Basic Data for Forecasting Small Areas

Helen Marklund, Statistics Sweden
Jeanette Bandel, Stockholm Office of Research and Statistics

Population forecasts of high quality are the best tool for the planning of a municipality, including child care and health care services for the elderly. For these applications it is necessary to make forecasts for smaller areas within municipalities. The statistical material for a single municipality may involve difficulties because it is based on too few observations to be reliable. Therefore basic data are needed on an overall base for the whole country, which can easily be updated. Experience has shown that it is best to analyse how risks for out-migration, age-distribution of in-migration and housing density varies for different residential types. Cluster analysis gives a division of different residential types with demographically similar classes. The division is defined by class of housing, tenure, year of construction, size of the town where the building is situated and also the taxation value.