Table of Contents
Preface v
Author's Biography vii
Acronyms ix
1 Demand Estimation for Differentiated-product Markets 1
1.1 Why Demand Analysis/Estimation? 1
1.2 Review: Demand Estimation 2
1.2.1 "Traditional" approach to demand estimation 3
1.3 Discrete-choice Approach to Modeling Demand 4
1.4 Berry (1994) Approach to Estimate Demand in Differentiated Product Markets 8
1.4.1 Measuring market power: Recovering markups 14
1.4.2 Estimating cost function parameters 16
1.5 Berry, Levinsohn, and Pakes (1995): Demand Estimation Using Random-coefficients Logit Model 17
1.5.1 Simulating the integral in Eq. (1.4) 21
1.6 Applications 22
1.7 Additional Details: General Presentation of Random Utility Models 24
Bibliography 26
2 Single-agent Dynamic Models: Part 1 29
2.1 Rust (1987) 29
2.1.1 Behavioral model 29
2.1.2 Econometric model 33
Bibliography 38
3 Single-agent Dynamic Models: Part 2 39
3.1 Alternative Estimation Approaches: Estimating Dynamic Optimization Models Without Numeric Dynamic Programming 39
3.1.1 Notation: Hats and Tildes 40
3.1.2 Estimation: Match Hats to Tildes 43
3.1.3 A further shortcut in the discrete state case 43
3.2 Semiparametric Identification of DDC Models 46
3.3 Appendix: A Result for MNL Model 50
3.4 Appendix: Relations Between Different Value Function Notions 52
Bibliography 53
4 Single-agent Dynamic Models: Part 3 55
4.1 Model with Persistence in Unobservables ("Unobserved State Variables") 55
4.1.1 Example: Pakes (1986) patent renewal model 55
4.1.2 Estimation: Likelihood function and simulation 58
4.1.3 "Crude" frequency simulator: Naive approach 59
4.1.4 Importance sampling approach: Particle filtering 60
4.1.5 Nonparametric identification of Markovian Dynamic Discrete Choice (DDC) models with unobserved state variables 64
Bibliography 71
5 Dynamic Games 73
5.1 Econometrics of Dynamic Oligopoly Models 73
5.2 Theoretical Features 74
5.2.1 Computation of dynamic equilibrium 76
5.3 Games with "Incomplete Information" 77
Bibliography 79
6 Auction Models 81
6.1 Parametric Estimation: Laffont-Ossard-Vuong (1995) 81
6.2 Nonparametric Estimation: Guerre-Perrigno-Vuong (2000) 85
6.3 Affiliated values Models 88
6.3.1 Affiliated PV models 88
6.3.2 Common value models: Testing between CV and PV 90
6.4 Haile-Tamer's "Incomplete" Model of English Auctions 92
Bibliography 94
7 Partial Identification in Structural Models 95
7.1 Entry Games with Structural Errors 96
7.1.1 Deriving moment inequalities 98
7.2 Entry Games with Expectational Errors 99
7.3 Inference Procedures with Moment Inequalities/Incomplete Models 100
7.3.1 Identified parameter vs. identified set 100
7.3.2 Confidence sets which cover "identified parameters" 101
7.3.3 Confidence sets which cover the identified set 103
7.4 Random Set Approach 105
7.4.1 Application: Sharp identified region for games with multiple equilibria 106
Bibliography 107
8 Background: Simulation Methods 109
8.1 Importance Sampling 110
8.1.1 GHK simulator: Get draws from truncated, multivariate normal (MVN) distribution 110
8.1.2 Monte Carlo integration using the GHK simulator 113
8.1.3 Integrating over truncated (conditional) distribution F(x
8.2 Markov Chain Monte Carlo (MCMC) Simulation 115
8.2.1 Background: First-order Markov chains 116
8.2.2 Metropolis-Hastings approach 117
8.2.3 Application to Bayesian posterior inference 120
Bibliography 121
9 Problem Sets 123
Bibliography 134
Index 135
