The interdependence of financial markets [Elektronische Ressource] : econometric modeling and estimation / vorgelegt von Dirk Baur

Informations

• Publié par: eberhard_karls_universitat_tubingen
• Publié le: 01 janvier 2004
• Nombre de lectures: 17
• Langue: English
• Poids de l'ouvrage: 1 Mo

Extrait

The Interdependence of Financial Markets -
Econometric Modeling and Estimation
Inaugural-Dissertation
zur Erlangung des Doktorgrades
der Wirtschaftswissenschaftlichen Fakultät
der Eberhard-Karls-Universität zu Tübingen
vorgelegt von
Dirk Baur
aus Stuttgart
2004Dekan: Prof. Dr. rer. pol. Renate Hecker
Erstberichterstatter: Prof. Dr. rer. pol. Gerd Ronning
Zweitberic Prof. Dr. rer. pol. Joachim Grammig
Tag der mündlichen Prüfung: 26. August 2003Contents
1 Introduction 9
2 Symmetric Interdependence: Correlations 13
2.1 Univariate GARCH Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.1 Asymmetric GARCH Models . . . . . . . . . . . . . . . . . . . . . . . . 18
2.1.2 The News-Impact Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.1.3 Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Multivariate GARCH Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.1 Existing Multivariate GARCH Models . . . . . . . . . . . . . . . . . . . 29
2.2.2 Bivariate Dynamic Correlations (BDC) . . . . . . . . . . . . . . . . . . 36
2.2.3 Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.2.4 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.2.5 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.2.7 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3 Asymmetric Interdependence: Spillovers 61
3.1 Mean and Volatility Spillovers . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.2 Constant Mean and Volatility Spillovers . . . . . . . . . . . . . . . . . . . . . . 67
33.2.1 Spurious Correlations and Spillovers . . . . . . . . . . . . . . . . . . . 67
3.2.2 The Econometric Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.2.3 The Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.2.4 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.2.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.3 Varying Mean Spillovers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.3.1 Time-varying Mean Spillovers . . . . . . . . . . . . . . . . . . . . . . . 84
3.3.2 Conditional Mean Spillovers . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.3.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.3.4 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.4 Mean and Volatility Contagion . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.4.1 Modeling Contagion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.4.2 Excess comovement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.4.3 Mean Contagion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.4.4 Volatility Contagion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.4.5 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.4.7 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4 Concluding Remarks 121
5 References 124List of Figures
2.1 Simulated GARCH(1,1) processes.......................... 17
2.2 News-Impact Curves ................................. 23
2.3 Daily Returns (Japan, UK, Germany, US) ..................... 46
2.4 Monthly Returns (Japan, UK, Germany, US) ................... 47
2.5 Asymmetric Time-varying Correlations (BDC ), (JAP, UK), (JAP, GER), (JAP,z
US), (UK, GER), (UK, US), (GER, US) ....................... 51
2.6 News-Impact Surfaces and frontal views, BDC model (JAP/ UK) (top) andz
(JAP/ GER) (bottom) ................................. 53
2.7 News-Impact Surfaces and frontal views, BDC model (JAP/ US) (top) andz
(UK/ GER) (bottom) .................................. 55
2.8 News-Impact Surfaces and frontal views, BDC model (UK/ US) (top) andz
(GER, US) (bottom) 56
3.1 Time-varying Correlations (Spillovers): top: (DAXNR , DOWDR ), inter-t t−1
mediate: (DOWNR , DAXDR ), (DAXDR , DOWDR ), bottom: (DOWDR ,t t t t−1 t
DAXDR ) ........................................ 90
t
3.2 Conditional Correlations (Spillovers): top: (DAXNR , DOWDR ), inter-t t−1
mediate: (DOWNR , DAXDR ), (DAXDR , DOWDR ), bottom: (DOWDR ,t t t t−1 t
DAXDR ) ........................................ 95
t
53.3 Simulated Correlation process (1)..........................103
3.4 Simulated Correlation process (2)104
3.5 Volatility Contagion..................................108
