Skip to content. | Skip to navigation

Personal tools

Christina Bohk, Roland Ewald, and Roland Rau (2012)

Towards a Generalized Subpopulation Support for Stochastic Population Projections

Poster, , Winter Simulation Conference 2012.

Demographic heterogeneity, i.e. differing mortality and fertility among subpopulations, is an important issue in stochastic demographic forecasting. Common approaches typically use the variables age and sex to construct subpopulations, but this might be insufficient and induce projection error. Many studies show significant differences in mortality and fertility among people with and without migration background, but also among people with different level of education or country of origin. So far, our model projects the autochthonous population, immigrants, emigrants, and their descendant generations with separate mortality and fertility. Hence, the subpopulations are build with the variables age, sex, and migration status. In this paper, we extend the model so that a forecaster can project an unlimited number of subpopulations. Next to age, sex, and migration background, a forecaster can use other characteristics like reason of migration or employment to construct subpopulations, and thus to increase projection accuracy.