Applications of point processes in empirical economics and finance [Elektronische Ressource] / vorgelegt von Kerstin Kehrle

Informations

• Publié par: eberhard_karls_universitat_tubingen
• Publié le: 01 janvier 2010
• Nombre de lectures: 20
• Langue: English
• Poids de l'ouvrage: 6 Mo

Extrait

Applications of Point Processes in Empirical
Economics and Finance
Inaugural-Dissertation
zur Erlangung des Doktorgrades
der Wirtschaftswissenschaftlichen Fakultat¨
der Eberhard-Karls-Universitat Tubingen¨ ¨
vorgelegt von
Kerstin Kehrle
aus Augsburg
2010Dekanin: Prof. Dr. rer. pol. Kerstin Pull
Erstberichterstatter: Prof. Dr. rer. pol. Joachim Grammig
Zweitberichterstatter: Prof. Dr. rer. pol. Martin Biewen
Tag der mu¨ndlichen Pru¨fung: 16. Dezember 2009Acknowledgments
Und jedem Anfang wohnt ein Zauber inne,
der uns beschu¨tzt und der uns hilft zu leben.
A magic dwells in each beginning,
protecting us tells us how to live.
(Hermann Hesse)
Although,IofficiallystartedmyPhDinOctober2005, statisticsandeconometricssparked
my interests and fascination already during my first semesters at the University of Tubingen¨
and Katholieke Universiteit Leuven. I want to thank my undergraduate teachers Gerd
Ronning and Robert Jung for introducing me to the field, for employing me as a student
assistant and for allowing me to collect my first teaching experiences as an undergraduate
tutor.
As a PhD student I had the pleasure to present my work at various seminars and
conferences. I gratefully acknowledge financial support from the German Economic
Association, the International Institute of Forecasters and the Graduate School of Economics
at the University of Tu¨bingen. I thank all seminar and conference participants for useful
comments. Inparticular,IwanttomentionMarceloFernandes,TilmannGneiting, Alexander
Kempf and Winfried Pohlmeier. I am grateful to Martin Biewen and Rainer Scho¨bel for
serving as members in my thesis committee.
Although, many teachers taught me, no one has had such a profound influence on my
academic thinkingand workingas my supervisorJoachim Grammig. I am deeply indebtedto
him for sharing his knowledge with me and for supporting me. I thank him for working with
meonajointproject. WhatIlearnedfromhimisinvaluableandwithouthisencouragements,
iii
comments and suggestions, this PhD thesis would not have been written. I could not have
wished for a better boss, coach and coauthor.
Many people accompanied me during the years of my PhD project at the chair of
Econometrics, Statistics and Empirical Economics. I want to thank our fantastic student
assistants who always delivered more than one expected (please forgive me if I forgot you):
Irina Dyshko, Tati Figueiredo, Benjamin Friedrich, Tobias Gummersbach, Benedikt Heid,
Tobias Langen, Felix Prothmann, Jantje Sorensken, Jan Starmans, Natascha Wagner and¨
Franziska Weiss. I thank also the secretaries Sylvia Bu¨rger and Angelika Hutt. I was
fortunate to work with great colleagues: Thomas Dimpfl, Stefan Frey, Luis Huergo, Stephan
Jank, Franziska J. Peter, Peter Schmidt, Miriam Sperl, Oliver Wunsche. In particular, I owe¨
Thomas Dimpfl and Stefan Frey my gratitude for installing an excellent IT infrastructure.
SpecialthanksarealsoreservedforFranziskaJ.Peter,mydoctoralsister. Shedidnothesitate
to start a common project and put all effort and passion in writing a joint paper with me.
I really appreciated working with her and hope that there will be future prospering projects
and the same unspoken thinking and understandingbetween us. I am very grateful to Oliver
Wunsche, who always gave me the feeling of being understood, who provided a shoulder to¨
cry on, who comforted me and confirmed my self-confidence in times I doubted, and who
made melaugh - also whenthere wasnothingtolaugh about. I cannot imagine that anybody
under the sun could possibly have a better team than I had during the time I wrote this
thesis.
Beyond academia there exists another world. I am grateful to my friends who are simply
there for me. Thanks Ina, Lisa, Manu, Solveig and Tina. Finally, I thank my family. My
parents, Anneliese and Karl and my brother, Jan-Michael. What would I be without them?
TheyarethebestandIthankthemfortheirconstantlove, indestructibletrustandinvaluable
advice.
Often, PhDstudents- includingme - livea life ona rollercoaster that is markedbyfailure
and success and some more failure. Those who are able to share these ups and downs with
