|Speciale - modelbyggervariant, 3. modul, 2007, id:360|
|Findes på RUb:||Ja|
In this thesis we present a population forecast for women in Denmark 50 years from now. We use mathematical modeling to produce the forecast. The mathematical model consists of an accounting model and separate submodels for fertility, death frequency, emigration frequency and number of immigrants. In the setup of each submodel we use statistical methods, and we focus on locating and describing possible soures of uncertainty. The resulting forecast shows both the uncertainty of the size of the total population and the uncertainty of one-year age classes 50 years from now. The uncertainty grows quickly during the perod. We therefore conclude that the value of the forecast as a basis for making decisions is limited after 30 yeras and almost non-existent after 50 years. The most certain mechanism of the population forecast is death, and the greatest source of uncertainty is linked to fertility. The strongest weakness of the forecast is the description of migration, which is based on ad hoc-assumptions. However, the forecast of the oldest age-groups makes sense throughout the entire time horizon, whereas the forecast of the youngest age-gruoups loses its applicability after 30 years. The results of the forecast are compared with corresponding population projections from DREAM and Statistics Denmark. Their expected projections are in accordance with the median curves of our forecast. Since the forecast shows that the uncertainty grows rapidly throughout the period, we conclude that the uncertainty grows rapidly throughout the period, we conclude that population projections are of limited value on time durations more than 5 to 10 years, without considerations of uncertainty.