ABSTRACT
From the vantage point of economic policy, assessing migration potential and predicting future migration streams are among the most relevant, yet least well understood topics of migration research. Most theoretical models and a large range of econometric studies successfully address heterogeneity at the individual level, with an emphasis on the detection of demographic and socioeconomic determinants of the individual migration decision, or on the identification of the appropriate decision unit. In the aggregate, though, many important explanatory factors are shared within the regions of origin and destination, rendering the individual-level results inappropriate as a predictive tool, and necessitating an analysis over time and space. The usual approach taken to address aggregatelevel prediction problems is to fit ad hoc specifications to historical data, and to extrapolate from these estimates on the basis of conditioning information that is assumed to be known with certainty.1
