贝叶斯博弈关于信息结构的比较静态分析及应用

The economic roles of information have been actively analyzed over the past several decades, which have led to the emergence of information economics as a major new field of economics. The analysis of issues pertaining to imperfections of information, costly information acquisition, and information asymmetries has had profound implications for thinking about economic behavior and has enabled economists to explain important economic and social phenomena, which otherwise would be difficult or impossible to understand. The model of Bayesian games introduced in Harsanyi's (1966/68) seminal work provides the theoretical foundations of information economics. Bayesian games are models of interactive situations in which a player's payoff depends on his own action, the actions of the others, and in addition, on some unknown fundamentals for which the players receive exogenous partial information. The information structure of a Bayesian game is exogenous.No general comparative static analysis of Bayesian games with respect to the exogenous information structures has ever been made. Yet, changing information structure can have significant behavioral implications and profound practical implications for the design of efficient economic organizations among many other areas. Indeed, in a competitive market context, improving agents' information may destroy the opportunity of mutually beneficial trade. The phenomenon that an increase in public information may be Pareto-worsening in the context of a competitive market with diverse information, known as the Hirshleifer effect, has been extensively investigated and has had important practical applications since the publication of Hirshleifer's (1971) seminal work. Competitive economies with incomplete information are special cases of Bayesian games. The main purse of this project is to provide the missing general comparative static analysis of Bayesian games with respect to information structures, and to explore applications to principal-agent problems in industrial economics and business administration. Given that information economics is grounded on the theory of Bayesian games, comparative static analysis of Bayesian games would be timely, highly original, and generally applicable. Our expected theoretical results include establishing criteria for ranking information structures in terms of private or social values based on intrinsic structures of the underlying Bayesian games. Structural properties such as strategic complementarity, strategic substitutability, Bayesian potentials will play useful roles for our analysis. By specifying feasible actions, information structures, the rewards (payoffs) to be paid, many problems are naturally modeled as Bayesian games. Our project provides a unified framework for analyzing problems of information in the presence of strategic behavior and the efficient design of mechanisms, in particular for the principal's efficient design of incentive schemes for the agents.

海萨尼提出的贝叶斯博弈模型描述的是多人互动决策问题。不同于单人决策问题，贝叶斯博弈的每一参与方所得收益不仅依赖于本身的决策，也赖于他人的决策，以及让各参与方拥有非对称信息的不确定性。贝叶斯博弈理论奠定了信息经济学的理论基础。贝叶斯博弈的信息结构是外生变量，贝叶斯博弈关于信息结构的一般性比较静态分析至今还未出现，但信息结构的变化对有效机制设计具有深刻的理论和实践意义。例如，在竞争性市场中，增加信息可能会使参与者丧失互惠交易的机会，导致帕累托恶化(Hirshleifer效应）。本课题旨在建立贝叶斯博弈关于信息结构的比较静态分析，并探讨其在企业理论中的应用。预期主要理论结果是基于贝叶斯博弈内在结构，建立评价信息结构的标准。考虑到贝叶斯博弈理论在信息经济学中的基础地位，本课题将会对信经济学加以拓展，丰富其应用范围。

作为经济学一重要分支，信息经济学对思考经济行为，解释经济社会现象等等具有深刻的影响。而贝叶斯博弈及贝叶斯-纳什均衡理论奠定了信息经济学的理论基础并提供标准的理论分析工具。在贝叶斯博弈中，各参与方所得收益即依赖于本身的决策，也依赖于其他参与方的决策，以及所有参与方对有关博弈的不确定性所拥有的信息。信息结构的变化如何影响参与方的行为对有效机制设计具有深刻的理论及实践意义。考虑到贝叶斯博弈及贝叶斯-纳什均衡理论在信息经济学中的基础地位，为拓展信息经济学并丰富其应用范围，本项目旨在建立评价贝叶斯博弈信息结构的信息性之标准，建立贝叶斯博弈关于信息结构的比较静态分析方法，探讨其在产业经济理论的应用，包括不完全信息下金融风险传导、不完全信息下企业交叉持股以及不同信息结构下的信息披露策略等问题。最后，鉴于势函数对计算完全信息下的纳什均衡的极大简化作用，本项目借助势函数在非对称信息下的推广来增强贝叶斯博弈及贝叶斯-纳什均衡理论的可应用性。. 项目组成员目前已在国际经济学SSCI期刊发表论文3篇，返修论文2篇，在投审稿论文4篇，部分论文结论被国际知名学者与顶级期刊引用，具有较高的学术价值。同时，项目组成员在国内外顶级会议做若干次学术报告，举办了4次学术会议，取得了较好的学术反响。本项目所获成果具有重要的国际学术价值，丰富和发展了非对称信息结构下的贝叶斯博弈及贝叶斯-纳什均衡理论及其应用。本项目提出的一些新概念、新模型、新方法，对分析与应用贝叶斯博弈及贝叶斯-纳什均衡理论带来极大的便利，具有重要的理论意义。同时对于不同信息结构下企业行为的分析，对国企改革、反垄断政策研究与监管等也都具有重要的应用价值。