未决赔款准备金评估的随机性Munich链梯法

Stochastic Munich Chain Ladder Method in Outstanding Claims Reserving

精算实务界通常采用链梯法等确定性方法评估未决赔款准备金,这些评估方法存在一定缺陷,一方面不能有效考虑保险公司历史数据中所包含的已决赔款和已报案赔款数据信息,另一方面只能得到未决赔款准备金的均值估计,不能度量不确定性。为了克服这些缺陷,本文结合Mack模型假设和非参数Bootstrap重抽样方法,提出了未决赔款准备金评估的随机性Munich链梯法,并应用R软件对精算实务中的实例给出了数值分析。

The chain ladder method and other deterministic methods are commonly used in outstanding claims reserving in actuarial practice. These assessment methods have some shortcomings. On the one hand, they do not effectively consider the information between the paid payments and incurred payments included in the historical data of insurance company. On the other hand, from these methods one can only obtain the mean estimation of claims reserves, but one cannot obtain the uncertainty measure of the reserves. In order to overcome these shortcomings, in combination with the assumptions of Mack model and non-parametric Bootstrap re-sampling method, the paper proposes a stochastic Munich chain ladder method in outstanding claims reserving, and further provides a numerical analysis with software R as an illustrations in actuarial practice.