多状态Markov模型在长护险精算中的运用——基于CHARLS数据  

Application of Multi-state Markov Model in Long-term Care Insurance’s Actuarial Science:Based on CHARLS data

作　　者：彭逸铭 揭佳豪 Peng Yiming;Jie Jiahao

机构地区：[1]中南财经政法大学金融学院,湖北武汉430073

出　　处：《保险职业学院学报》2021年第4期66-73,共8页Journal of Insurance Professional College

摘　　要：本文运用CHARLS在2013年和2015年的数据,选择60-80岁的男性和50-80岁的女性作为样本,依照国际上常用的ADLs体系统计出各年龄段内健康、失能、死亡状态的人数,使用多状态Markov模型获得间隔为2年的状态转移概率,进而得到间隔为1年的一步转移概率矩阵,从而计算出男性60-80岁和女性50-80岁逐年趸交纯保费表。研究结果显示:女性的失能概率大于男性失能概率,且随着年龄的增加,差值增大;长护险趸交保费对于普通家庭来说将是一笔巨大的额外支出,而商业保险模式保费高昂,很难具备普惠性质,因此,长护险的发展需要合适的筹资机制,鼓励国家资本与社会资本共同参与。The paper uses CHARLS data in 2013 and 2015,choosing 60-80 year-old men and 50-80 year-old women as samples and in accordance with international common ADLs system statistics in all ages,calculates the number of the state of health,disability,death,and multi-state Markov model to obtain the in⁃terval of 2 years of state transition probability,then multi-state Markov model is used to obtain the probability of state transition with an interval of 2 years to calculate the net single premium of 60-80 year-old males and 50-80 year-old females.The results show that the probability of female disability is greater than that of male disability,and the difference increases with the increase of age.For ordinary families,the premium for longterm care insurance will be a huge extra expenditure.However,the commercial insurance model is difficult to have inclusive nature due to its high premium.Therefore,the development of long-term care insurance needs an appropriate financing mechanism to encourage the participation of state capital and social capital.

关 键 词：长护险 精算定价 多状态MARKOV模型 转移概率矩阵 

分 类 号：F840.48[经济管理—保险]