The accuracy of Statistics Norway’s national population projections

Abstract

Statistics Norway projects the population by age, sex and immigrant background at the national level. This paper examines the accuracy of the Norwegian population projections produced between 1996 and 2018. We assess deviations between projected and registered numbers, both for the total population and for several key components, such as age structure, total fertility rate and number of births, period life expectancy at birth and number of deaths, as well as net international migration. While fertility proved to be the most difficult component to project during the post-war period, net international migration has been the main source of inaccuracy in the national population projections produced since 1996. The projections produced in 1996, 1999, 2002 and 2005 underestimated longterm population growth due to the unforseen increase in immigration following EU expansion in 2004. Subsequent projections have not, however, produced a consistent under- or overprojection of net migration, and the deviations between the projected and registered total populations have been moderate. Overall, the projections for life expectancy at birth have proven to be consistently lower than the real development in life expectancy in Norway, at least until very recently. Fertility, on the other hand, has continued to be overprojected since observed fertility started to decline in 2009. Nevertheless, the deviations between projected and realised trends in births and deaths have been small compared to the observed deviations for net international migration.