Industrial Organization and Experimental Economics

Publish in



Please download to get full document.

View again

of 38
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Industrial Organization and Experimental Economics. Huanren (Warren) Zhang. Why lab experiments?. Test theoretical prediction in controlled experiments In the field all kinds of factors intertwined together, hard to pin down determinative factors
Industrial Organization and Experimental EconomicsHuanren(Warren) ZhangWhy lab experiments?
  • Test theoretical prediction in controlled experiments
  • In the field all kinds of factors intertwined together, hard to pin down determinative factors
  • When without theories, observed regularities in lab experiments can help discover new theories
  • Two Papers to be Presented
  • I. Firm Behavior
  • Transparency in Markets for Experience Goods (Bastian Henze and Florian Schuett)
  • II. Consumer Behavior
  • Understanding Consumers' Choice of Pricing Schemes (Natalia Shestakova)I. Transparency: Motivating Example
  • European lawmakers mandate disclosure of information on Internet Service Providers’ network management (increase transparency for an experience good)
  • Theory predicts firms will voluntarily disclose the information because producer surplus is higher when quality is observable.
  • Firm Behavior: Relevant Literature
  • (Shaked and Sutton, 1982): When quality is observable, firms are predicted to engage in vertical differentiation in order to relax price competition.
  • (Akerlof, 1970): When consumers cannot observe quality, firms will supply inefficiently low quality
  • So the authors want to…
  • Investigate the role of transparency in the market for an experience good using a lab experiment
  • To do that, they
  • Vary the degree to which consumers are informed about quality (four different treatments) and compare the results
  • Model:
  • Two firms simultaneously choose a level of quality
  • Observing each other’s quality level, firms simultaneously post a price
  • consumers make purchase decision
  • Model
  • Firm ’s per-unit cost of providing quality is .
  • Consumer ’s utility function
  • Marginal value for quality is normally distributed over .
  • Four Treatments
  • no-info: none of the buyers observe the quality on offer
  • full-info: all buyers perfectly observe quality
  • subset: half of the buyers perfectly observes quality while the other half does not
  • signal: all buyers observe a signal about quality where is uniformly distributed on *
  • Parameters
  • Marginal cost of quality
  • Consumer’s taste distribution
  • Fixed component of consumer utility
  • Number of consumers
  • Support of signaling noise distribution *
  • Quality levels
  • Theoretical Prediction: Full Information
  • Solve the game backward:
  • Assume (without loss of generality)
  • Marginal consumer with taste for quality
  • Theoretical Prediction: Full Information
  • Suppose . Denote
  • Probability that prefers B is
  • Probability that prefers A is
  • Expected demand
  • Theoretical Prediction: Full Information
  • At the pricing stage, firm j solves
  • Bes responses:
  • Theoretical Prediction: Full Information
  • Solving for the Nash equilibrium prices
  • Theoretical Prediction: Full Information
  • Profit functions
  • ==(Note: the high-quality firm’s profit is increasing in its own quality whatever its rival’s quality and the low-quality firm’s profit is decreasing in its own qualityTheoretical Prediction: Full Information
  • Two subgame perfect pure-strategy Nash equilibrium:
  • Firm A offers highest possible quality while firm B offers the lowest
  • Firm B offers highest possible quality while firm A offers the lowest
  • For the given parameters, the equilibria are and .
  • Theoretical Prediction: No information
  • Perfect Bayesian equilibrium (PBE):
  • players’ strategies must be optimal given beliefs
  • beliefs must be derived from equilibrium strategies using Bayes’ rule whenever possible.
  • Theoretical Prediction: No information
  • One pooling PBE:
  • Both firms choose the lowest quality and price at marginal cost .
  • Consumers believe the quality for the products on offer is at the lowest level
  • Consumers’ out-of-equilibrium beliefs: any firm choosing a different price also has the lowest quality.
  • Theoretical Prediction: No information
  • No theory has addressed whether there can be a signaling equilibrium when firms choose from a set of more than two possible levels of quality.
  • But it is unlikely that separation would remain an equilibrium when there are more than two levels of quality.
  • Theoretical Prediction: Subset
  • The literature has not analyzed the case of multiple quality levels.
  • The presence of informed consumers makes the existence of a separating equilibrium, where prices signal qualities, more likely.
  • The informed consumers exert a positive informational externality on uninformed consumers: creating a cost of mimicking high-quality firms.
  • Theoretical Prediction: Imperfect Signal
  • No theories exist dealing with the case of imperfect information about quality.
  • Intuitively we would expect a similar argument as in the subset treatment to apply, and the theoretical predictions for both treatments would be similar.
  • Experimental Design
  • 30 periods followed 2 practice periods
  • Subjects randomly assigned to be either a seller or a buyers, and the roles remain the same for the entire experiment
  • Each session has three groups of sellers and three groups of buyers.
  • With probability of 1/3, sellers are rematchedwith buyers in each period.
  • Experimental Design
  • Fully informed buyers (full-info and subset treatments) were played by the computer.
  • full-info: 4 sessions, 6 subjects each (12 automated buyers)
  • subset: 5 sessions, 12 subjects each (6 automated buyers)
  • no-info : 3 sessions, 18 subjects each
  • signal: 4 sessions, 18 subjects each
  • ResultsResultsResultsConclusion: Transparency
  • Firms do not differentiate quality under full information.
  • Under no information, quality is low. At the same time, firms manage to maintain prices substantially above marginal cost*.
  • In the subset and signal treatments, quality is significantly above the no-information level
  • Consumer’s surplus increases with the increase of quality transparency
  • II. Consumer Choice: Motivating Example
  • Suppose you consume 80 minutes/month, which plan will you choose?
  • Consumer Choice: Background
  • Traditional economics assumes people are rational (like a powerful calculating machine)
  • Experimental economics of individual behavior is closely related to behavioral economics
  • Behavioral economics claims that people’s rationality is bounded
  • Instead of calculating the optimal solution, people use heuristics (rules of thumb) to find a satisfactory solution
  • The Model
  • Pricing Scheme
  • Demand uniform distributed over
  • Consumer Problem
  • The Model: Illustration“Expected Demand” Heuristic (EDH)
  • Instead of minimizing expected expenditure,
  • where   is the expected demand.
  • Prediction EDH: following expected demand heuristic, consumers tend to choose the scheme with an expenditure function which is nonlinear over the demand range, even when this scheme is not first-best.
  • EDH: IllustrationResultsResultsResults
  • Result 1: n the 27 main experimental tasks, only 55.6% of choices are first-best.
  • Result 2: Expected demand heuristic is the best in predicting subjects’ choices.
  • Match heuristic gives reasonable prediction.
  • Conclusion: Consumer Choice
  • The “irrationalities” observed in experiments on individual decision-making gives rise to behavioral economics
  • People use heuristics to speed up the decision-making process, but this sometimes gives suboptimal solution
  • Final Comments
  • Experiments can test the predictive power of a theory in a controlled environment
  • When no relevant theories are available, experiments can give valuable empirical implication
  • Empirical regularities observed in lab experiments can help improve current theories or develop new theories
  • Related Search
    Related Documents
    View more...
    We Need Your Support
    Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

    Thanks to everyone for your continued support.

    No, Thanks