.40176739 21.11362474 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 1994 Variables in the Equation B S.E. Wald Step 1 SET 0.421364264 0.580402523 0.527056068 PROFIT 19.90443372 3.583042448 30.85998753 INVEST1 -2.04817657 0.49123395 17.3843333 Constant 0.828505261 0.357005393 5.385685487 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 1995 Variables in the Equation B S.E. Wald Step 1 SET 0.113335512 0.592852892 0.036545856 PROFIT 22.39143171 3.755078097 35.55705463 INVEST1 0.315013545 0.402183652 0.613493023 Constant 0.252276658 0.355153006 0.504572 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 1996 Variables in the Equation B S.E. Wald Step 1 SET 1.012963284 0.528232751 3.677365182 PROFIT 15.37261702 2.856105429 28.96991001 INVEST1 -0.207842807 0.39772938 0.273082991 Constant 0.283130654 0.278999929 1.029830142 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 1997 Variables in the Equation B S.E. Wald Step 1 SET 1.956362569 0.441452572 19.63950727 PROFIT 1.834539726 0.471427319 15.14344723 INVEST1 -0.189181745 0.238378865 0.629829439 Constant 0.244061097 0.228288642 1.142953358 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Step 1 df 1 1 1 1 Sig. 0.17512598 0.32138764 0.17007783 0.00126034 Exp(B) 2.482403223 0.054842904 1.407358736 5.038748017 df 1 1 1 1 Sig. 0.94922956 0.00206804 0.00977724 0.00018886 Exp(B) 1.052828036 92108.56532 0.220724282 6.645295967 df 1 1 1 1 Sig. 0.77885492 1.2946E-06 4.4489E-08 4.3284E-06 Exp(B) 0.822506966 3305898.486 0.053596579 6.335087699 df 1 1 1 1 Sig. 0.46784719 2.7733E-08 3.0533E-05 0.02030261 Exp(B) 1.52403933 440946326.1 0.128969857 2.289893388 df 1 1 1 1 Sig. 0.84839264 2.4769E-09 0.43347565 0.47749808 Exp(B) 1.120007646 5302433516 1.370277871 1.286952033 df 1 1 1 1 Sig. 0.05515595 7.3511E-08 0.60127113 0.31019857 Exp(B) 2.753749077 4745071.331 0.81233472 1.327278565 df 1 1 1 1 Sig. 9.3515E-06 9.9644E-05 0.42741788 0.28502913 Exp(B) 7.073550661 6.262251128 0.827636075 1.276422314 282 DBA Thesis: Malinee Ronapat Logit regression 1998 Variables in the Equation B S.E. Wald SET 2.490787506 0.481848774 26.72094666 PROFIT 1.567042419 0.717224534 4.773657229 INVEST1 0.438409329 0.211117835 4.312307558 Constant -2.365617422 0.308193704 58.91711793 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 1999 Variables in the Equation B S.E. Wald Step 1 SET 0.750371887 0.453141114 2.74211821 PROFIT 3.974769365 1.008424878 15.53591247 INVEST1 0.518463167 0.429289108 1.458599559 Constant -1.367565114 0.2777623 24.24093846 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 2000 Variables in the Equation B S.E. Wald Step 1 SET 1.324993571 0.439054176 9.107335127 PROFIT 5.060612144 1.187881476 18.14930031 INVEST1 0.80011661 0.334073122 5.736189623 Constant -1.487770427 0.270983376 30.14300498 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 2001 Variables in the Equation B S.E. Wald Step 1 SET 1.761378872 0.389166374 20.48495192 PROFIT -0.069944305 0.089979679 0.604248864 INVEST1 0.009309897 0.016001202 0.338520153 Constant -1.271783961 0.230614679 30.41254689 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Logit regression 2002 Variables in the Equation B S.E. Wald Step 1 SET 1.51588267 0.384204861 15.56701939 PROFIT 2.202351906 0.85389976 6.65211343 INVEST1 0.082779556 0.102829378 0.648054844 Constant -1.045317985 0.224369269 21.70551518 a Variable(s) entered on step 1: SET, PROFIT, INVEST1. Step 1 df 1 1 1 1 Sig. 2.3506E-07 0.02889837 0.0378376 1.6446E-14 Exp(B) 12.07077819 4.792453142 1.550239335 0.093891312 df 1 1 1 1 Sig. 0.09773514 8.0953E-05 0.22715308 8.5006E-07 Exp(B) 2.117787449 53.23783699 1.67944464 0.254726435 df 1 1 1 1 Sig. 0.00254586 2.0424E-05 0.01661881 4.0133E-08 Exp(B) 3.762161169 157.6870139 2.225800463 0.225875701 df 1 1 1 1 Sig. 6.0102E-06 0.43696147 0.56068467 3.4926E-08 Exp(B) 5.820457532 0.932445751 1.009353369 0.280331076 df 1 1 1 1 Sig. 7.9632E-05 0.00990383 0.42080898 3.1788E-06 Exp(B) 4.553438537 9.046264467 1.086302313 0.351580004 Source: Developed for this research Note: PROFIT represents Et/At ratio. INVEST1 represents dAt/At ratio. 