破产概率视角下的巨灾保险偿付能力分析——基于混合分布模型

Analysis of Catastrophe Insurance Solvency from the Ruin Probability Perspective:Based on Mixed Distribution Model

破产概率是分析保险公司巨灾风险偿付能力的一个重要数量指标,而巨灾损失数据的拟合模型则是准确估计破产概率的关键。选取我国1990-2018年震级5级以上地震灾害为研究样本,用非参数核密度估计方法和广义帕累托分布构建了地震造成的直接经济损失的混合分布模型,并用峰度法确定该模型中的阈值,进而给出保险公司破产概率的尾渐近估计。实证结果表明,所构建的混合分布模型对巨灾损失数据的拟合效果良好,可以为我国保险公司经营巨灾风险提供一种新的精算模型。

Ruin probability is an important quantitative indicator to analyze the catastrophe insurance solvency,while the catastrophe loss data fitting model is the key to accurately estimate ruin probability.This paper collects earthquake disasters with magnitude five or above during the years 1990-2018 in China as the research sample.A mixed distribution model of earthquake direct economic losses is constructed by combining non-parametric kernel density estimation methods with generalized Pareto distribution.Moreover,the threshold used in this model is determined by Kurtosis method.Then the last asymptotic estimation for ruin probability is derived.The empirical results show that the mixed distribution model constructed in this paper fitting the catastrophe loss data well and provide a new actuarial method for insurance companies to manage catastrophic risks in China.