风险信息共享背景下的个体风险评估研究

项目名称： 风险信息共享背景下的个体风险评估研究

项目编号： No.71303045

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 管理科学

项目作者： 谢远涛

作者单位： 对外经济贸易大学

项目金额： 20万元

中文摘要： 本项目将在风险信息共享的背景下讨探讨个体风险的评估和管理。以违约率为研究的主线，从两个视角出发，其一是从未定权益（CCA）的分析框架出发，计算财务违约率；其二是从结构方程框架出发，基于扩展的广义线性混合模型和真实数据计算风险信息违约率；按照信度模型得到实际违约率的信度估计；基于个人财富贴现的最大化构建动态规划模型，计算最优保费水平；根据实际保费水平与最优保费水平的差异的极小化建立优化模型，校准模型参数，估计违约率，为个体风险评估和量化风险管理服务。信息共享与个体风险评估是北京保险行业协会重点关注的内容，本项目将得到协会的业务和技术支持，最终实证结果可以作为政策参考。

中文关键词： 广义线性模型；信度；广义线性混合模型；风险依赖；风险评估

英文摘要： This project mainly discusses individual risk assessment and focuses on probability of default under the background of risk information share. Probability of default for financial can be obtained from the point of view of CCA, while the probability of default for risk information can be estimated from generalized linear mixed models within the framework of structure equations models. Then real probability of default can be estimated by way of credibility theory from both probability of default. Best premium level can be solved by constructing dynamic programming problem, parameter adjustment can be achieved by way of minimizing the difference between real probability of default and theory probability of default, which serve the risk assessment and risk management. This project is of important to the work of Beijing Insurance Association and may be supported by the latter. The result of the empirical analysis may serve as the political reference.

英文关键词： Generalized linear models；credibility；Generalized linear mixed models；risk dependence；risk evaluation