Estimation,Prediction And Application Of Chinese Urban Population Mortality CBD Model Under Irregular Data

Since the 21 st century,from China's age-specific population data released by demographic information,we can see that the proportion of the population over 65 years old is getting bigger and bigger,which shows that the problem of population aging in China is increasing.The reason for the aging of the population is the decrease in the birth rate and the extension of the life expectancy.The fertility rate will change and stabilize with a series of social,economic and cultural levels,such as social,economic level and birth policy.Under certain fertility levels,the continuous decline in mortality and the prolonged life expectancy have become long-term influencing factors for the aging population.The extension of life expectancy will lead to a gap in the personal accounts of pensions,which will bring huge payment pressure to China's social pension insurance system.Mortality prediction has always been the focus of demographics.The Lee-Carter model and the CBD model are two classic models of random mortality prediction.Considering that the CBD model has a better predictive effect on high-age mortality data,this paper mainly establishes a CBD model for the age-disaggregated data of Chinese urban population.The upper age limit of the statistics of high-age population in China is irregular,which brings trouble to the establishment of the CBD model.In this paper,an iterative weighted least squares method is used to estimate the model parameters.According to the AIC criterion and other statistical discriminant methods,the most suitable age range for modeling is determined.This is an innovation point of this paper.In addition,the life expectancy calculated by the CBD model based on the original data is significantly different from the life expectancy announced by the National Bureau of Statistics in the same period.Therefore,the time factor of the model is adjusted according to the life expectancy announced by the National Bureau of Statistics.Life expectancy in national statistics is late revision data and should be more credible than the original data,so the mortality model adjusted for life expectancy should be more reasonable.Adjusting the parameters of the CBD model based on life expectancy is another innovation in this paper.It should be noted that the life expectancy published by the National Bureau of Statistics is the life expectancy at birth,while the CBD model is a model of high-age population.The life expectancy of the model can only be old,such as 60-year life expectancy.Therefore,the Lee-Carter model is used in the adjustment of the time factor based on life expectancy.We can use the life expectancy of high age data calculated by Lee-Carter model adjusted by the time factor to adjust CBD model.According to the established time factor model,single-valued and randomized predictions of mortality and life expectancy can be obtained.In order to verify the effect of the model parameter adjustment,in this paper,the life expectancy values of different age range modeling,no parameter adjustment and parameter adjustment are calculated and compared.Finally,based on the adjusted life expectancy forecast,the impact of life expectancy extension on pension personal accounts and population aging is discussed,and corresponding suggestions are put forward.