Lee-Carter模型在中国城市人口死亡率预测中的应用与改进  

Application and improvement of Lee-Carter Model in mortality projection of China′s urban population

作　　者：韩猛[1] 王晓军[2] 

机构地区：[1]内蒙古财经学院统计与数学学院,内蒙古呼和浩特010051 [2]中国人民大学统计学院,北京100872

出　　处：《保险研究》2010年第10期3-9,共7页Insurance Studies

摘　　要：由死亡率下降带来的长寿风险给社会、政治以及经济带来了新的挑战。为了更加准确地对长寿风险进行评估和管理,需要对未来死亡率趋势进行预测。本文针对我国死亡率数据样本量小以及数据存在缺失的实际情况,对Lee-Carter模型进行了改进,通过一个双随机过程对Lee-Carter模型中的时间项进行建模。在模型中考虑了样本量不足对预测结果造成的影响,使得改进后的Lee-Carter模型更加适合目前中国的人口死亡率预测。The longevity risk, due to mortality improvement, is imposing significant social, political and economic challenges for us. To better assess and manage the longevity risk, we need to make a projection of future mortality trends. Due to the actual situation of limited and incomplete mortality sample data in China, the author improved on the Lee-Carter model, and applied a double stochastic process to model its time index. The new model made an allowance for the impacts of sample data insufficiency on the projection result, so that the improved Lee-Carter model better suited the Chinese population mortality prediction.

关 键 词：死亡率预测 长寿风险 Lee-Carter模型 

分 类 号：F840[经济管理—保险]