others are lucky - lucky like me.
Kerstin Kehrle, Tubingen, 15. January 2010¨Contents
Acknowledgments i
Contents iii
List of Figures vi
List of Tables vii
1 Introduction 1
2 A Model for the Federal Funds Rate Target 7
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Institutional Details and Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Econometric Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.1 The ACH-ACM Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.2 Evaluating Probability Function Forecasts of DMPP Models . . . . . . 19
2.4 Estimation Results and Diagnostic Checks . . . . . . . . . . . . . . . . . . . . 22
2.4.1 Empirical Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4.2 Estimation Results and Goodness of Fit . . . . . . . . . . . . . . . . . 24
2.4.3 Comparing Short Term Interest Rate Forecasts . . . . . . . . . . . . . . 27
2.4.4 In-sample Probability Forecasts . . . . . . . . . . . . . . . . . . . . . . . 29
2.4.5 Out-of-sample Forecast Evaluation . . . . . . . . . . . . . . . . . . . . . 32
2.5 Conclusion and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
A.1 Four Category ACH-ACM Model . . . . . . . . . . . . . . . . . . . . . . . . . . 36
A.1.1 Four Category Estimation Results . . . . . . . . . . . . . . . . . . . . . 36
iiiCONTENTS iv
A.1.2 In- and Out-of-sample Four Category ACH-ACM Forecast Results . . . 38
A.2 Simulation of Multi-step Probability Forecasts . . . . . . . . . . . . . . . . . . 42
A.3 ACH and OP Estimation Results . . . . . . . . . . . . . . . . . . . . . . . . . 44
A.4 Bayesian Type Model Averaging . . . . . . . . . . . . . . . . . . . . . . . . . . 45
A.5 Additional Bayesian Type Model Averaging Results . . . . . . . . . . . . . . . 45
3 Forecasting Return Volatility 47
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.2 Theoretical Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.3 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.4 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.1 Econometric Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.2 Forecast Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.5 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.5.1 Estimation Results and Residual Diagnostics . . . . . . . . . . . . . . . 61
3.5.2 Density Forecast Evaluation . . . . . . . . . . . . . . . . . . . . . . . . 63
3.5.3 Out-of-sample Point Forecast Performance . . . . . . . . . . . . . . . . 68
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
B.1 Density Forecasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
B.2 Additional Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
B.3 Additional Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4 A Unique Intensity Based Information Share 77
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.2.1 The Autoregressive Conditional Intensity Model . . . . . . . . . . . . . 80
4.2.2 Impulse Response Functions and Information Shares . . . . . . . . . . . 83
4.3 The Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.4 Estimation, Information Shares and Results . . . . . . . . . . . . . . . . . . . 89
4.4.1 Estimation Results and Diagnostics . . . . . . . . . . . . . . . . . . . . 89
4.4.2 Information Shares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93CONTENTS v
C.1 Deseasonalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
C.2 Additional Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
C.3 VECM and Hasbrouck Shares . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
5 Conclusion 101
Bibliography 104List of Figures
2.2.1 Federal funds rate target, effective federal funds rate and time series of target changes. . . . 14
2.3.1 Simulation of probability forecasts for the ACH-ACM model. . . . . . . . . . . . . . . . . 21
2.4.1 Effect of a target change shock on state probabilities. . . . . . . . . . . . . . . . . . . . 27
2.4.2 Histograms of the continued PIT sequence: ACH-PSACM, in-sample forecast. . . . . . . . 31
2.4.3 Autocorrelations of the continued PIT sequence: ACH-PSACM, in-sample forecast. . . . . . 32
2.4.4 Histograms of the continued PIT sequence: ACH-DACM, out-of-sample forecast. . . . . . . 34
A.1.1 Histograms of the continued PIT sequence: four category ACH-ACM, in-sample forecast. . . 41
A.1.2 Autocorrelations of the continued PIT sequence: four category ACH-ACM, in-sample forecast. 41
A.1.3 Hist