Mortality Prediction Of Hong Kong Based On Lee-Carter Model

According to the latest data from the C&SD,HK's population of 65 years and over reached 1.354 million in 2019,accounting for 18.0%of the overall population,the proportion far exceeding the internationally defined aging standard of 10%.The current aging population is increasing,which may seriously affect the development and living standards of HK and bring tremendous pressure on the labor market,the government's public finances and medical security.The increase in life expectancy per capita is the main cause of population aging.In 2014,HK surpassed Japan to become the region with the highest average life expectancy in the world.The average life expectancy reached 83.98 years,surpassing Japan 's 83.58 years.The reduction in mortality is the main reason for the increase in average life expectancy.The change in mortality has an important impact on the social security system:for the local government,the reduction in the death rate of the population will increase the government 's spending on social security for the pension and medical care,thereby increasing the government 's financial pressure;for insurance companies,the reduction in mortality affects the pricing of life insurance products and annuity products;for residents,the extension of life expectancy makes personal savings unable to better afford pension needs.Therefore,predicting the mortality rate of the population has wide application and important research value.Lee and Carter(1992)initially proposed a method for predicting future mortality for the US population mortality data.This method is a nonlinear model that covers age factors and time control factors,and establishes a logarithmic relationship with the central mortality rate.The model is simple in form and wide in application.Therefore,,this paper chooses the Lee-Carter model after analyzing a variety of mortality models.This article chooses the Lee-Carter mortality model to fit and predict HK population mortality data to explore the applicability and manifestation of the Lee-Carter model in Hong Kong population mortality.Based on the mortality data of Hong Kong population by sex and age from 1971 to 2018,SVD,WLS and MLE were used to estimate the parameters of the model.Comparing the prediction results of these three methods,we found that the MLE is the optimal parameter estimation method.Finally,we use the optimal fitting model to predict the future mortality rate of HK from 2019 to 2023,and then calculate the average life expectancy of newborn babies.We did a comparative study:one is to directly use the life expectancy of infants for time series modeling,the other is the average life expectancy calculated based on the mortality data predicted by the Lee-Carter model.The comparison found that the Lee-Carter model predicted more conservative.