Estimating and forecasting trend output and modelling labor input [Elektronische Ressource] : application of panel techniques and factor models / vorgelegt von Marcus Heribert Kappler

Informations

• Publié par: goethe_universitat_frankfurt_am_main
• Publié le: 01 janvier 2007
• Nombre de lectures: 23
• Langue: English

Extrait

Estimating and Forecasting
Trend Output and
Modelling Labor Input:
Application of Panel Techniques and
Factor Models
Inaugural-Dissertation
zur Erlangung des Doktorgrades
des Fachbereichs Wirtschaftswissenschaften
der Johann Wolfgang Goethe-Universitat¨
Frankfurt am Main
vorgelegt von
Dipl.-Volksw. Marcus Heribert Kappler
aus Stuttgart
2007Acknowledgements
First and foremost I am indebted to my supervisor Prof. Dr. Uwe Hassler for
finding interest in my research topics and for giving me considerable freedom to
write my doctoral thesis. I am also very grateful to Prof. Dr. Dieter Nautz, who
kindly consented to be my second supervisor. Furthermore, I would like to thank
Prof. Dr. Dr. h.c. mult. Wolfgang Franz, President of the Centre for European
Economic Research (ZEW), for continuously supporting this dissertation and for
providing excellent working and research conditions. My thanks also go to all my
colleagues from the ZEW Research Department Corporate Taxation and Public Fi-
nance and to the colleagues from the Research Department International Finance
and Financial Management for the pleasant and constructive working environment
and beneficial discussions. The many helpful comments from Dr. Friedrich Heine-
mann, Dr. Michael Schroder,¨ Katrin Ullrich, Steffen Osterloh, Andreas Schrimpf
and from several anonymous referees and from participants in international con-
ferences are highly appreciated. I am also very grateful to the student assistants,
notably to Qianying Chen and Heidi Hellerich, who provided able help in collecting
and processing data and in editing this thesis.
I also thank Fritz B¨ohringer and the Bohr¨ inger-Ilsfeld Foundation for providing
financial support for essential parts of this work within the framework of the project
“ZEW Growth Monitor”.
Finally, Iwouldliketoexpressmydeepgratitudetomyfamily, andinparticular
to my wife Juliane for unquestioningly supporting me throughout the work of this
doctoral thesis and for believing in me.
IIITable of Contents
List of figures VII
List of tables X
Preface 1
1 Projecting Medium-Term GDP Growth 5
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Approaches for predicting medium-term growth . . . . . . . . . . . . 8
1.3 Analysis of the production function approach . . . . . . . . . . . . . 12
1.3.1 Implementing the production function approach . . . . . . . 13
1.3.2 In-sample estimates of potential output . . . . . . . . . . . . 20
1.3.3 Multi-step forecasts and analysis of errors . . . . . . . . . . . 28
1.4 Out-of-sample results . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
1.4.1 Germany . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
1.4.2 USA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
1.4.3 United Kingdom . . . . . . . . . . . . . . . . . . . . . . . . . 43
1.4.4 France . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
1.4.5 Italy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.4.6 Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.4.7 Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
1.5 Summary and conclusion . . . . . . . . . . . . . . . . . . . . . . . . 54
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2 Estimating Trend Growth Using Panel Techniques 63
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.2 The sources of economic growth . . . . . . . . . . . . . . . . . . . . . 64
2.2.1 The proximate sources of growth . . . . . . . . . . . . . . . . 64
2.2.2 The wider sources of growth . . . . . . . . . . . . . . . . . . . 66
2.3 Theoretical embedding and empirical strategy . . . . . . . . . . . . . 70
2.3.1 Derivation of steady-state dynamics . . . . . . . . . . . . . . 71
2.4 Econometric approach and specification search . . . . . . . . . . . . 73
VVI TABLE OF CONTENTS
2.4.1 Panel integration and cointegration tests . . . . . . . . . . . . 73
2.4.2 Panel Error Correction Model (ECM) and estimation . . . . 77
2.4.3 Deriving trend output estimates: an EMVF approach . . . . 80
2.4.4 Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2.5 Data and results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
2.5.1 Results of the panel unit root tests . . . . . . . . . . . . . . . 85
2.5.2 Results of the panel ECMs . . . . . . . . . . . . . . . . . . . 87
