Demand models with price endogeneity and advertising [Elektronische Ressource] / vorgelegt von Christoph Nagel

Informations

• Publié par: universitat_mannheim
• Publié le: 01 janvier 2010
• Nombre de lectures: 9
• Langue: English
• Poids de l'ouvrage: 1 Mo

Extrait

Demand Models with Price
Endogeneity and Advertising
Inauguraldissertation
zur Erlangung des akademischen Grades
eines Doktors der Wirtschaftswissenschaften
der Universit at Mannheim
vorgelegt von
Christoph Nagel
Juli 2010Abteilungssprecher: Prof. Dr. Martin Peitz
Referent: Prof. Dr. Carsten Trenkler
Korreferent: Prof. Stefan Hoderlein, Ph.D.
Tag der mundlic hen Prufung: 21. Oktober 2010
iiAcknowledgments
I would like to thank my supervisor Stefan Hoderlein for letting me follow my own
research ideas and the encouraging and fruitful feedback he gave me. Despite busy times
in Mannheim and abroad, he was very much supporting me, widened my perspective
towards econometrics and introduced me to many enlightening scholars.
I would also like to thank Carsten Trenkler for his insightful comments on my work and
his immediate willingness to join the thesis committee. Further, I would like to thank
Enno Mammen for his support and important comments.
I am especially grateful to Elu von Thadden for giving me the opportunity to spend valu-
able time at the Economics Department of the University College London, CEMMAP
and IFS, London.
I am indebted to Kai Kopperschmidt, Martin Schniedermeier and Marc Rossbach of
A.C. Nielsen for supplying the dataset used in this work.
I very much enjoyed interacting with my colleagues at the Chair for Statistics and the
graduate program of CDSE. I would like to thank Ste en Reinhold and Hannes Ullrich
for instructive comments on the third and fourth chapter of this dissertation, Christoph
Rothe for inspiring discussions about econometrics. I also thank Renate Bent for always
solving my administrative problems.
Many other colleagues made my stay in Mannheim a memorable time, especially Gon-
zague Vannoorenberghe, Daniel Harenberg, Christoph Rothe, Melanie Schienle, Michal
Kowalik, Bj orn Sass, Heiner Schumacher and Malte Hubner.
My warmest thanks go to my family, in particular to my wife and my children for sup-
porting me during the past years to make my research ideas come true.
Christoph Nagel
iiiContents
1 Introduction 1
2 Introducing a Household Panel Dataset with Consumption and Adver-
tising 5
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Data Structure, Collection and Use for Research . . . . . . . . . . . . . 6
2.2.1 Collection Process . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.2 Using the Dataset for Research . . . . . . . . . . . . . . . . . . . 11
2.2.3 Relation to other Datasets . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Data Description and Details by File . . . . . . . . . . . . . . . . . . . . 12
2.3.1 General Purchase Data . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.1.1 Data Modi cations . . . . . . . . . . . . . . . . . . . . . 13
2.3.1.2 Remarks on created Variables . . . . . . . . . . . . . . . 14
2.3.1.3 Summary Statistics . . . . . . . . . . . . . . . . . . . . . 15
2.3.2 Category Purchase Data . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.2.1 Data Modi cations . . . . . . . . . . . . . . . . . . . . . 18
2.3.2.2 Remarks on created Variables . . . . . . . . . . . . . . . 19
2.3.2.3 Summary Statistics . . . . . . . . . . . . . . . . . . . . . 20
2.3.3 Demographic Data . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3.3.1 Data Modi cations . . . . . . . . . . . . . . . . . . . . . 25
2.3.3.2 Remarks on created Variables . . . . . . . . . . . . . . . 25
2.3.3.3 Summary Statistics . . . . . . . . . . . . . . . . . . . . . 26
2.3.4 Advertising Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.3.4.1 Data Modi cations . . . . . . . . . . . . . . . . . . . . . 29
vvi CONTENTS
2.3.4.2 Remarks on created Variables . . . . . . . . . . . . . . . 30
2.3.4.3 Summary Statistics . . . . . . . . . . . . . . . . . . . . . 30
2.4 Combined Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.4.1 Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.4.2 Inferring Prices in Step 5 . . . . . . . . . . . . . . . . . . . . . . . 35
2.4.3 Remarks on created Variables . . . . . . . . . . . . . . . . . . . . 37
