Bayesian Probabilistic Projection of International Migration.

Area of Science:

• Demography
• Population Studies
• Computational Social Science

Background:

• Accurate population projections require robust modeling of international migration.
• Existing models often lack joint projections across all countries or detailed age-sex breakdowns.
• Global net migration is a critical constraint for demographic consistency.

Purpose of the Study:

• To introduce a novel method for joint probabilistic migration projections for all countries.
• To incorporate age and sex structure into migration projections.
• To ensure global demographic consistency by enforcing zero net migration.

Main Methods:

• Developed a probabilistic model for joint migration projections across all countries.
• Constrained model outputs to satisfy zero global net migration.
• Validated the model using out-of-sample testing and compared it to persistence and gravity models.

Main Results:

• The proposed model provides joint probabilistic projections of migration, disaggregated by age and sex.
• Out-of-sample validation demonstrated the model's predictive capability.
• Point projections showed improvements over a persistence model and a state-of-the-art gravity model.

Conclusions:

• The new method offers a robust framework for projecting international migration with demographic consistency.
• This approach enhances the accuracy of population projections by accounting for age-sex-specific migration flows.
• The model provides a valuable tool for demographic research and policy planning.