3.6 Asian markets, first 180 trading days .......................114List of Tables
2.1 Simulation Results - Multivariate GARCH Models . . . . . . . . . . . . . . . . 42
2.2 Simulations Results - BDC Model (two-step procedure) . . . . . . . . . . . . 43z
2.3 Descriptive Statistics (daily data) . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.4 Descriptive Statistics (monthly data) . . . . . . . . . . . . . . . . . . . . . . . . 45
2.5 Unconditional Correlation (daily data) . . . . . . . . . . . . . . . . . . . . . . . 45
2.6 Correlation (monthly data) . . . . . . . . . . . . . . . . . . . . . 46
2.7 Daily Data: BDC MODEL, Asymmetric Volatility . . . . . . . . . . . . . . . . 48z
2.8 Daily Data: BDC MODEL, Asymmetric Correlations . . . . . . . . . . . . . . 48z
2.9 Monthly Data: BDC MODEL, Asymmetric Correlations . . . . . . . . . . . . 48z
2.10 Daily Data: Parameter Comparison . . . . . . . . . . . . . . . . . . . . . . . . 49
2.11 Daily Data: Asymmetric Diagonal BEKK MODEL . . . . . . . . . . . . . . . . 49
3.1 Timing of Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2 Classification of Spillovers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3 Timing of the opening and the closing of the Frankfurt and the New York
stock market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.4 DAX overnight returns and the preceding daytime returns of DOW . . . . . . 73
3.5 DOW overnight returns and the preceding daytime returns of DAX . . . . . . 74
73.6 DAX morning returns: spillovers from the preceding day in New York . . . . 76
3.7 DOW morning returns: spillovers from the morning in Frankfurt . . . . . . . 77
3.8 Simulation Results - Pure Mean Spillover . . . . . . . . . . . . . . . . . . . . . 79
3.9 Simulation Results - Pure Volatility Spillover . . . . . . . . . . . . . . . . . . . 81
3.10 Simulation Results - Mean and Volatility Spillover . . . . . . . . . . . . . . . 82
3.11 Time-varying Correlation (DAXNR , DOWDR ) ............... 88t t−1
3.12 Correlation (DOWNR , DAXDR)................. 88t t
3.13 Time-varying Spillover (DAXDR , DOWDR ). 89t t−1
3.14 Spillover (DOWDR , DAXDR ).................. 91
t t
3.15 Conditional Correlation (DAXNR , DOWDR ) ................ 93t t−1
3.16 Correlation (DOWNR , DAXDR). 94t t
3.17 Conditional Spillover (DAXDR , DOWDR ).................. 94t t−1
3.18 Spillover (DOWDR , DAXDR )................... 94
t t
3.19 Descriptive Statistics (Asian Markets) . . . . . . . . . . . . . . . . . . . . . . . 111
3.20 Correlations (Asian Markets) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.21 Crises correlations (Asian Markets) . . . . . . . . . . . . . . . . . . . . . . . . 111
3.22 Mean and Volatility Contagion (Hong Kong Crisis) . . . . . . . . . . . . . . . . 112
3.23 Mean and Volatility Contagion (Thailand Crisis) . . . . . . . . . . . . . . . . . 113
3.24 Simulation Results - Contagion Tests . . . . . . . . . . . . . . . . . . . . . . . 118Chapter 1
Introduction
We live in an era of interdependence.
Keohane and Nye, 2002
The above statement expresses a widespread feeling that the world we live in is now
more interconnected than it was before. However, such statements do usually not deliver
precise definitions of the words entailed and examples of interdependent phenomena in
the medical, social, political and economic aspects of our existence, not to mention the
economic structures, are infinite (see Drouet and Kotz, 2001).
We exclusively study economic interdependence and focus on the interdependence of
financial markets.
It is noteworthy that the finance literature has neither provided a generally accepted
definition or description for "interdependence" nor for "dependence". However, a thorough
analysis of "interdependence" requires an accurate definition of the term before examining
the sources and the constituting factors of this phenomenon in a static and also a dynamic
sense. After the discussion of such issues, it is surely interesting to evaluate the results of
such an interconnectedness. Any increased knowledge could lead to a better understand-
ing of the functioning of the international financial system and could answer the question
whether the financial markets are part of an appropriate financial architecture or not. An
9analysis of interdependence can potentially also assign a new role to the Internation