Bayesian Estimation and Projections of Internal Migration Flows in Italy

Emanuela Furfaro, University of California, Davis
Federico Castelletti, Università Cattolica del Sacro Cuore
Adrian Raftery, University of Washington, Seattle

We use a Bayesian Hierarchical model to estimate and forecast internal migration flows. In particular, we use data on sub-national migration in Italy and we develop a Bayesian Extended Gravity Model, which includes demographic and geographic covariates, for testing theories of migration. The model is well calibrated, and we find that the Total Fertility Rate is a relevant predictor. We also propose several Bayesian models for forecasting migration flows between all pairs of provinces, and we find that Bayesian autoregressive models perform well and give well-calibrated intervals.

Keywords: Bayesian methods / estimation, Internal migration, Population projections, forecasts, and estimations

See paper.

  Presented in Session 170. Understanding Migration: Applying New Data and Methods