我国商业车险奖惩系统研究  

The Bonus-Malus System for Commercial Automobile Insurance in China

作　　者：张永霞[1,2] 孟生旺[1,2] 

机构地区：[1]中国人民大学统计学院,北京100872 [2]中国人民大学应用统计科学研究中心,北京100872

出　　处：《保险研究》2016年第10期3-15,共13页Insurance Studies

基　　金：教育部重点研究基地重大项目"基于大数据的精算统计模型与风险管理问题研究"

摘　　要：商业车险的费率主要由先验费率和浮动费率两部分构成。先验费率通常使用广义线性模型进行厘定,而浮动费率则是通过奖惩系统(Bonus-Malus System,BMS)对先验费率进行的一种后验调整。我国现行的商业车险BMS存在转移规则不够合理和奖惩系数不尽准确的问题。本文使用动态转移规则,在线性约束下,通过最小化索赔频率真实值与预测值之间的均方误差,求得了BMS的最优奖惩系数,并基于我国某财产保险公司2015年的一组商业车险数据进行了实证研究。结果表明,本文的方法可以有效改善我国现行商业车险BMS存在的不足。The premium rate of commercial auto insurance is mainly composed of two parts, the priori premium rate and the floating premium rate. The priori premium rate is usually determined by the generalized linear models, and the floating premium rate is a posteriori adjustment on the priori premium rate achieved by the Bonus-Malus System （BMS）. In China＇ s current BMS,the transfer rule is unreasonable and the Bonus-Malus coefficients are inaccu- rate. In this paper, by using the dynamic transfer rules, the optimal Bonus-Malus coefficients of BMS were solved under the linear constraint and by minimizing the mean squared error between the true value and the predictive val- ue of the claim frequency. The BMS was applied to a set of commercial auto insurance data of a property insurance company in China, and the result showed that the proposed method could effectively overcome the existing problems of the BMS in China.

关 键 词：商业车险 先验费率 浮动费率 BMS 

分 类 号：F222.3[经济管理—国民经济]