Prediction Of China's Population Age Structure

In the past twenty years,our population age structure is undergoing significant changes.The decline of fertility and the extension of life expectancy accelerate the progress of population aging.From 2005 to 2013,the proportion of the people over 60 increased from 10.45 percent to 10.45 percent.So,the rapid ageing tendency has brought stronger stress to our old-age security system.In addition,our aging population structure has also caused many other problems,such as the increase of the sex ratio at birth and old age dependency ratio,the high level of sex ratio of the gross,reduction of labor resources of our country etc.All of these issues have a great impact on the development of economic and social.Therefore,whether we could find some scientific and accurate methods to forecast the population age structure of our country is very important,those accurate research results can provide some beneficial guidance to the government strategy.It can also offer theoretical and empirical basis for China's next step population development planning.Domestic studies on population age structure start late,most of the research is limited to qualitative analysis for a long time,these studies can only predict the roughly distribution of the age structure,cannot be used to analyze the concrete numerical value of population age structure.With the improvement of precision requirements for population planning,methods of quantitative calculation of all kinds of index of population are receiving more and more attention.And the current domestic mainstream population age structure prediction models contains logistic model,neural network,the regression equation and Leslie matrix model and so on,but most of them take a certain point of view to analyze static statistics data,which is very limited to reflect problems.Based on the problems above,we applied a method named functional data analysis which derived from continuous random population forecast model to the forecasting of population age structure.Therefore,we chose cohort-component method and take it as a basic model,then apply functional data analysis to it.Based on this,we established the age-specific mortality rate and age-specific fertility rate respectively to predict the future deaths and births,so I can get the total population and the population age structure.But we cannot estimate the influence of immigration rates for the lacking of immigration data and its' small proportion.So we chose to ignore this factor.Based on this,we establish the functional data model of mortality and fertility respectively,It takes the observed data as a whole,and considered the process of mortality and fertility rate change with age as a continuous stochastic process.So,every yearly mortality or fertility rate corresponding to a curve,whose horizontal axis is age and vertical axis is mortality rate or fertility rate.And the sample data will be a series of curve.Then,we can predict the future population age structure according to the population of birth and death.The results showed that accuracy of functional data model is higher,the prediction accuracy of male mortality function model is above 90%,female mortality function model is above 86%,fertility rate function model is above 88%.This paper is arranged as follows:Chapter1:Introduction.Mainly contains the background and significance of the research,research status of population age structure and functional data analysis in China and abroad,research ideas and framework,and innovation place.Chapter2:Population age structure model and functional data model.The first part is cohort-component forecast model,it contains the introduction of the basic theory of the model and the interpretation of some relevant indicators.The second part is about the functional data model,it contains some fundamental theory and the specific methods of applying it to the study of population age structure.Chapter3:The population age structure prediction model based on functional data.The first part is about the sources and pretreatment of data.The second part is the establishment prediction and of the model,including functional mortality model,functional fertility rate prediction model and population age structure prediction model.Chapter4:Conclusion and policy recommendations.The first part is the main conclusion of the article,including the future population age structure,labor population,the total dependency ratio and population aging process.The second part is the policy and suggestion part,including some suggestions about issues brought by the changes of population age structure and the advantages and disadvantages of functional data analysis methods and the attention should be paid while using this functional analysis method.The research of this paper expanded application fields of functional data analysis method in our country,enriched the demographic research methods and provides a broader thinking for the research of population age structure.Compared with the traditional model,functional data analysis model make full use of historical data information without any subjective judgment.And the continuous random mortality and fertility function improve the prediction precision greatly.