基于次可加测度的代理模型及应用

发布时间：2018-06-05 03:07

本文选题：代理理论 + 可信性测度　；参考：《天津大学》2014年博士论文

【摘要】：在经济管理领域的各个方面普遍存在着委托代理问题,在这些委托代理问题中由于诸多不确定因素的影响不可避免地存在着不完全信息,而不完全信息的存在又从根本上对委托人和代理人的最优决策产生着不可忽视的影响.为了降低信息不完全性对决策者最优收益的不利影响,则需要事前在委托人与代理人之间签订最优合同,该合同不仅可以对委托人与代理人的预期行为加以规范,而且还可以对代理人进行有效的事前激励.因此,在不完全信息下寻找激励的影响因素并设计最优合同已经成为当今国际研究的热点前沿问题之一. 不完全信息中通常包含具有主观特性的不确定因素,同时Ellsberg悖论也已表明在某些特定情形下,决策者所做决策与基于具有可加性的概率测度的决策准则并不一致.因此,本文将利用具有次可加性的可信性测度和不确定测度来度量具有主观特性的不完全信息.在此基础上,重点研究私人信息和公共未知信息下的委托代理问题,并分析决策者对待风险的主观态度及信息多样性对最优合同的影响.本文的主要研究内容包括以下五个方面: 1)基于可信性测度,研究代理人的能力是其私人信息时的委托代理问题.由于委托人无法观测到代理人的能力,故将委托人对代理人能力的估计刻画为模糊变量.为了激励代理人在合作过程中按照自己的利益取向进行决策,委托人须根据代理人的能力特征设计最优的转移支付.于是,从委托人的角度出发,基于可信性测度建立了私人信息下的代理模型,分别给出了当委托人是风险规避和风险中性时的最优合同,并分析了委托人对待风险的主观态度对最优合同的影响.由于代理人具有信息优势,故其在合作过程中能够从委托人处攫取一定的信息租金,但其对待风险的主观态度对最优合同没有影响.最后,将所得结果应用于供应链采购系统中,其结果表明该最优合同可以实现对代理人的有效激励. 2)基于可信性测度,研究代理人的能力为公共未知信息时的委托代理问题.由于委托人和代理人处于同等的信息状态,故两者合作的目的是为了通过积极有效的合作发掘代理人潜在的能力,从而尽可能多地提高双方的收益.于是,委托人根据代理人为其带来的收益状况设计最优的转移支付,进而基于可信性测度建立公共未知信息下的代理模型,并分别给出当代理人是风险规避和风险中性时的最优合同.因为处于同等的信息状态,所以代理人无法再从委托人处攫取信息租金.同时,由于公共未知信息的不可预测性,决策者的主观态度对最优合同均没有影响.最后,基于可信性测度将私人信息下的最优合同与公共未知信息下的最优合同进行了比较,突出了两种情形下最优合同的不同之处. 3)基于不确定测度,研究多重公共未知收益信息下的委托代理问题,重点分析代理人是否采取与委托人相同的决策准则对最优合同的影响.由于公共未知信息的不可观测性,故将代理人对未知信息的估计刻画为相互独立的不确定变量.为了探究代理人是否采取与委托人相同的决策准则对最优合同的影响,本文在两种情形下分别建立了不确定期望值代理模型,通过求解模型给出了最优合同存在的充分必要条件.进一步地,将上述模型应用于投资组合问题,其研究结果验证了模型的有效性,也表明了不同的决策准则将产生不同的最优合同,且信息的多样性增加了合同的复杂性.最后,将两种情形下的不确定期望值代理模型进行了比较,突出了两类代理模型之间的区别及其原因. 4)基于不确定测度,研究多重公共未知收益信息下的委托代理问题,重点探讨当代理人的努力水平对于委托人可观测和不可观测时的最优合同及其影响因素.当代理人的努力水平对于委托人可观测时,委托人只需要通过参与约束确保代理人与自己合作的意愿.然而,当代理人的努力水平对于委托人不可观测时,最优合同的设计取决于代理人是否愿意与委托人共同承担未知风险.鉴于此,在三种情形下分别建立了不确定信度代理模型,通过求解模型给出了对应情形下委托人与代理人之间的最优合同,分析了最优合同的唯一性、复杂性及其影响因素,并进一步将其应用于投资组合问题中验证了所得结论的合理性. 5)将本文的研究工作与已有结论进行了比较.首先,在可信性理论框架下完成了两方面的比较工作: i)将本文基于可信性测度所做的研究与相关文献进行了比较; ii)将基于可信性测度的代理模型与基于概率测度和容度的代理模型进行了比较.其次,在不确定理论框架下也完成了两方面的比较工作: i)将本文基于不确定测度所做的研究与相关文献进行了比较; ii)将基于不确定测度的代理模型与基于概率测度的代理模型进行了比较.比较结果不仅突出了本文研究内容的创新之处,还验证了基于可信性测度和不确定测度对代理问题进行研究的可行性,同时也为实际代理问题如何选择最佳的求解方法提供了理论依据.
[Abstract]:The principal-agent problem exists in all aspects of the economic management field. In these principal-agent problems, the incomplete information inevitably exists because of the influence of many uncertain factors, and the existence of incomplete information has an unnegligible influence on the optimal decision of the principal and agent. The adverse effect of low information incompleteness on the optimal return of a decision maker requires an optimal contract between the principal and the agent in advance. The contract not only regulates the expected behavior of the principal and agent, but also can effectively stimulate the agent in advance. Therefore, it will find the shadow of incentive under the incomplete information. The design of optimal contract has become one of the hot topics in international research.
Incomplete information usually contains uncertain factors with subjective characteristics, and the Ellsberg paradox has also shown that decision makers do not agree with the decision criteria based on an additive probability measure in some specific cases. Therefore, this paper will measure the reliability and uncertainty measures with sub additivity. On the basis of this, we focus on the study of the principal-agent problem of private information and public unknown information, and analyze the subjective attitude of the decision-makers on risk and the influence of information diversity on the optimal contract. The main contents of this paper include the following five aspects:
1) based on the credibility measure, the ability of the agent is the principal-agent problem of his personal information. Because the principal can not observe the agent's ability, the agent's estimation of agent's ability is described as a fuzzy variable. According to the agent's capability characteristics, the optimal transfer payment is designed. From the client's point of view, the agent model under the personal information is established based on the credibility measure, and the optimal contract when the principal is risk aversion and risk neutral is given respectively, and the influence of the subjective attitude of the principal on the optimal contract is analyzed. Because the agent has the information advantage, it can grab a certain information rent from the client in the process of cooperation, but the subjective attitude towards the risk has no effect on the optimal contract. Finally, the results are applied to the supply chain procurement system. The results show that the best contract can realize the effective incentive to the agent.
