Handbook of Industrial Organization

Contributors
Introduction to the series
Preface
1 Foundations of demand estimation
	1 Introduction
		1.1 Why estimate demand?
		1.2 Our focus
	2 The challenges of demand estimation
		2.1 The first fundamental challenge
		2.2 The second fundamental challenge
		2.3 Demand is not regression
		2.4 A surprisingly difficult case: exogenous prices
		2.5 Many common tools fall short
			2.5.1 Controls, including fixed effects
			2.5.2 Control function
			2.5.3 Average treatment effects
		2.6 Balancing flexibility and practicality
		2.7 Demand or utilities?
	3 Discrete choice demand
		3.1 Random utility models
		3.2 The canonical model
		3.3 Why random coefficients?
	4 Market-level data
		4.1 The BLP estimator
		4.2 Instruments
			4.2.1 Cost shifters and their proxies
			4.2.2 BLP instruments
			4.2.3 Waldfogel-Fan instruments
			4.2.4 Exogenous measures of market structure
			4.2.5 Optimal instruments
			4.2.6 Evaluating instruments
		4.3 Using a supply side
		4.4 Computing the BLP estimator and standard errors
	5 Nonparametric identification: market-level data
		5.1 Insights from parametric models
			5.1.1 Multinomial logit
			5.1.2 Nested logit
			5.1.3 The BLP model
			5.1.4 Index, inversion, and instruments
		5.2 Nonparametric demand model
			5.2.1 A nonparametric index
			5.2.2 Inverting demand
		5.3 Identification via instruments
		5.4 Discussion
			5.4.1 Why 2J instruments?
			5.4.2 Why BLP instruments?
			5.4.3 Why the index?
			5.4.4 Further restrictions and tradeoffs
	6 Micro data, panels, and ranked choices
		6.1 Micro data
		6.2 Consumer panels
		6.3 Ranked choice data
		6.4 Hybrids
	7 Nonparametric identification with micro data
		7.1 Nonparametric demand model
		7.2 Identification
			7.2.1 Identification of the index function
			7.2.2 Identification of demand
		7.3 Discussion
	8 Some directions for future work
	References
2 Empirical models of demand and supply in differentiated products industries
	1 Introduction
	2 A motivating example
		2.1 Model
			2.1.1 Supply
			2.1.2 Demand
		2.2 Estimation and results
		2.3 Discussion
	3 Demand
		3.1 Background
		3.2 Discrete choice demand models
			3.2.1 Price elasticity and substitution patterns
			3.2.2 Consumer welfare
	4 Demand estimation
		4.1 The estimation problem
		4.2 What variation in the data can identify the parameters?
			4.2.1 Intuition from individual-level data
			4.2.2 The informational content of E[ξ|Z]=0
		4.3 The general estimation procedure
			4.3.1 Instrumental variables
				BLP instruments
				Hausman instruments
				Waldfogel instruments
			4.3.2 Additional sources of variation
				Multiple markets
				Micro moments and second choice data
				Supply-side moments
			4.3.3 Efficiency
			4.3.4 Computational algorithms
				Nested fixed point
				Mathematical programming with equilibrium constraints (MPEC):
				Approximate BLP (ABLP):
		4.4 Extensions
			4.4.1 Error in market shares
			4.4.2 Non-parametric and flexible estimation
	5 Supply
		5.1 The workhorse model of horizontal competition
		5.2 Distinguishing between models of competition
		5.3 Adding retailers into the mix
		5.4 Models of bargaining
	6 Extensions of the demand model
		6.1 Extensions to the static demand model
			6.1.1 Multiple goods
			6.1.2 General characteristics demand models
		6.2 Dynamic demand
			6.2.1 Storable products
			6.2.2 Durable products
	7 Concluding comments
	References
3 An industrial organization perspective on productivity
	1 A productivity primer
		1.1 Background and focus
		1.2 Productivity conceptualized
	2 Empirical facts about productivity at the producer level
		2.1 Dispersion
		2.2 Persistence within producers
		2.3 Correlations
	3 A simple model of equilibrium productivity dispersion
		3.1 Demand
		3.2 Supply
		3.3 Equilibrium
		3.4 Empirical implications
	4 Measurement of output and inputs
		4.1 Output measurement
		4.2 Input measurement
		4.3 Data sources
	5 Recovering productivity from the data
		5.1 Operating environment and unit of analysis
			5.1.1 Market structure
			5.1.2 Unit of analysis
			5.1.3 Output and input data
			5.1.4 Trade-offs across approaches
			5.1.5 Notation and setup
		5.2 Factor shares
		5.3 Production function estimation (producer level)
			5.3.1 Perfect competition (A.1)
				Control function approach
				Selection bias
				Procedure
				Dynamic panel
				Discussion
			5.3.2 Imperfect competition (B.1)
				Homogeneous good
				Product differentiation
				Deflating revenue
				Adding demand-side information
				Pass-through
				Beyond price data: how to compare quantities?
