巨灾风险厚尾分布:POT模型及其应用

Heavy-tailed Distribution of Catastrophe Risk : POT Model and Its Application

针对Gamma，LognormM和Weibull等传统厚尾分布拟合巨灾风险的不足，本文一方面从理论上分析探讨了POT模型及其拟合巨灾厚尾风险的相对优势，另一方面应用POT模型和GPD分布，对我国1952年～2008年间地震直接经济损失数据进行拟合，发现了POT模型拟合巨灾风险厚尾部分的效果比Gamma、LognormM和Weibull等分布的拟合效果更为理想，最后探索了POT模型在VaR和超赔再保险的定价等方面的具体应用。

The paper presented the shortcomings of some traditional heavy-tailed distribution theories such as Gamma, Lognormal and Weibull in fitting catastrophe risks. It discussed the POT model and its comparative advantage to fit catastrophe risk theoretically, and then applied the POT model and General Pareto Distribution to fit the earthquake direct economic losses in China from year 1952 to 2008. It came to the conclusion that the POT model was better than Gamma, Lognormal and Weibull in fitting catastrophe risks, especially its fat tail section. At last, it explored some applications of POT model on VaR and excess of loss reinsurance.