2.5.3 Results of an extended panel ECM . . . . . . . . . . . . . . . 95
2.5.4 Results of the EMVF . . . . . . . . . . . . . . . . . . . . . . 101
2.6 Summary and conclusion . . . . . . . . . . . . . . . . . . . . . . . . 109
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3 Panel Tests for Unit Roots in Hours Worked 113
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.3 Single country analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 116
3.4 Panel analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
3.4.1 The panel unit root framework . . . . . . . . . . . . . . . . . 118
3.4.2 Cross section dependence in the panel of hours worked . . . . 120
3.4.3 Panel unit root tests for cross sectionally dependent panels . 122
3.5 Summary and conclusion . . . . . . . . . . . . . . . . . . . . . . . . 136
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Bibliography 154
Curriculum Vitae 155List of Figures
1.1 Potential and actual GDP . . . . . . . . . . . . . . . . . . . . . . . . 20
1.2 Potential and actual GDP for Germany . . . . . . . . . . . . . . . . 21
1.3 Potential output and the output gap . . . . . . . . . . . . . . . . . . 22
1.4 Output gaps from the PFA . . . . . . . . . . . . . . . . . . . . . . . 23
1.5 BMWA projections for 5-year GDP growth in Germany . . . . . . . 41
1.6 CBO and Troika projections for 5-year GDP growth in the USA . . 43
1.7 HMT projections for 3-year GDP growth in the United Kingdom . . 45
1.8 IMF projections for 3-year GDP growth in the France . . . . . . . . 47
1.9 IMF projections for 3-year GDP growth in Italy . . . . . . . . . . . 48
1.10 IMF projections for 3-year GDP growth in Japan . . . . . . . . . . . 51
1.11 IMF projections for 3-year GDP growth in Canada . . . . . . . . . . 52
2.1 Trend, actual and fitted values of GDP per capita (I) . . . . . . . . . 102
2.2 Trend, actual and fitted values of GDP per capita (II) . . . . . . . . 103
2.3 Trend, actual and fitted values of GDP per capita (III) . . . . . . . . 104
2.4 Growth of trend and actual GDP per capita (I) . . . . . . . . . . . . 105
2.5 Growth of trend and actual GDP per capita (II) . . . . . . . . . . . 106
2.6 Growth of trend and actual GDP per capita (III) . . . . . . . . . . . 107
3.1 PANIC decomposition for Japan . . . . . . . . . . . . . . . . . . . . 136
3.2 PANIC decomposition for Germany. . . . . . . . . . . . . . . . . . . 136
3.3 PANIC decomposition for Norway . . . . . . . . . . . . . . . . . . . 137
3.4 Annual hours worked per worker (I) . . . . . . . . . . . . . . . . . . 139
3.5 Annual hours worked per worker (II) . . . . . . . . . . . . . . . . . . 140
3.6 Annual hours worked per worker (III) . . . . . . . . . . . . . . . . . 141
VIIList of Tables
1.1 Labor shares from National Accounts Data . . . . . . . . . . . . . . 16
1.2 Factor share estimates from the Cobb-Douglas function . . . . . . . 18
1.3 Properties of the PFA output gap estimates . . . . . . . . . . . . . . 25
1.4 Size properties of Newey-West based tests of forecast unbiasedness . 31
1.5 Size properties of tests for equal forecasts accuracy . . . . . . . . . . 36
1.6 Results of forecast evaluation for Germany . . . . . . . . . . . . . . . 40
1.7 Results of forecast evaluation for the USA . . . . . . . . . . . . . . . 42
1.8 Results of forecast evaluation for the United Kingdom . . . . . . . . 44
1.9 Results of forecast evaluation for France . . . . . . . . . . . . . . . . 46
1.10 Results of forecast evaluation for Italy . . . . . . . . . . . . . . . . . 49
1.11 Results of forecast evaluation for Japan . . . . . . . . . . . . . . . . 50
1.12 Results of forecast evaluation for Canada . . . . . . . . . . . . . . . 53
2.1 Summary of panel unit root tests . . . . . . . . . . . . . . . . . . . . 86
2.2 PMG estimates and cointegration test results for various models (I) 89
2.3 PMG es and coin test results for various models (II) 90
2.4 PMG and MG estimation results for an extended ECM . . . . . . . 97
2.5 Estimates of the error correction coefficients φ . . . . . . . . . . . . 98i
2.6 Orders of lags in the ARDL model . . . . . . . . . . . . . . . . . . . 99
2.7 Diagnostic statistics for the extended panel ECM . . . . . . . . . . . 99
2.8 Alternative PMG and MG estimation results of the extended ECM
model for different ARDL specifications . . . . . . . . . . . . . . . . 100
2.9 Averages of trend growth and actual growth of GDP per capita . . . 109
2.10 Description of data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2.11 Correlation matrix of the explanatory variables . . . . . . . . . . . . 112
3.1 Individual ADF(l ) test statistics . . . . . . . . . . . . .