On Currency Crises, Exchange Rate Regimes and Contagion Victor Pontines A thesis submitted to the University of Adelaide in fulfilment of the requirements for the degree of Doctor of Philosophy School of Economics February 2006 Table of Contents List of Tables and Figures lv Abstract vii Declaration xi Acknowledgements xii Chapter 1 Introduction Chapter 2 An Extreme Value Approach in Identifying Speculative 9 Attacks in East Asia and Latin America 2,1 Introduction 9 2.2 Review of the Related Empirical Literature 13 2.2.1 Definition of Currency Crises l3 2.2.2 The Statistical Distribution of Financial or Asset Returns (Speculative Prices) 22 2.3 The Concept of an Exchange Market Pressure Index 23 2.3.I Eichengreen, Rose and Wyplosz (1995, 1996) 25 2.3.2 Sachs, Tornell and Velasco (1996) 25 2.3.3 Kaminsky, Lizondo and Reinhart (1998, 1999) 26 2,4 Extreme Value Theory 27 2.4.1 The Hill and the HKPP Estimators 29 2.5 Data and Some Preliminary Results 30 2.5.I Data 30 2.5.2 Preliminary Results 31 2.5.2.1 Summary Statistics 31 2.5.2.2 The Hill and HKKP Estimators 43 2.6 Results and Analyses 54 2.7 Brief Conclusion 66 Chapter 3 A Re-examination of the Evidence of Post-Crisis Dollar-Peg Reversion in the East Asia-S Using Regime Switching ARCH 68 3.1 Introduction 68 3.2 Literature Review 73 3.3 Exchange Market Pressure and Exchange Rate Policy 79 3.4 Markov-Regime Switching ARCH (SWARCH) 81 83 3 .4.I Smoothed Probabilities and the Index of Intervention 3.5 Data Description and Preliminary Results 85 3.6 Results and Analysis 88 3.6.1 Indonesia 88 3.6.2 Korea 94 3.6.3 Philippines 98 3.6.4 Singapore 101 3.6.5 Thailand 102 ll 3.7 Brief Conclusion 106 Chapter 4 ARegime Switching Approach to Correlation-based Test of Contagion : The Case of the East Asia-7 108 4.1 Introduction 108 4.2 Defining Contagion 111 4.3 Brief Theoretical Literature on the International Transmission Mechanisms 113 4.3.I Crisis-Contingent versus Non-Crisis Contingent Theories 113 4.4 Some Empirical Evidence on the International Transmission Mechanisms 115 4.5 Tie-Up to the Existing Empirical Literature 118 4.6 Method t22 4.6.1 The Markov Switching VAR r22 4.6.2 Causality Tests Using the Markov Switching VAR t24 4.7 Data and Preliminary Results t25 4.8 Results and Analyses 132 4.9 Conclusion 184 5. Conclusion : Policy Implications and Scope for Further Research 187 Appendix 4.1: Sensitivity Test Results Using the Nominal Exchange Rate in the Construction of the EMP Indices t94 (Appendix to Chapter 2) Appendix 4.2 Crisis Episodes Identified by Conventional and EVT Approaches with Corresponding Chronologies of Political 220 and Economic Events (Appendix to Chapter 2) Appendix B: Sensitivity Test Results'Without the Interest Rate Controls in the Estimation of the MS-VAR model 24r (Appendix to Chapter 4) References 285 lll List of Tables and Figures Tables Table 2.1: Various Issues in the Definition of Currency Crises 18 Table 2.2:Desuiptive Statistics of Individual EMP Measures (East Asian Sample) JJ Table 2.3: Descriptive Statistics of Individual EMP Measures (Latin American Sample) 34 Table 2.4: Normality Tests for the Individual EMP Measures (East Asian Sample) 35 Table 2.5: Normality Tests for the Individual EMP Measures (Latin American Sample) 36 Table 2.6:IJnitRoot tests for the Individual EMP Measures (East Asian Sample) 44 Table 21:IJnitRoot tests for the Individual EMP Measures (Latin American Sample) 45 Table 2.8: Ljung-Box Q-statistics for East Asian and Latin American Countries 46 Table 2.9: Number and Proportion of crises episodes according to the ERW- EMP 55 Table 2.10: Number and Proportion of crises episodes according to the KLR- EMP 56 Table 2.1 1: Number and Proportion of crises episodes according to the STV- EMP 57 Table 2.12: Number of monthly episodes of crises and incidence of crises using the extreme value theory and ERW-EMP 59 Table 2.13: Number of monthly episodes of crises and incidence of crises using the extreme value theory and KLR-EMP 60 Table 2,14: Number of monthly episodes of crises and incidence of crises using the extreme value theory and STV-EMP 6l Table 2.15: Crisis Episodes According to Conventional Method (East Asian and Latin American SamPle 63 Table 2.16: Crises Episodes According to Extreme Value Theory (East Asian Sample 64 Table 2.16: Crises Episodes According to Extreme Value Theory (Latin American SamPle) 65 Table 3.1: De Jure Exchange Rate Regime Classification 69 Table 3.2:IJnivariate Statistics on Exchange Rates (EXR), Reserves and Interest Rates (INT) 86 Table 3.3: 2-state regime switching ARCH regressions for percentage changes in exchange rates 89 Table 3.4: 2-state regime switching ARCH regressions for percentage changes in reserves 90 Table 3.5:2-state regime switching ARCH regressions for f,rrst-difference in in interest rates 9l lv Table 4.1: 1997 East Asian Crisis: Unconditional Correlation Coefficients 109 Table 4.2: