Unintended Impacts of Automation for Integration? Simulating Integration Outcomes of Algorithm-Based Asylum Seeker Allocation in Germany

Economic integration is one among various important integration outcomes for asylum seekers, affecting individual well-being and economic development. When assigning asylum seekers to resettlement locations within a country, taking into account asylum seekers’ skills and characteristics could improve the chances of integration into the local labor markets. Researchers thus developed algorithmic procedures that match asylum seekers to resettlement locations based on asylum seekers’ characteristics. However, attempts to optimize solely for labor market integration may lead to unintended consequences for other integration dimensions. To investigate such effects, we use agent-based models to simulate different allocation mechanisms of asylum seekers to locations in Germany and their impact on integration outcomes. Concretely, we compare the current approach using the Königsteiner Schlüssel with the algorithm-based procedure GeoMatch. We study the procedures’ impacts on labor market and social integration, while taking effects on inequalities between subgroups of asylum seekers into account. Decision models and agents’ characteristics are based on the IAB-BAMF-SOEP survey of asylum seekers and refugees in Germany. This project contributes to the literature by showing to which extent algorithm-based approaches for optimizing labor market integration may impact another key integration outcome. More generally, it shows how agent-based models can be used to study unintended consequences of algorithm-based allocations of asylum seekers in dynamically changing demographic environments.