Bayesian projection of international migration for all countries Seminar

In this talk I present a method for probabilistic projection of international migration. The goal of this work is to produce projections of net migration for all countries extending out to the year 2100, similar to the projections in the United Nation's World Population Prospects. Migration is notoriously hard to predict, so quantification of uncertainty is especially important. We accomplish this with a Bayesian hierarchical autoregressive model fit to time series data for all countries.