283 DBA Thesis: Malinee Ronapat APPENDIX E: T-STATISTICS (TWO INVESTMENT VARIABLES) Table E1: t-statistics for 1991-1996, 1997-2002, 1991-2002 (two investment variables) 284 DBA Thesis: Malinee Ronapat t-stat results 1991-1996 One-Sample Test Test Value = 0 CONSTANT SET PROFIT INVEST1 INVEST2 t df 4.2243059 3.2457428 3.7868216 -3.4593826 -1.7689473 5 5 5 5 5 Sig. (2-tailed) 0.008293448 0.022803365 0.01280004 0.018054303 0.137133446 Mean Difference 1.378333333 0.866666667 14.54333333 -0.358333333 -0.92 95% Confidence Interval of the Difference Lower Upper 0.539587586 2.21707908 0.180279063 1.55305427 4.670983385 24.41568328 -0.624601993 -0.092064674 -2.256916753 0.416916753 t-stat results 1997-2002 One-Sample Test Test Value = 0 CONSTANT SET PROFIT INVEST1 INVEST2 t df -2.2812049 9.4480851 3.2708552 -3.6908656 1.053322 5 5 5 5 5 Sig. (2-tailed) 0.071430364 0.000224296 0.022177572 0.014132908 0.340410564 Mean Difference -0.775 1.851666667 3.145 -0.59 0.176666667 95% Confidence Interval of the Difference Lower Upper -1.648310806 0.098310806 1.347875617 2.355457716 0.673328419 5.616671581 -1.000918046 -0.179081954 -0.254479846 0.607813179 t-stat results 1991-2002 One-Sample Test Test Value = 0 CONSTANT SET PROFIT INVEST1 INVEST2 t df 0.7642389 6.2704339 3.4649563 -4.873616 -1.2045706 11 11 11 11 11 Sig. (2-tailed) 0.46080455 6.08364E-05 0.00528676 0.000491864 0.253646707 Mean Difference 0.301666667 1.359166667 8.844166667 -0.474166667 -0.371666667 95% Confidence Interval of the Difference Lower Upper -0.567124248 1.170457582 0.882085539 1.836247794 3.226237404 14.46209593 -0.68830619 -0.260027144 -1.050774076 0.307440742 Source: Developed for this research Note: PROFIT represents Et/At ratio. INVEST1 represents dAt/At ratio and INVEST2 represents Vt/At. 285 DBA Thesis: Malinee Ronapat APPENDIX F: T-STATISTICS (ONE INVESTMENT VARIABLE) Table F1: t-statistics for 1991-1996, 1997-2002, 1991-2002 (one investment variable) t-statistic results 1991-1996 One-Sample Test 286 DBA Thesis: Malinee Ronapat Test Value = 0 CONSTANT SET INVEST2 PROFIT t df Sig. (2-tailed) 3.588465874 1.918639616 -1.812972014 3.710863396 5 5 5 5 0.015733521 0.113125827 0.129575803 0.01384251 t df Sig. (2-tailed) -3.526400773 6.766639201 1.827821418 3.2533474 5 5 5 5 0.016804068 0.001071098 0.127123447 0.022611779 t df Sig. (2-tailed) -0.118025689 4.193188702 -1.087213535 3.27070573 11 11 11 11 0.908174919 0.001502741 0.30019141 0.007455828 Mean Difference 1.12 0.383333333 -1.006666667 13.53333333 95% Confidence Interval of the Difference Lower Upper 0.317692898 1.922307102 -0.130254365 0.896921031 -2.434002019 0.420668686 4.158549858 22.90811681 t-statistic results 1997-2002 One-Sample Test Test Value = 0 CONSTANT SET INVEST2 PROFIT Mean Difference -1.218333333 1.633333333 0.276666667 2.426666667 95% Confidence Interval of the Difference Lower Upper -2.10644156 -0.330225107 1.012845544 2.253821123 -0.112427312 0.665760645 0.509273752 4.344059582 t-statistic results 1991-2002 One-Sample Test Test Value = 0 CONSTANT SET INVEST2 PROFIT Mean Difference -0.049166667 1.008333333 -0.365 7.98 95% Confidence Interval of the Difference Lower Upper -0.966044206 0.867710872 0.479063877 1.537602789 -1.103916098 0.373916098 2.609947471 13.35005253 Source: Developed