2) based on the credibility measure, the principal agent problem is studied when the agent's ability is public unknown information. Because the principal and agent are in the same information state, the purpose of the cooperation is to explore the potential ability of the agent through active and effective cooperation, thereby increasing the benefit of the two parties as much as possible. Thus, the principal is the principal. The optimal transfer payment is designed according to the benefit of the agent, and then the agent model under the public unknown information is established based on the credibility measure, and the optimal contract when the agent is the risk aversion and the risk neutral is given respectively. Because the agent is in the same information state, the agent can not seize the trust from the client again. At the same time, because of the unpredictability of public unknown information, the subjective attitude of the decision-maker has no influence on the optimal contract. Finally, based on the credibility measure, the optimal contract under the private information is compared with the optimal contract under the public unknown information, and the difference of the optimal contract under the two circumstances is highlighted.
3) based on the uncertainty measure, this paper studies the principal-agent problem under the information of multiple public unknown income, and focuses on whether the agent adopts the same decision criteria as the principal to the optimal contract. Because of the unobservability of the public unknown information, the agent's estimation of the unknown information is described as an independent uncertain variable. In order to investigate whether the agent adopts the same decision criteria as the principal to the optimal contract, this paper establishes an uncertain expectation value agent model in two cases, and gives the sufficient and necessary conditions for the existence of the optimal contract by solving the model. The validity of the model is verified, and it is also shown that different decision criteria will produce different optimal contracts, and the diversity of information increases the complexity of the contract. Finally, we compare the uncertain expectation value agent model under the two cases, and highlight the difference between the two kinds of agent models and their reasons.
4) based on the uncertainty measure, this paper studies the principal-agent problem under the information of multiple public unknown income, and focuses on the optimal contract and its influencing factors when the agent's effort level is observable and unobservable. When the agent's effort level is observable to the client, the client only needs to be ensured through the participation constraint. The agent's willingness to cooperate with himself. However, when the agent's level of effort is unobservable, the design of the optimal contract depends on whether the agent is willing to share the unknown risk with the principal. In this case, the uncertain reliability model is established in three cases, and the corresponding situation is given through the solution model. The optimal contract between the principal and the agent is analyzed, and the uniqueness, complexity and influencing factors of the optimal contract are analyzed, and the reasonableness of the conclusion is verified by applying it to the portfolio problem.
5) compare the research work of this paper with the existing conclusions. First, we have completed two aspects of comparative work under the framework of credibility theory: I) compare the research based on the credibility measure to the related literature; ii) make the proxy model based on the credibility measure and the agent model based on the probability measure and the tolerance. Secondly, in the framework of uncertainty theory, we have completed two aspects of comparative work: I) compared the research based on the uncertainty measure to the related literature; ii) compared the agent model based on the uncertainty measure to the agent model based on the probability measure. The comparison results not only highlight the study in this paper. At the same time, the feasibility of the research on the agent problem based on the credibility measure and the uncertainty measure is verified, and the theoretical basis is provided for how to choose the best solution for the actual agent problem.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：F274;F224

【参考文献】