Summary of Empirical Evidence on Contagion versus Interdependence using The cross-Market correlation Returns As Their Testing MethodologY tt9 Table 4.3: Descriptive Statistics of Stock Returns in East Asia, January 1994- December 2003 t27 Table 4.4: Test for Regime Switching: Linear VAR versus Markov Switching VAR r28 Table 4.5: Test for Regime-dependent Heteroscedasticity 130 Table 4.6: Granger Causality Test Using the MS-VAR Model 131 Table 4.7: MS-VAR Estimation Results, 1994-98 134 Table 4.8: MS-VAR Estimation Results, 1994-99 t4r Table 4.9: MS-VAR Estimation Results, 1994-00 148 Table 4.10: MS-VAR Estimation Results, 1994-0I 155 Table 4,1 1: MS-VAR Estimation Results,1994-02 162 Table 4.12: MS-VAR Estimation Results ,1994-03 t69 Table 4.9: Contagion versus Interdependence: 1994-1998 178 Table 4 .9 : Contagion versus Interdep end enc e : 1 9 9 4 - 19 9 9 t79 Table 4.9: Contagion versus Interdependence: 1994-2000 180 Table 4.10: Contagion versus Interdependence: 1994-2001 181 Table 4.1 1: Contagion versus Interdependence: 1994-2002 r82 Table 4 .12: Contagion versus Interdepend ence: 199 4-2003 183 Figures Figure 2.1: Approaches to building leading indicator models of currency crises 10 Figure 2.2:Histogram of country ERW-EMP Measures and corresponding Normal Probability Density Function (East Asia) JI Figure 2.3: Histogram of Country ERW-EMP Measures and Corresponding Normal Probability Density Function (Latin America) 38 Figure 2.4:Histogram of country KLR-EMP Measures and corresponding Normal Probability Density Function (East Asia) 39 Figure 2.5: Histogram of country KLR-EMP Measures and corresponding Normal Probability Density Function (Latin America) 40 Figure 2.6: Histogram of country STV-EMP Measures and corresponding Normal Probability Density Function (East Asia) 4t Figure 2.7: Histogram of country STV-EMP Measures and corresponding Normal Probability Density Function (Latin America) 42 Figure 2.8: Recursive Residuals for East Asia using ERW as the EMP Measure 48 Figure 2.9: Recursive Residuals for Latin America using ERW as the EMP Measure 49 Figure 2.10: Recursive Residuals for East Asia using KLR as the EMP Measure 50 Figure 2.1 1: Recursive Residuals for Latin America using KLR as the EMP Measure 51 Figure 2.12: Recursive Residuals for East Asia using STV as the EMP Measure 52 Figure 2.13: Recursive Residuals for Latin America Using STV as the EMP V Measure 53 Figure 3. 1 : Index of Intervention Estimates for Indonesia 92 Figure 3.2: Index of Intervention Estimates for Korea 97 Figure 3.3: Index of Intervention Estimates for Philippines 100 Figure 3.4: Index of Intervention Estimates for Singapore 104 Figure 3.5: Index of Intervention Estimates for Thailand 105 v1 On Currency Crises, Exchange Rate Regimes and Contagion Abstract The enormous literature that the East Asian crisis generated has provided the necessary fertile inputs that brought forth specific proposals on how to strengthen the architecture of the international monetary and financial system. Viewed in this context, this thesis attempts to revisit or re-assess prevailing conventional wisdom or recent empirical evidence on some key issues that have arisen in discussions/debates regarding reform efforts towards shengthening the international financial architecture. Towards this end, the core chapters rely on recently developed research methods in time-series econometrics and examine the major East Asian countries that were acknowledged to have either directly or indirectly been impacted by the crisis of t997-98. The starting point or the f,rrst core chapter (chapter 2) focuses on speculative attacks and crises identification, one of the basic ingredients of early warning system models that seek to identiff and measure the determinants of a countries' vulnerability to crises. The underlying contention is that the current design of these early warning models in terms of empirically def,rning or identiffing currency crises is that it implicitly assumes that either the changes in exchange rates alone or an index of exchange market pressure are normally distributed. If, on the contrary, however, these variables are not normally distributed, this signihes that the conventional method of employing the mean and some arbitrary multiples of the standard deviation in current designs of early warning systems underestimates the frequency or incidence of speculative attacks. Using three indices of exchange market pressure popularly adopted in the literature and employing an alternative approach that makes far fewer parametric assumptions, the results reveal that these indices are non-noÍnal, while, in vll the process, there is evidence to suggest that the identification of currency crises ts sensitive to the choice one adopts with regard to the weights of the index of exchange market pressure.
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages307 Page
-
File Size-