Forecast Of Elderly Mortality Based On Classical Model

Since 1950,the survival rate of the world’s elderly population has been continuously increasing.As a developing country,China has enjoyed rapid economic growth over the past decades,with its GDP constantly rising and its medical and health care levels constantly improving.More and more young people choose to have fewer children,and the cultivation principle of high standards makes the dependency ratio of the total population increase year by year.This has a huge impact on China’s public pension system and medical security system.Compared with foreign demographic data,China’s statistical data on the mortality rate of the elderly population are relatively scarce,which not only limits the accuracy of model parameter estimation,but also influences the accuracy of model prediction.Based on the actual situation of insufficient demographic data in China,this thesis conducts statistical modeling and prediction analysis of mortality.Firstly,the extreme value theory method was used to fit and predict the mortality rate of the elderly population in China and Japan.The optimal threshold age and model parameters were selected by using the least square and repeated tests,and overcome any subjective judgement,ensure older mortality and extrapolation of smoothness,provides favorable data support for the future.Secondly,this thesis adpots the latest demographic data from 1995 t02015,and uses Lee-Carter(LC)model and Cairs-Blake-Dowd(CBD)model to fit the mortality data in China.In the process of fitting,the classical model was compared according to the bayesian information criterion and the red-pool information criterion,the LC model was found to be superior to the CBD model.As for the residual diagram,the choice of the model depends on the situation.Finally,the traditional ARIMA method and the double-random walk process were used to predict the mortality data of Chinese population,and the predicted results were compared with the real values,so that the traditional ARIMA was better.According to the result,the mortality of Chinese population in the next five years was predicted,which provided favorable data support for China’s old-age undertakings and industrial development planning.