Contents
Preface page xv
Notation and Abbreviations xix
List of Contributors xxv
1 Initial Tasks and Overview 1
Helmut L¨utkepohl
1.1 Introduction 1
1.2 Setting Up an Econometric Project 2
1.3 Getting Data 3
1.4 Data Handling 5
1.5 Outline of Chapters 5
2 Univariate Time Series Analysis 8
Helmut L¨utkepohl
2.1 Characteristics of Time Series 8
2.2 Stationary and Integrated Stochastic Processes 11
2.2.1 Stationarity 11
2.2.2 Sample Autocorrelations, Partial Autocorrelations,
and Spectral Densities 12
2.2.3 Data Transformations and Filters 17
2.3 Some Popular Time Series Models 22
2.3.1 Autoregressive Processes 22
2.3.2 Finite-Order Moving Average Processes 25
2.3.3 ARIMA Processes 27
2.3.4 Autoregressive Conditional Heteroskedasticity 28
2.3.5 Deterministic Terms 30
2.4 Parameter Estimation 30
2.4.1 Estimation of AR Models 30
2.4.2 Estimation of ARMA Models 32
2.5 Model Specification 33
ix
x Contents
2.5.1 AR Order Specification Criteria 33
2.5.2 Specifying More General Models 35
2.6 Model Checking 40
2.6.1 Descriptive Analysis of the Residuals 40
2.6.2 Diagnostic Tests of the Residuals 43
2.6.3 Stability Analysis 47
2.7 Unit Root Tests 53
2.7.1 Augmented Dickey–Fuller (ADF) Tests 54
2.7.2 Schmidt–Phillips Tests 57
2.7.3 A Test for Processes with Level Shift 58
2.7.4 KPSS Test 63
2.7.5 Testing for Seasonal Unit Roots 65
2.8 Forecasting Univariate Time Series 70
2.9 Examples 73
2.9.1 German Consumption 73
2.9.2 Polish Productivity 78
2.10 Where to Go from Here 85
3 Vector Autoregressive and Vector Error Correction Models 86
Helmut L¨utkepohl
3.1 Introduction 86
3.2 VARs and VECMs 88
3.2.1 The Models 88
3.2.2 Deterministic Terms 91
3.2.3 Exogenous Variables 92
3.3 Estimation 93
3.3.1 Estimation of an Unrestricted VAR 93
3.3.2 Estimation of VECMs 96
3.3.3 Restricting the Error Correction Term 105
3.3.4 Estimation of Models with More General Restrictions
and Structural Forms 108
3.4 Model Specification 110
3.4.1 Determining the Autoregressive Order 110
3.4.2 Specifying the Cointegrating Rank 112
3.4.3 Choice of Deterministic Term 120
3.4.4 Testing Restrictions Related to the Cointegration
Vectors and the Loading Matrix 121
3.4.5 Testing Restrictions for the Short-Run Parameters
and Fitting Subset Models 122
3.5 Model Checking 125
3.5.1 Descriptive Analysis of the Residuals 125
3.5.2 Diagnostic Tests 127
3.5.3 Stability Analysis 131
Contents xi
3.6 Forecasting VAR Processes and VECMs 140
3.6.1 Known Processes 141
3.6.2 Estimated Processes 143
3.7 Granger-Causality Analysis 144
3.7.1 The Concept 144
3.7.2 Testing for Granger-Causality 148
3.8 An Example 150
3.9 Extensions 158
4 Structural Vector Autoregressive Modeling and Impulse
Responses 159
J¨org Breitung, Ralf Br¨uggemann, and Helmut L¨utkepohl
4.1 Introduction 159
4.2 The Models 161
4.3 Impulse Response Analysis 165
4.3.1 Stationary VAR Processes 165
4.3.2 Impulse Response Analysis of Nonstationary VARs
and VECMs 167
