Stochastic population forecasts using functional data models


22 February 2007


University of Melbourne


Abstract: I will present a new approach to age-specific forecasting of population based on separate forecasts of mortality, fertility and net migration. Functional data models with time series coefficients will be used to model mortality and fertility rates, and net migration. These models are then used in a Monte Carlo simulation of future mortality, fertility and net migration, to give age-specific forecast distributions of the population by sex. The forecast distributions allow the calculation of prediction intervals for the population size of any age group, and for any derived variables such as old-age dependency and life expectancy. The method is demonstrated by making 20-year forecasts for Australia.

Download pdf