Navigation and service

Oliver Lipps and Frank Betz (2005)

Stochastic Population Projection for West and East Germany*

In: Zeitschrift für Bevölkerungswissenschaft, Vol. 30, 1/2005, p. 3-44, Wiesbaden: VS Verlag für Sozialwissenschaften, ISSN: 0340-2398

Demographic projections are traditionally based on deterministic models. In these models, uncertainty is accounted for by calculating different scenarios. Given varying assumptions on the future development of demographic rates, future population parameters are obtained by simply updating the population of the base year. Since this technique makes unrealistic assumptions on the correlation of demographic rates and parameters, it is not able to consistently represent the uncertainty of the projection. Furthermore, this approach only yields results concerning the magnitude of the respective population parameter, but not on its probability distribution. In this paper, we develop a stochastic population projection based on time series analysis. This approach leads to probability distributions of future demographic rates. Using simulation studies, it is possible to consistently model the uncertainty of the projection, and thus to construct projection intervals for population parameters. In this framework, one only has to decide on the parameterization of the time series model and on the length of time series used for forecasting purposes. It is implicitly assumed that developments observed in the past will also be valid in the future. Consequently, we have chosen the historical time series such that no structural breaks, for instance revolutions, epidemics, baby boom, etc., have taken place. We apply this method to separately project the population of western and eastern Germany until 2050. To parameterize age-specific fertility rates, a model based on a bell-shaped Gaussian distribution is used. Mortality is projected by the well-known approach developed by Lee and Carter.

* peer-reviewed article (original German title: Stochastische Bevölkerungsprojektion für West- und Ostdeutschland)

This Page

© Federal Institute for Population Research - 2017