[Demographic population forecasts: theoretical framework, assumptions, and prediction uncertainty].

Population forecasts are not only highly demanded by scientists, politicians, and economists, but also by the general public. Thereby the wish of the users for only one forecast runs contrary to the uncertainty of the future developments of the population. In the past, but also today, population forecasts were and are predominantly accomplished by applying a deterministic approach: the Cohort Component Method. To counteract the uncertainty of future trends in the demographic processes in fertility, mortality, and migration, different scenarios are applied. Many studies have analyzed ex post the accuracy of past population projections. They show that, in addition to other factors, the time horizon and the level of regional aggregation influence the accuracy of the forecast outcomes. In particular, errors in the assumptions about future trends in fertility, mortality, and migration determine the accuracy of the forecasts. In many cases, these assumptions under- or overestimated the real trends. Progress in the question on uncertainty was made in recent years using a new approach: probabilistic forecasts which include probabilities of future trends in demographic processes.