for this research Note: PROFIT represents Et/At ratio. INVEST2 represents Vt/At. 287 DBA Thesis: Malinee Ronapat APPENDIX G: EXPECTED PERCENTAGE OF PAYERS: CONTROLLING TWO INVESTMENT VARIABLES (CROSSTABS) Table G1: Expected percentage of payers: controlling two investment variables (crosstabs) 1991 CODE2 Total 288 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total 1.00 Expected Count 3.6 28.4 32.0 % within PAYER 15.6% 84.4% 100.0% % within CODE2 20.0% 13.6% 14.3% Expected Count 21.4 170.6 192.0 % within PAYER 10.4% 89.6% 100.0% % within CODE2 80.0% 86.4% 85.7% Expected Count 25.0 199.0 224.0 % within PAYER 11.2% 88.8% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G2: Expected percentage of payers: controlling two investment variables (crosstabs) 1992 CODE2 .00 PAYER .00 1.00 Total Total 1.00 Expected Count .3 23.7 24.0 % within PAYER 8.3% 91.7% 100.0% % within CODE2 66.7% 8.8% 9.4% Expected Count 2.7 227.3 230.0 % within PAYER .4% 99.6% 100.0% % within CODE2 33.3% 91.2% 90.6% Expected Count 3.0 251.0 254.0 % within PAYER 1.2% 98.8% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G3: Expected percentage of payers: controlling two investment variables (crosstabs) 1993 CODE2 .00 PAYER .00 1.00 Total Expected Count Total 1.00 4.1 39.9 44.0 % within PAYER 20.5% 79.5% 100.0% % within CODE2 32.1% 12.8% 14.6% Expected Count 23.9 233.1 257.0 % within PAYER 7.4% 92.6% 100.0% % within CODE2 67.9% 87.2% 85.4% Expected Count 28.0 273.0 301.0 % within PAYER 9.3% 90.7% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G4: Expected percentage of payers: controlling two investment variables (crosstabs) 1994 CODE2 Total 289 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total Expected Count 1.00 1.3 51.7 53.0 % within PAYER 11.3% 88.7% 100.0% % within CODE2 75.0% 14.6% 16.1% Expected Count 6.7 269.3 276.0 % within PAYER .7% 99.3% 100.0% % within CODE2 25.0% 85.4% 83.9% Expected Count 8.0 321.0 329.0 % within PAYER 2.4% 97.6% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G5: Expected percentage of payers: controlling two investment variables (crosstabs) 1995 CODE2 .00 PAYER .00 1.00 Total Expected Count Total 1.00 .9 53.1 54.0 % within PAYER 11.1% 88.9% 100.0% % within CODE2 100.0% 14.2% 15.7% Expected Count 5.1 284.9 290.0 % within PAYER .0% 100.0% 100.0% % within CODE2 .0% 85.8% 84.3% Expected Count 6.0 338.0 344.0 % within PAYER 1.7% 98.3% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G6: Expected percentage of payers: controlling two investment variables (crosstabs) 1996 CODE2 .00 PAYER .00 1.00 Total Expected Count Total 1.00 1.5 70.5 72.0 % within PAYER 9.7% 90.3% 100.0% % within CODE2 87.5% 17.6% 19.0% Expected Count 6.5 299.5 306.0 % within PAYER .3% 99.7% 100.0% % within CODE2 12.5% 82.4% 81.0% Expected Count 8.0 370.0 378.0 % within PAYER 2.1% 97.9% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G7: Expected percentage of payers: controlling two investment variables (crosstabs) 1997 CODE2 Total 290 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total Expected Count 1.00 42.2 69.8 112.0 % within PAYER 58.0% 42.0% 100.0% % within CODE2 43.9% 19.2% 28.5% Expected Count 105.8 175.2 281.0 % within PAYER 29.5% 70.5% 100.0% % within CODE2 56.1% 80.8% 71.5% Expected Count 148.0 245.0 393.0 % within PAYER 37.7% 62.3% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G8: Expected percentage of payers: controlling two investment variables (crosstabs) 1998 CODE2 .00 PAYER .00 1.00 Total Expected Count Total 1.00 38.0 240.0 278.0 % within PAYER 18.3% 81.7% 100.0% % within CODE2 96.2% 67.8% 71.6% Expected Count 15.0 95.0 110.0 % within PAYER 1.8% 98.2% 100.0% % within CODE2 3.8% 32.2% 28.4% Expected