Towards causal forecasting of international migration

International migration is difficult to predict because of uncertainties. The identification of sources of uncertainty and the measurement and modelling of uncertainties are necessary, but they are not sufficient. Uncertainties should be reduced by accounting for the heterogeneity of migrants, the reasons why some people leave their country while most stay, and the causal mechanisms that lead to those choices. International migration takes place within a context of globalisation, technological change, growing interest in migration governance, and the emergence of a migration industry. Young people are more likely than older people to respond to these contextual factors, as they are better informed, have greater self-efficacy, and are more likely to have a social network abroad than previous generations. My aim in this paper is to present ideas for the causal forecasting of migration. Wolfgang Lutz’s demographic theory of socioeconomic change is a good point of departure. The cohort-replacement mechanism, which is central to Lutz’s theory, is extended to account for cohort heterogeneity, life-cycle transitions, and learning. I close the paper by concluding that the time has come to explore the causal mechanisms underlying migration, and to make optimal use of that knowledge to improve migration forecasts.