摘要
针对台风数据匮乏的特点,采用贝叶斯推断方法来保证广东省台风巨灾损失频数分布和损失强度分布参数估计的准确性。在已知损失频数分布和损失强度分布的情况下,利用蒙特卡洛模拟法扩充样本对总损失分布模型进行拟合。利用零息债券定价公式对不同本金偿还比例的台风巨灾债券进行定价。实证结果表明:运用贝叶斯推断可以较好地解决台风数据不足的问题;台风巨灾债券的收益与风险成正比。
Monte Carlo simulation method is used to fit the total loss distribution model after expanding the samples in the case of knowing the loss frequency distribution and loss intensity distribution,whose accurate data can also be deduced by Bayesian inference method with deficient typhoon data.Zero coupon bond pricing formula is used to price the typhoon catastrophe bonds of different principal repayment ratios.By taking Guangdong Province as an example,it is concluded that Bayesian inference is preferable in solving the problem of lacking typhoon data and that the risk of Typhoon catastrophe bond is proportional to the revenue.
作者
展凯
丁冬
ZHAN Kai,DING Dong(School of Finance,Guangdong University of Foreign Studies,Guangzhou,Guangdong 510006,Chin)
出处
《长沙理工大学学报(社会科学版)》
2018年第3期88-95,102,共9页
Journal of Changsha University of Science and Technology:Social Science
关键词
台风巨灾债券
贝叶斯推断
共轭先验分布
typhoon catastrophe bond
Bayesian inference
conjugate prior distribution