The labour market differences between the asylum population and other migrants

Abstract This article examines economic disparities between the asylum population and other migrants accounting for selection on unobservables. With administrative datasets of arrivals since the 1940s, descriptive evidence suggests that the asylum population is 7%–11% less likely to be employed and earn 15%–19% less than other migrants from 2010 to 2014 in Switzerland. Comparing bias-adjusted results considering selection on unobservables to conventional approaches such as decomposition and matching, I find that the bias-adjusted gaps bias towards zero. This suggests that conventional approaches may overestimate the outcome gaps. My results highlight that the outcome gaps persist, although discrimination and other unexplained factors are decreasing over time.