《保险研究》20200208-《考虑状态停留时长的我国中老年人口状态转移概率测算》（刘乐平、唐爽、程瑞华）

［摘   要］健康状态转移概率测算是长期护理保险定价的基础,对我国建立发展长期护理保障制度具有重要意义。目前假设健康状态转移服从Markov过程的建模方法使用较为广泛,本文使用Semi-Markov多状态模型对其进行扩展，在中国健康与养老追踪调查数据基础上，借鉴国际相关研究数据， 进一步精算不同状态停留时长下的我国中老年人口状态转移概率。研究表明,在初始状态的停留时长对转移概率有显著影响,女性比男性更需要长期护理保障。

［关键词］长期护理；转移概率；Semi-Markov多状态模型

［基金项目］本文为国家自然科学基金项目《基于机器学习的长期护理保险精算预测模型与风险分析》(71771163)的阶段性研究成果。

［作者简介］刘乐平,天津财经大学科研处处长,大数据统计研究中心主任、统计学院教授,研究方向:精算与风险管理、机器学习；唐爽(通讯作者),天津财经大学统计学院博士研究生,研究方向:精算与风险管理、机器学习；程瑞华,天津财经大学大数据统计研究中心博士后,研究方向:精算与风险管理、机器学习。

The Estimation of State Transition Probabilities of Middle-aged and Elderly People in China with the Consideration of the Length of Sojourn

LIU Le-ping,TANG Shuang,CHENG Rui-hua

Abstract：Estimation of the transition probabilities is the basis of long-term care (LTC) insurance pricing,which is very important for the establishment of the LTC security system in China. Currently,the modeling method that assumes that transitions obey the Markov stochastic processes is widely used. Based on the data of China Health and Retirement Longitudinal Study,this paper combined relevant international research data and used the Semi-Markov multistate model to extend the basic assumptions,and further calculated the transition probabilities of middle-aged and elderly people in China under different lengths of sojourn. The research results show that the sojourn time in the initial state has a significant impact on the transition probabilities,and women need LTC more than men.

Key words：long-term care; transition probabilities; Semi-Markov multistate model