			5.3.3 Impact on the coefficients of interest
		5.4 Multi-product production
			5.4.1 Allocation of inputs to products
				Explicit aggregation from product to producer level
			5.4.2 Estimate transformation function (A.2)
			5.4.3 Product differentiation and imperfect competition (B.2.2)
				Illustration
		5.5 Cost versus production functions
		5.6 Measurement and specification errors
			5.6.1 Measurement error
			5.6.2 Model misspecification
				Productivity process
				Technology heterogeneity
				Functional form
	6 Productivity analysis
		6.1 Producer-level productivity analysis
			6.1.1 Identifying producer-level drivers
				Exogenous drivers
				Endogenous drivers
			6.1.2 Sources of productivity differences
				Managerial practices
				Unobservable input quality
				Intangible capital
				Firm structure
				Product-side differences
		6.2 Aggregate analysis: resource (re/mis)allocation
			6.2.1 What does theory predict?
			6.2.2 Empirical work
				Decomposing industry aggregate productivity
			6.2.3 Exogenous drivers: reallocation
				Deregulation
				Technology
			6.2.4 Endogenous drivers and aggregation: market power
		6.3 Misallocation
	7 Looking ahead
		7.1 Market power and productivity data
			7.1.1 Measuring market power using production data
				Applications
			7.1.2 Integrating product and factor markets using productivity data
				Vertical linkages
				Labor market power
		7.2 Technological change and market-level outcomes
			7.2.1 Factor-biased technological change
			7.2.2 Endogenous productivity growth
	8 Conclusion
	References
4 Dynamic games in empirical industrial organization
	1 Introduction
		1.1 Role of dynamic games in empirical industrial organization
		1.2 Organization of this chapter
	2 Models
		2.1 Basic framework
		2.2 Markov perfect Nash equilibrium
			2.2.1 Definition
			2.2.2 Equilibrium existence
			2.2.3 Incomplete information
			2.2.4 Multiple equilibria
		2.3 Examples
		2.4 Extensions of the basic framework
			2.4.1 Continuous time
			2.4.2 Oblivious equilibrium
			2.4.3 Large state spaces
			2.4.4 Persistent asymmetric information
			2.4.5 Firms\' biased beliefs
	3 Identification and estimation
		3.1 Data
		3.2 Identification
			3.2.1 Non-identification result
			3.2.2 A set of sufficient conditions for identification
			3.2.3 Relaxing restrictions (ID.1) to (ID.8)
			3.2.4 Identification of mixed continuous-discrete choice models
		3.3 Estimation methods
			3.3.1 Full solution methods
			3.3.2 Two-step CCP methods
			3.3.3 Bajari-Benkard-Levin (BBL) method
			3.3.4 Large state space and finite dependence
			3.3.5 Unobserved market heterogeneity
		3.4 The promise of machine learning
	4 Empirical applications
		4.1 Earlier empirical work on dynamic games
			4.1.1 Competition in the hospital market
			4.1.2 Dynamic output competition with learning by doing
			4.1.3 Dynamics in auctions
			4.1.4 Environmental regulations in concentrated industries
			4.1.5 Demand shocks and market structure
			4.1.6 Subsidizing entry
		4.2 Innovation and market structure
			4.2.1 Microprocessor innovation: Intel vs AMD
			4.2.2 Hard drive innovation: new products and cannibalization
			4.2.3 Car innovation and quality ladders
			4.2.4 Data on innovation
		4.3 Antitrust policy towards mergers
			4.3.1 Endogenous mergers
			4.3.2 Evolving market structure and mergers
			4.3.3 Revealed merger efficiencies
		4.4 Dynamic pricing
			4.4.1 Competition with price adjustment costs
			4.4.2 Limit pricing
			4.4.3 Dynamic pricing with network effects
		4.5 Regulation
			4.5.1 Environmental regulation
			4.5.2 Land use regulation
			4.5.3 Product variety
			4.5.4 Industrial policy
		4.6 Retail
			4.6.1 Economies of density and cannibalization
			4.6.2 Chains
			4.6.3 Unobserved heterogeneity and entry in retail
