Parsimonious stochastic forecasting of international and internal migration on the NUTS-3 level: An outlook of regional depopulation trends in Germany

Abstract

Substantiated knowledge of future demographic changes that are derived from sound statistical and mathematical methods is a crucial determinant of regional planning. Among the components of demographic developments, migration is the one that shapes regional demographics the strongest in the short term. Yet, despite its importance, existing approaches model future regional migration based on deterministic assumptions that do not sufficiently account for its highly probabilistic nature. Given this shortcoming in the literature, our paper uses age- and gender-specific migration data for German NUTS-3 regions over the period 1995-2019 and compares the performance of a variety of forecasting models in backtests. Using the best-performing model specification and drawing on Monte Carlo simulations, we present a stochastic forecast of regional migration dynamics across German regions until 2040 and analyze their role in regional depopulation. The results provide evidence that well-known age-specific migration patterns across the urban-rural continuum of regions, such as education-induced migration of young adults, are highly probable to persist and shape future regional (de)population dynamics.