Estimating and Forecasting Bilateral Migration Flows from Europe to South America, 1986-2060

Andrea Aparicio-Castro , University of Manchester
Dilek Yildiz , International Institute for Applied Systems Analysis
Michaela Potancokova, International Institute for Applied Systems Analysis (IIASA)
Arkadiusz Wisniowski, University of Manchester

Europe hosted more than 30% of the global migrant stocks between 1990 and 2020. Given the importance of Europe as a destination, research usually focuses on migration towards the region, neglecting the need of analysing emigration. Emigration is a widely debatable topic amongst policymakers since it may represent the loss/waste of human capital. South America(SA) has been one of the usual receiving regions of migrants from Europe, even producing the change of the direction of Europe-SA flows in some periods, e.g., of economic crisis. The absence of a complete-time series of these flows does not allow drawing conclusions about them. Therefore, we aim at (a) estimating a complete and consistent time series of bilateral migration flows from Europe to SA from 1986 to 2019, and (b) quantifying how likely these patterns will remain across time, forecasting these flows until 2060. We use census data as an input to estimate flows. However, using censuses entails (a)working with five-year transition data that are not always comparable as they do not always refer to the same five years, (b)measurement errors caused by measuring migration indirectly, and (c) missingness in the intercensal periods. We propose a two-part hierarchical Bayesian model, which (a) translates five-year transition census data into one-year values; (b) corrects the measurement errors that using census data involves; (c) imputes flows for the intercensal periods; and, (d) forecasts these flows until 2060. The output is a set of synthetic estimates of bilateral Europe-SA flows from 1986 to 2060 with measures of uncertainty.

See extended abstract

 Presented in Session P1. Postercafe