Files in this item
|(no description provided)|
|Title:||Information transmission in dynamic Bayesian duopoly models|
|Author(s):||Urbano, Maria Amparo|
|Doctoral Committee Chair(s):||Mirman, L.J.|
|Department / Program:||Economics|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Abstract:||In economic models in which agents are asymmetrically informed about the structural parameters of the economy the acquisition or manipulation of information plays a crucial role. The incentive to affect the flow of information is especially important in models in which choice variables, and hence market variables, generate information used for future decisions. In this paper we study two different reasons why agents might change their (myopically) optimal decisions when they take account of the informational content of their decisions. The first is when informed agents manipulate the informational content of observed market variables through their own decisions in order to influence the learning of uninformed agents. The second is when uninformed agents "experiment" in order to influence the flow of information on which their own future decisions are based.
This thesis presents two models in which both of these possibilities are present. We consider for both models a duopolistic market characterized by a linear stochastic demand function with an unknown parameter. The information about demand is asymmetric. It is assumed that there is one informed firm and one uninformed firm. The informed firm knows the expected value of the demand parameter and the uninformed firm has subjective probability distribution over the possible values of the mean of the unknown parameter.
We solve the two-period model duopoly model as a dynamic game of incomplete information. Firms choose their output level in the first period based upon their priors and their understanding of how information is used to update their priors. After the first period firms, having observed the market signals, use this information to update their priors on the unknown parameter. The solution concept we employ is the Bayesian Nash equilibrium (sequential equilibrium) of Harsanyi (H) and study the properties of a separating equilibrium.
The essential difference between the models is in the information structure of them. In the second model, only prices are observed. In the first model, both prices and quantities are observed; however, output is random.
|Rights Information:||Copyright 1989 Urbano, Maria Amparo|
|Date Available in IDEALS:||2011-05-07|
|Identifier in Online Catalog:||AAI8916317|