Abstract:Since the 20th century, the most striking feature of the world population development is the prolongation of life expectancy caused by the continuous decline of population mortality. The longevity risk associated with the mortality decline already became the serious social risks faced by government and individuals. Therefore, it is very important to the quantification and management of life expectancy by predicting accurately the mortality improvement. The Lee-Carter model has received significant attention for its ability to model and project mortality rates since 1992. Because its computation is straightforward and it achieves strong accuracy in its predictions, the Lee-Carter model probably is the most popular approach for population projections. However, there is limitation restricting Lee-Carter model. Lee-Carter model exists significant heterogeneity. This paper proposes a mortality model based on change-point detection to project future mortality using principal component analysis. Comparisons of the proposed principal component analysis and change-point detection model with the well-known model --Lee-Carter model--employ empirical studies of mortality data from eight countries in Asia, Europe and North America. The mortality data come from the human mortality database. The proposed principal component analysis and change-point detection model produces smaller prediction errors for almost all illustrated countries in its mean absolute percentage error. Meanwhile we find that Lee-Carter model is a particular case of the proposed principal component approach, and by using two main components, heteroskedasticity could be reduced.

Dynamic principal component regression for forecasting functional time series in a group structure

DYNAMIC PRINCIPAL COMPONENT REGRESSION: APPLICATION TO AGE-SPECIFIC MORTALITY FORECASTING

Grouped multivariate and functional time series forecasting: An application to annuity pricing

FORECASTING MULTIPLE FUNCTIONAL TIME SERIES IN A GROUP STRUCTURE: AN APPLICATION TO MORTALITY

Grouped Functional Time Series Forecasting: An Application to Age-Specific Mortality Rates

High-dimensional functional time series forecasting: An application to age-specific mortality rates

Forecasting High-Dimensional Functional Time Series: Application to Sub-National Age-Specific Mortality

Multipopulation mortality modelling and forecasting: The multivariate functional principal component with time weightings approaches

Mortality and life expectancy forecasting for a group of populations in developed countries: A multilevel functional data method

A Forecast-driven Hierarchical Factor Model with Application to Mortality Data

Functional principal component analysis for longitudinal observations with sampling at random

A robust functional time series forecasting method

Mortality and life expectancy forecasting for a group of populations in developed countries: A robust multilevel functional data method

Mortality Forecasting using Factor Models: Time-varying or Time-invariant Factor Loadings?

Reconciling Forecasts of Infant Mortality Rates at National and Sub-National Levels: Grouped Time-Series Methods

Mortality Prediction Based On Chang-Point Detection Of The Principal Component Scores

Is the group structure important in grouped functional time series?

Dynamic modelling and coherent forecasting of mortality rates: a time-varying coefficient spatial-temporal autoregressive approach

Dynamic Function-on-Scalars Regression

Forecasting Age Distribution of Deaths: Cumulative Distribution Function Transformation

Forecasting age distribution of death counts: an application to annuity pricing