vii TABLE OF CONTENTS CHAPTER 1 2 TITLE PAGE DECLARATION ii DEDICATION iii ACKNOWLEDGEMENTS iv ABSTRACT v ABSTRAK vi TABLES OF CONTENTS vii LIST OF TABLES x LIST OF FIGURES xii LIST OF APPENDICES xiv INTRODUCTION 1 1.0 Introduction 1 1.1 Background of Study 2 1.2 Problem Statements 4 1.3 Research Objective 6 1.4 Research Questions 6 1.5 Scope of the Study 7 1.6 Significance of the Study 7 1.7 Organisation of the Report 8 LITERATURE REVIEW 9 2.0 Introduction 9 2.1 Dynamic Regression Model 9 2.2 Autoregressive Distributed Lag (ARDL) 11 viii Model 3 2.3 Related Studies on ARDL 13 2.4 Related Studies on Palm Oil Price 19 2.5 Summary 23 RESEARCH METHODOLOGY 26 3.0 Introduction 26 3.1 Testing for Stationarity 26 3.2 ARDL model 29 3.2.1 Model Specification 30 3.2.2 ARDL Bound Test for Cointegration 30 3.2.3 Expressing ARDL Model as Error 31 Correction Model (ECM) 3.2.4 Estimation of the Long-run 33 Coefficients 3.2.5 Choosing Lag Length on First 35 Difference Variables 3.2.6 Estimation of ARDL model 37 Box-Jenkins ARIMA model 37 3.3.1 Stationary Time Series Model 38 3.3.2 Non-Stationary Time Series Model 39 3.3.3 Model Identification 40 3.3.4 Parameter Estimation 41 Model Diagnostic 45 3.4.1 Stability Test 46 3.5 Model Evaluation 49 3.6 Operational Framework 51 3.7 Summary 54 3.3 3.4 4 DATA ANALYSIS 55 4.0 Introduction 55 4.1 Data Description 55 4.2 Analysis by using ARDL model 59 ix 4.2.1 Stationary Testing 59 4.2.2 Existence of the Long-run Relationship 69 4.2.3 Estimation of the ARDL model 70 4.2.4 Estimation of the Long-run 74 Coefficients 4.2.5 Estimation of the Error Correction 76 Model 4.3 4.4 4.2.6 Model Diagnostics 78 Analysis by using ARIMA model 81 4.3.1 Stationary Testing 82 4.3.2 Model Identification 84 4.3.3 Parameter Estimation 86 4.3.4 Model Diagnostics 88 Data Modelling 89 4.4.1 Modelling using ARDL (2,2,0,0,2) 89 4.4.2 Modelling using ARDL (1,2,0,0,2) 91 4.4.3 Modelling using ARIMA (1,1,6) 92 4.4.4 Modelling Performances of ARDL 93 and ARIMA Model 4.5 5 Summary 94 CONCLUSION AND RECOMMENDATIONS 95 5.0 Introduction 95 5.1 Summary 95 5.3 Conclusions 96 5.3 Suggestion for Future Study 98 REFERENCES Appendices A – B 101 106-131 x LIST OF TABLES TABLE NO. TITLE PAGE 2.1 Summary of studies for ARDL model 23 4.1 ADF test of the logarithmic price at level 60 4.2 ADF test of the logarithmic price at first difference 61 4.3 ADF test of the logarithmic production at level 62 4.4 ADF test of the logarithmic production at first 63 difference 4.5 ADF test of the logarithmic import at level 65 4.6 ADF test of the logarithmic export at level 66 4.7 ADF test of the logarithmic stock at level 67 4.8 Integrating order of all the logarithmic variables 68 4.9 F-statistics for testing the existence of a long-run price 70 equation 4.10 Estimated ARDL (2,2,0,0,2) 71 4.11 Estimated ARDL (1,2,0,0,2) 72 4.12 Long-run coefficients estimates for ARDL (2,2,0,0,2) 74 4.13 Long-run coefficients estimates for ARDL (1,2,0,0,2) 75 4.14 Error correction representation of ARDL Model 1 76 4.15 Error correction representation of ARDL Model 2 77 4.16 ADF test of the original data for price 82 4.17 ADF test of the price at first difference 83 4.18 Possible ARIMA models for price 85 4.19 Equations of ARIMA (p,d,q) for price 86 xi 4.20 Summary of Q-statistics for ARIMA model 88 4.21 Corresponding AIC for the adequate model 88 4.22 Modelling performance of ARDL (2,2,0,0,2) 90 4.23 Modelling performance of ARDL (1,2,0,0,2) 92 4.24 Modelling performance of ARIMA (1,1,6) 93 4.25 Modelling performance of ARDL and ARIMA models 93 5.1 Modelling performance of ARDL and ARIMA models 98 xii LIST OF FIGURES FIGURE NO. 3.1 TITLE PAGE The process in developing an ARDL model using 52 ARDL bound test 3.2 The process in developing an ARIMA model 53 4.1 Monthly prices of crude palm oil from January 2000 56 until December 2013 4.2 Monthly production of crude palm oil from January 57 2000 until December 2013 4.3 Monthly import volume of crude palm oil from 57 January 2000 until December 2013 4.4 Monthly export volume of crude palm oil from 58 January 2000 until December 2013 4.5 Monthly closing stock of crude palm oil from 58 January 2000 until December 2013 4.6 Asymptotic critical value bounds for the F-statistic 69 for Case III (unrestricted intercepts; no trends) 4.7 ACF, PACF and Q-statistics of the residuals for 78 ARDL (2,2,0,0,2) 4.8 ACF, PACF and Q-statistics of the residuals for 79 ARDL (1,2,0,0,2) 4.9 Plot for a) CUSUM and b) CUSUMQ of the 80 recursive residual for ARDL (2,2,0,0,2) 4.10 Plot for a) CUSUM and b) CUSUMQ of the recursive residual for ARDL (1,2,0,0,2) 80 xiii 4.11 ACF and PACF for price at first difference level 85 4.12 Modelling graph for palm oil price using ARDL 90 (2,2,0,0,2) 4.13 Modelling graph for palm oil price using ARDL 91 (1,2,0,0,2) 4.14 Modelling graph for palm oil price using ARIMA (1,1,6) 92 xiv LIST OF APPENDICES APPENDIX A TITLE Monthly price, production, closing stock, PAGE 106 import and export volume of palm oil B All ARDL and ARIMA models involved in modelling palm oil price 121