残差相关条件下非寿险准备金风险分析

统计研究

残差相关条件下非寿险准备金风险分析

刘乐平等

出版日期:2015-10-15 发布日期:2015-10-26

Non-life Reserving Risk Analysis with Residual Correlative Structure

Liu Leping et al.

Online:2015-10-15 Published:2015-10-26

摘要/Abstract

摘要： 本文研究Solvency II框架下非寿险准备金风险的度量与控制问题。针对精算实务中赔付数据残差三角形不满足独立性假设情形，首先基于多重假设检验理论与错误发现率（FDR）控制过程，对残差三角形是否具有相关性进行识别和检验；然后提出两阶段分区域Bootstrap方法，模拟获得未决赔款的预测分布，分析残差相关性对准备金风险波动的影响；最后，保险公司真实索赔数据的实证结果表明：消除残差三角形相关性的影响，可以增强准备金估计的稳健性，从而有效提高准备金风险度量的准确性。

关键词: 非寿险准备金风险, 残差相关性, FDR, 两阶段分区域Bootstrap方法

Abstract: The Paper focuses on the measurement and control of non-life reserving risk under the framework of Solvency II. For the situation of residual triangle of claim data not fitting independence hypothesis, the paper firstly identifies and tests if the residual triangle is correlative based on multiple hypothesis testing theory and FDR process, and then designs a two stage block Bootstrap method to simulate the prediction distribution of outstanding liabilities and analyze the influence of residual correlation on reserving risk fluctuation. Empirical results of the real claims payment data of insurance companies show that the steadiness of reserving estimation can be improved by eliminating the influence of residual correlation so as to increase the veracity of measurement of reserving risk effectively.

Key words: Non-life Reserving Risk, Residual Correlation, False Discovery Rate(FDR), Two Stage Block Bootstrap

刘乐平等. 残差相关条件下非寿险准备金风险分析[J]. 统计研究, 2015, 32(10): 82-91.

Liu Leping et al.. Non-life Reserving Risk Analysis with Residual Correlative Structure[J]. Statistical Research, 2015, 32(10): 82-91.