Count 53.0 335.0 388.0 % within PAYER 13.7% 86.3% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G9: Expected percentage of payers: controlling two investment variables (crosstabs) 1999 CODE2 .00 PAYER .00 1.00 Total Expected Count Total 1.00 51.2 218.8 270.0 % within PAYER 26.3% 73.7% 100.0% % within CODE2 97.3% 63.8% 70.1% Expected Count 21.8 93.2 115.0 % within PAYER 1.7% 98.3% 100.0% % within CODE2 2.7% 36.2% 29.9% Expected Count 73.0 312.0 385.0 % within PAYER 19.0% 81.0% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G10: Expected percentage of payers: controlling two investment variables (crosstabs) 2000 CODE2 Total 291 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total Expected Count 1.00 34.9 203.1 238.0 % within PAYER 22.3% 77.7% 100.0% % within CODE2 98.1% 58.9% 64.7% Expected Count 19.1 110.9 130.0 % within PAYER .8% 99.2% 100.0% % within CODE2 1.9% 41.1% 35.3% Expected Count 54.0 314.0 368.0 % within PAYER 14.7% 85.3% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G11: Expected percentage of payers: controlling two investment variables (crosstabs) 2001 CODE2 .00 PAYER .00 1.00 Total Expected Count Total 1.00 24.9 198.1 223.0 % within PAYER 18.4% 81.6% 100.0% % within CODE2 97.6% 54.5% 59.3% Expected Count 17.1 135.9 153.0 % within PAYER .7% 99.3% 100.0% % within CODE2 2.4% 45.5% 40.7% Expected Count 42.0 334.0 376.0 % within PAYER 11.2% 88.8% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research Table G12: Expected percentage of payers: controlling two investment variables (crosstabs) 2002 CODE2 .00 PAYER .00 1.00 Total Expected Count Total 1.00 17.2 190.8 208.0 % within PAYER 15.4% 84.6% 100.0% % within CODE2 100.0% 49.6% 53.7% Expected Count 14.8 164.2 179.0 % within PAYER .0% 100.0% 100.0% % within CODE2 .0% 50.4% 46.3% Expected Count 32.0 355.0 387.0 % within PAYER 8.3% 91.7% 100.0% % within CODE2 100.0% 100.0% 100.0% Source: Developed for this research 292 DBA Thesis: Malinee Ronapat APPENDIX H: EXPECTED PERCENTAGE OF PAYERS: CONTROLLING ONE INVESTMENT VARIABLE (CROSSTABS) Table H1: Expected percentage of payers: controlling one investment variable (crosstab) 1991 CODE4 Total 293 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total Expected Count 1.00 .1 31.9 32.0 % within PAYER .0% 100.0% 100.0% % within CODE4 .0% 14.3% 14.3% Expected Count .9 191.1 192.0 % within PAYER .5% 99.5% 100.0% % within CODE4 100.0% 85.7% 85.7% Expected Count 1.0 223.0 224.0 % within PAYER .4% 99.6% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H2: Expected percentage of payers: controlling one investment variable (crosstab) 1992 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 .2 23.8 24.0 % within PAYER 8.3% 91.7% 100.0% % within CODE4 100.0% 8.7% 9.4% Expected Count 1.8 228.2 230.0 % within PAYER .0% 100.0% 100.0% % within CODE4 .0% 91.3% 90.6% Expected Count 2.0 252.0 254.0 % within PAYER .8% 99.2% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H3: Expected percentage of payers: controlling one investment variable (crosstab) 1993 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 3.1 40.9 44.0 % within PAYER 11.4% 88.6% 100.0% % within CODE4 23.8% 13.9% 14.6% Expected Count 17.9 239.1 257.0 % within PAYER 6.2% 93.8% 100.0% % within CODE4 76.2% 86.1% 85.4% Expected Count 21.0 280.0 301.0 % within PAYER 7.0% 93.0% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H4: Expected percentage of payers: controlling one investment variable (crosstab) 1994 CODE4 Total 294 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total Expected Count 1.00 .8 52.2 53.0 % within PAYER 5.7% 94.3% 100.0% % within CODE4 60.0% 15.4% 16.1% Expected Count 4.2 271.8 276.0 % within PAYER .7% 