			4.6.4 Effect of Walmart on rival grocers
			4.6.5 Exit in declining industries
			4.6.6 Repositioning
			4.6.7 Advertising
		4.7 Uncertainty and firms\' investment decisions
			4.7.1 Firm investment under uncertainty
			4.7.2 Uncertainty and oil drilling in Texas
			4.7.3 Uncertainty in shipping
		4.8 Network competition in the airline industry
		4.9 Dynamic matching
		4.10 Natural resources
	5 Concluding remarks
	References
5 Moment inequalities and partial identification in industrial organization
	1 Introduction
	2 Definitions and background
	3 Revealed preference
		3.1 Primitive assumptions
		3.2 Paths to estimators
		3.3 Examples
			3.3.1 Richer assumptions on disturbances
	4 Generalized discrete choice approaches
		4.1 Models of discrete games with complete information
		4.2 Simple game example
		4.3 Using both necessary and sufficient conditions for Nash equilibrium
			4.3.1 Empirical applications
			4.3.2 Assumptions on information
		4.4 Models of auctions
		4.5 Alternative assumptions
	5 Estimation and inference
		5.1 Estimation
		5.2 Overview of inference
		5.3 Moment inequality approach
		5.4 Criterion function approach
		5.5 Random set approach
		5.6 Bayesian approach
	6 Implementation of partial identification
		6.1 Computational considerations
		6.2 Simulation based approaches
		6.3 Reporting empirical results from a partially identified model
			6.3.1 The overall identified set, marginal identified sets, or the identified sets for objects of interest
			6.3.2 Counterfactuals in models with incompleteness and/or multiple equilibria
	7 Conclusions
	References
6 Frictions in product markets
	1 Introduction
	2 Transaction costs
		2.1 Vertical differentiation
		2.2 Market power and secondary markets
		2.3 Empirical research
		2.4 Role of intermediaries in overcoming transaction costs
	3 Asymmetric information
		3.1 Theory
			3.1.1 Static adverse selection, exogenous ownership
			3.1.2 Dynamic trading with exogenous initial ownership
			3.1.3 Endogenous initial ownership
		3.2 Empirics
		3.3 Role of intermediaries in overcoming asymmetric information
	4 Search frictions
		4.1 Theoretical models
			4.1.1 Information clearinghouse
			4.1.2 Simultaneous search
			4.1.3 Sequential search
			4.1.4 Sellers\' non-price behavior
			4.1.5 Product differentiation
		4.2 Empirical contributions
			4.2.1 Descriptive papers
			4.2.2 Estimation of simultaneous-search models
			4.2.3 Estimation of sequential-search models
		4.3 The role of intermediaries in search markets
			4.3.1 Theoretical contributions
			4.3.2 Empirical contributions
	5 Matching frictions
		5.1 The role of intermediaries in matching markets
	References
7 Two-sided markets, pricing, and network effects
	1 Introduction
		1.1 Terminology and background
	2 Monopoly
		2.1 Basic framework and notation
		2.2 Profit-maximizing prices
			2.2.1 Homogeneous interaction benefits (Armstrong, 2006)
			2.2.2 Homogeneous stand-alone benefits (Rochet and Tirole, 2003)
		2.3 Welfare-maximizing prices
		2.4 Distortions
		2.5 Miscellaneous: chicken & egg problem, non-negative prices, distortionary taxation
			2.5.1 Chicken & egg
			2.5.2 Vertical integration
			2.5.3 Non-negative prices
			2.5.4 Distortionary taxation and two-part tariffs
		2.6 Dynamic pricing
	3 Competition for the market
		3.1 Divide-and-conquer strategies
		3.2 Congestion within sides
		3.3 Multi-homing
		3.4 Dynamic competition
	4 Competition on the market
		4.1 Single-homing
			4.1.1 Homogeneous interaction benefits
			4.1.2 Dispersed information
			4.1.3 Heterogeneous interaction benefits
		4.2 Multi-homing
			4.2.1 Competitive bottleneck
			4.2.2 Multi-homing on both sides
		4.3 Concentration, merger and collusion
			4.3.1 Concentration and entry of platforms
			4.3.2 Horizontal mergers
			4.3.3 Collusion
		4.4 Exclusivity and bundling