4.4 Estimation of Structural Parameters 172
4.4.1 SVAR Models 172
4.4.2 Structural VECMs 175
4.5 Statistical Inference for Impulse Responses 176
4.5.1 Asymptotic Estimation Theory 176
4.5.2 Bootstrapping Impulse Responses 177
4.5.3 An Illustration 179
4.6 Forecast Error Variance Decomposition 180
4.7 Examples 183
4.7.1 A Simple AB-Model 183
4.7.2 The Blanchard–Quah Model 185
4.7.3 An SVECM for Canadian Labor Market Data 188
4.8 Conclusions 195
5 Conditional Heteroskedasticity 197
Helmut Herwartz
5.1 Stylized Facts of Empirical Price Processes 197
5.2 Univariate GARCH Models 198
5.2.1 Basic Features of GARCH Processes 199
5.2.2 Estimation of GARCH Processes 201
5.2.3 Extensions 203
5.2.4 Blockdiagonality of the Information Matrix 206
5.2.5 Specification Testing 207
5.2.6 An Empirical Illustration with Exchange Rates 207
5.3 Multivariate GARCH Models 212
xii Contents
5.3.1 Alternative Model Specifications 214
5.3.2 Estimation of Multivariate GARCH Models 217
5.3.3 Extensions 218
5.3.4 Continuing the Empirical Illustration 220
6 Smooth Transition Regression Modeling 222
Timo Ter¨asvirta
6.1 Introduction 222
6.2 The Model 222
6.3 The Modeling Cycle 225
6.3.1 Specification 225
6.3.2 Estimation of Parameters 228
6.3.3 Evaluation 229
6.4 Two Empirical Examples 234
6.4.1 Chemical Data 234
6.4.2 Demand for Money (M1) in Germany 238
6.5 Final Remarks 242
7 Nonparametric Time Series Modeling 243
Rolf Tschernig
7.1 Introduction 243
7.2 Local Linear Estimation 245
7.2.1 The Estimators 245
7.2.2 Asymptotic Properties 248
7.2.3 Confidence Intervals 250
7.2.4 Plotting the Estimated Function 251
7.2.5 Forecasting 254
7.3 Bandwidth and Lag Selection 254
7.3.1 Bandwidth Estimation 256
7.3.2 Lag Selection 258
7.3.3 Illustration 261
7.4 Diagnostics 262
7.5 Modeling the Conditional Volatility 263
7.5.1 Estimation 264
7.5.2 Bandwidth Choice 265
7.5.3 Lag Selection 266
7.5.4 ARCH Errors 267
7.6 Local Linear Seasonal Modeling 268
7.6.1 The Seasonal Nonlinear Autoregressive Model 269
7.6.2 The Seasonal Dummy Nonlinear Autoregressive
Model 270
7.6.3 Seasonal Shift Nonlinear Autoregressive Model 271
Contents xiii
7.7 Example I: Average Weekly Working Hours in the United
States 272
7.8 Example II: XETRA Dax Index 280
8 The Software JMulTi 289
Markus Kr¨atzig
8.1 Introduction to JMulTi 289
8.1.1 Software Concept 289
8.1.2 Operating JMulTi 290
8.2 Numbers, Dates, and Variables in JMulTi 290
8.2.1 Numbers 290
8.2.2 Numbers in Tables 291
8.2.3 Dates 291
8.2.4 Variable Names 292
8.3 Handling Data Sets 292
8.3.1 Importing Data 292
8.3.2 Excel Format 292
8.3.3 ASCII Format 293
8.3.4 JMulTi .dat Format 293
8.4 Selecting, Transforming, and Creating Time Series 293
8.4.1 Time Series Selector 293
8.4.2 Time Series Calculator 295
8.5 Managing Variables in JMulTi 296
8.6 Notes for Econometric Software Developers 296
8.6.1 General Remark 296
8.6.2 The JStatCom Framework 297
8.6.3 Component Structure 297
8.7 Conclusion 299
References 301
Index 317