99.3% 100.0% % within CODE4 40.0% 84.6% 83.9% Expected Count 5.0 324.0 329.0 % within PAYER 1.5% 98.5% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H5: Expected percentage of payers: controlling one investment variable (crosstab) 1995 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 .8 53.2 54.0 % within PAYER 9.3% 90.7% 100.0% % within CODE4 100.0% 14.5% 15.7% Expected Count 4.2 285.8 290.0 % within PAYER .0% 100.0% 100.0% % within CODE4 .0% 85.5% 84.3% Expected Count 5.0 339.0 344.0 % within PAYER 1.5% 98.5% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H6: Expected percentage of payers: controlling one investment variable (crosstab) 1996 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 1.7 70.3 72.0 % within PAYER 9.7% 90.3% 100.0% % within CODE4 77.8% 17.6% 19.0% Expected Count 7.3 298.7 306.0 % within PAYER .7% 99.3% 100.0% % within CODE4 22.2% 82.4% 81.0% Expected Count 9.0 369.0 378.0 % within PAYER 2.4% 97.6% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H7: Expected percentage of payers: controlling one investment variable (crosstab) 1997 CODE4 Total 295 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total Expected Count 1.00 42.7 69.3 112.0 % within PAYER 58.9% 41.1% 100.0% % within CODE4 44.0% 18.9% 28.5% Expected Count 107.3 173.7 281.0 % within PAYER 29.9% 70.1% 100.0% % within CODE4 56.0% 81.1% 71.5% Expected Count 150.0 243.0 393.0 % within PAYER 38.2% 61.8% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H8: Expected percentage of payers: controlling one investment variable (crosstab) 1998 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 36.5 241.5 278.0 % within PAYER 16.9% 83.1% 100.0% % within CODE4 92.2% 68.5% 71.6% Expected Count 14.5 95.5 110.0 % within PAYER 3.6% 96.4% 100.0% % within CODE4 7.8% 31.5% 28.4% Expected Count 51.0 337.0 388.0 % within PAYER 13.1% 86.9% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H9: Expected percentage of payers: controlling one investment variable (crosstab) 1999 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 51.9 218.1 270.0 % within PAYER 26.7% 73.3% 100.0% % within CODE4 97.3% 63.7% 70.1% Expected Count 22.1 92.9 115.0 % within PAYER 1.7% 98.3% 100.0% % within CODE4 2.7% 36.3% 29.9% Expected Count 74.0 311.0 385.0 % within PAYER 19.2% 80.8% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H10: Expected percentage of payers: controlling one investment variable (crosstab) 2000 CODE4 Total 296 DBA Thesis: Malinee Ronapat .00 PAYER .00 1.00 Total 1.00 % within PAYER 20.6% 79.4% 100.0% % within CODE4 98.0% 59.4% 64.7% % within PAYER 100.0% .8% 99.2% % within CODE4 2.0% 40.6% 35.3% % within PAYER 13.6% 86.4% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H11: Expected percentage of payers: controlling one investment variable (crosstab) 2001 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 24.9 198.1 223.0 % within PAYER 18.4% 81.6% 100.0% % within CODE4 97.6% 54.5% 59.3% Expected Count 17.1 135.9 153.0 % within PAYER .7% 99.3% 100.0% % within CODE4 2.4% 45.5% 40.7% Expected Count 42.0 334.0 376.0 % within PAYER 11.2% 88.8% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research Table H12: Expected percentage of payers: controlling one investment variable (crosstab) 2002 CODE4 .00 PAYER .00 1.00 Total Expected Count Total 1.00 14.5 193.5 208.0 % within PAYER 13.0% 87.0% 100.0% % within CODE4 100.0% 50.3% 53.7% Expected Count 12.5 166.5 179.0 % within PAYER .0% 100.0% 100.0% % within CODE4 .0% 49.7% 46.3% Expected Count 27.0 360.0 387.0 % within PAYER 7.0% 93.0% 100.0% % within CODE4 100.0% 100.0% 100.0% Source: Developed for this research 297 DBA Thesis: Malinee Ronapat 298
© Copyright 2024 Paperzz