2.4.4 Basic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Appendix A: Tables and Figures . . . . . . . . . . . . . . . . . . . . . . . 43
Appendix B: Value Labels of Coded Variables . . . . . . . . . . . . . . . 46
3 Estimating the Magnitude of Causes for Consumer Price Endogeneity 49
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.1.1 Related Literature . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.2.1 The Empirical Model of Choice . . . . . . . . . . . . . . . . . . . 55
3.2.2 Endogeneity Correction Model . . . . . . . . . . . . . . . . . . . 60
3.2.3 Causes of Price Endogeneity . . . . . . . . . . . . . . . . . . . . 62
3.2.3.1 Industry Structure . . . . . . . . . . . . . . . . . . . . . 62
3.2.3.2 Cause 1 - Product Characteristics and Variety Informa-
tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.3.3 Cause 2 - Retail Activity . . . . . . . . . . . . . . . . . 65
3.2.3.4 Cause 3 - TV Advertising . . . . . . . . . . . . . . . . . 66
3.2.3.5 Cause 4 - Individual Inventories . . . . . . . . . . . . . . 68
3.2.3.6 Cause 5 - State Dependence . . . . . . . . . . . . . . . . 69
3.2.3.7 Point 6 - Consumer Heterogeneity . . . . . . . . . . . . 70
3.2.3.8 Point 7 - Data Aggregation . . . . . . . . . . . . . . . . 71
3.2.4 Choice and Discussion of Instrumental Variables . . . . . . . . . 71
3.2.5 Method and Interpretation . . . . . . . . . . . . . . . . . . . . . 73
3.3 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.1 Data Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74CONTENTS vii
3.3.1.1 Relation to other Datasets . . . . . . . . . . . . . . . . . 79
3.3.2 Variables of Interest in the Analysis . . . . . . . . . . . . . . . . 79
3.3.3 Construction of Instrumental Variables . . . . . . . . . . . . . . . 82
3.3.4 De nition of Estimation Samples . . . . . . . . . . . . . . . . . . 83
3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.1 Price Elasticity Results . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.2 Parameter Estimates and Estimation Details . . . . . . . . . . . 88
3.4.3 Speci cations with nominal Prices . . . . . . . . . . . . . . . . . 90
3.4.3.1 Robustness Checks . . . . . . . . . . . . . . . . . . . . 91
3.4.3.2 Comparison to PT Setup . . . . . . . . . . . . . . . . . 92
3.4.4 Speci cations with e ciency Prices . . . . . . . . . . . . . . . . . 92
3.4.5 Relation to Literature . . . . . . . . . . . . . . . . . . . . . . . . 93
3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Appendix A: Model Details - Estimation of Pooled Case . . . . . . . . . 96
Appendix B: Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Elasticity Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Parameter . . . . . . . . . . . . . . . . . . . . . . . . . 102
4 Local Advertising in a Discrete Choice Demand Model 117
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.2 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.2.1 Product dummies . . . . . . . . . . . . . . . . . . . . . . . . . . 123
4.2.2 Brand dummies . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.2.3 Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.2.4 The Simple Logit Model . . . . . . . . . . . . . . . . . . . . . . . 127
4.3 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.3.1 Data Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.3.2 Advertising . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.3.3 Construction of a Market Level Dataset . . . . . . . . . . . . . . 130
4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134viii CONTENTS
4.4.1 Simple Logit Model . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.4.2 Random Coe cients Model . . . . . . . . . . . . . . . . . . . . . 140
4.4.2.1 Results without Demographics . . . . . . . . . . . . . . 142
4.4.2.2 Results with . . . . . . . . . . . . . . . . 146
4.4.3 Discussion of Details for other Model Variants . . . . . . . . . . . 149
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Appendix A: Berry’s (1994) Model . . . . . . . . . . . . . . . . . . . . . 153
Appendix B: Tables and Figures . . . . . . . . . . . . . . . . . . . . . . . 154
Model Setup Information . . . . . . . . . . . . . . . . . . . . . . . 154
Simple Logit Model Results . . . . . . . . . . . . . . . . . . . . . 160
Random Coe cient Model Results without D