The Effect of Social Networks on Migrants’ Labor Market Integration: A Quasi-Experiment (joint work with Klarita Gërxhani)

Empirically identifying the causal effect of social ties on migrants’ economic prospects is a challenging task. The ethnic social homophily argument predicts the non-random sorting of migrants into locations with greater opportunities for (previous) connections. We address endogeneity and self-selection issues by relying on a quasi-experiment distilled from a rich dataset of the IAB-SOEP Migration Sample. The experiment builds on the random residential allocation of various migrant groups by German authorities, minimizing the role of pre-existing social ties in migrants’ inflow. Our analyses are based on a random sample and a non-random sample, which were determined according to whether refugees and other migrants reported being subject to allocation policies. The empirical results obtained from these samples imply that the failure to control for potential (unobserved) confounders in the observational studies “pulls” the observed association between job-related social ties and migrants’ labor market outcomes away from the true association. Consequently, the effect of job-related social ties on migrants’ labor market entry is overestimated, and the effect on wages in their first jobs is concealed. Caution is warranted when interpreting previous results on the role of social ties in migration outcomes.

Keywords: social ties, refugees, first-generation migrants, labor market, quasi-experiment

Discussion Paper