Bayesian Hierarchical Model for Estimating Age and Sex Patterns of International Migration.

AYSHA BASHEER, University of Manchester
Arkadiusz Wisniowski, University of Manchester
Maciej J. Danko, Max Planck Institute for Demographic Research (MPIDR)

This study presents a Bayesian hierarchical model to analyze the age and sex patterns of international migration across Europe. The model incorporates age, sex, origin, destination, and time to address the challenges of inconsistent and incomplete migration data. Data from 31 European countries between 2008 and 2022 were used, focusing on 18 five-year age groups (0 to 85+) for both sexes. We implemented three variations of the log-linear model, with the most complex version, including interaction and time terms, showing the best fit according to diagnostic measures. Preliminary results highlight higher migration proportions among males, particularly in the 20-34 age group, with an increase in migration observed between 2009 and 2015. These estimates were compared to reported values and will be further evaluated against the Rogers-Castro migration schedule. Furthermore, we plan to extend the model to add covariates and incorporate splines for smoothing flows. This study provides valuable insights into the age and sex patterns of European migration flows, offering a robust approach to harmonizing inconsistent migration data and potentially improving the accuracy of migration statistics, which are crucial for policy-making and social service planning.

Keywords: Bayesian methods , International Migration, Data and Methods

See extended abstract.