Integrating Traditional and Social Media Data to Predict Bilateral Migrant Stocks in the European Union

Abstract

Although up-to-date information on the nature and extent of migration within the European Union (EU) is important for policymaking, timely and reliable statistics on the number of EU citizens residing in or moving across other member states are difficult to obtain. In this paper, we develop a statistical model that integrates data on EU migrant stocks using traditional sources such as census, population registers and Labour Force Survey, with novel data sources, primarily from the Facebook Advertising Platform. Findings suggest that combining different data sources provides near real-time estimates that can serve as early warnings about shifts in EU mobility patterns. Estimated migrant stocks match relatively well to the observed data, despite some overestimation of smaller migrant populations and underestimation for larger migrant populations in Germany and the United Kingdom. In addition, the model estimates missing stocks for migrant corridors and years where no data are available, offering timely now-casted estimates.