			4.4.1 Exclusivity
			4.4.2 Bundling
	5 Alternative modeling of competition, coordination and beliefs
		5.1 Richer price structures
		5.2 Quantity competition
		5.3 Passive beliefs
	6 Matching design
		6.1 Second-degree price discrimination and matching design
			6.1.1 One-to-one matching
			6.1.2 Many-to-many matching
		6.2 Targeting and third-degree price discrimination
		6.3 Dynamic arrivals and evolving private information
	7 Identification of network effects
		7.1 Direct network effects
			7.1.1 Base model
			7.1.2 Consumer heterogeneity
			7.1.3 Contextual effects
		7.2 Some solutions for direct network effects
			7.2.1 Random assignment
			7.2.2 Heterogeneous networks
			7.2.3 Nonlinear models
			7.2.4 Dynamics
			7.2.5 Variance
			7.2.6 Other approaches
		7.3 Indirect network effects
	8 Estimating indirect network effects
		8.1 Dynamics
		8.2 Exclusions in two-sided markets
	9 Empirical work on pricing in platform studies
		9.1 Price as endogenous
		9.2 Price as exogenous
		9.3 Matching
	10 Conclusion
	References
8 Information markets and nonmarkets
	1 Introduction
	2 Buying and selling information
		2.1 Value of information and experiment
		2.2 Selling experiments
		2.3 Selling realizations
		2.4 Information products
		2.5 Returns from information and data
	3 Markets for information
		3.1 Value of information in a normal quadratic environment
		3.2 Selling information to competing firms
		3.3 Information sharing among competing firms
		3.4 Data intermediation
		3.5 Data markets and data externalities
	4 Instruments to trade and monetize information
		4.1 Third-degree price discrimination
		4.2 Ratchet effect and privacy
		4.3 Ratings and recommender systems
		4.4 Certification and expert markets
	5 Forecasting and aggregation of predictions
		5.1 Forecasting nonmarkets
			5.1.1 Scoring rules
			5.1.2 Forecasting contests
			5.1.3 Reputational forecasting
			5.1.4 Expert aggregation
		5.2 Prediction markets
			5.2.1 Equilibrium with heterogeneous beliefs and limited wealth
				Manski bounds
			5.2.2 Equilibrium with heterogeneous beliefs and risk-averse traders
				Logarithmic preferences
				CARA preferences
			5.2.3 Aggregation of information and beliefs
				Bounded wealth
				Decreasing absolute risk aversion
				Revelation of private information under rational expectations
				Link to no-trade theorem
		5.3 Automated market maker
		5.4 Performance of prediction markets
	6 Science as nonmarket
		6.1 Organization of science
		6.2 Measuring influence
		6.3 Selection, publication, and funding
		6.4 Economics of statistical inference
	7 Conclusion
	References
9 Structural empirical analysis of contracting in vertical markets
	1 Introduction
	2 Theory
		2.1 Multilateral settings with externalities
			2.1.1 Non-cooperative bargaining models
				The offer game
					Public offers
					Private offers
				The bidding game
				Other non-cooperative bargaining processes and more general vertical structures
			2.1.2 Nash-in-Nash bargaining
				Extensions of the Nash-in-Nash concept
				Contracting dynamics
	3 An empirical framework
		3.1 Supply: modeling vertical contracting
			3.1.1 Preliminaries: contracts, actions, and payoffs
				Contracts
				Payoff-relevant actions
				Payoffs
			3.1.2 Modeling contract formation
				Offer games
				Nash-in-Nash bargaining
				Remarks on timing
		3.2 Supply: estimation and identification
		3.3 Demand
			3.3.1 Bundles and usage models
			3.3.2 Upstream choice only
			3.3.3 Consumer selection
			3.3.4 Joint estimation of demand with supply
	4 Recent applications
		4.1 Horizontal mergers in vertical markets
		4.2 Effects of vertical integration and mergers
		4.3 Price discrimination
		4.4 Nonlinear contracts
		4.5 Exclusive vertical contracts
	5 Concluding remarks
	References
Index