This paper examines how local labor market conditions and native attitudes at the time of refugees’ arrival shape their subsequent integration in Germany. During the large influx of asylum-seekers to Germany between 2013 and 2016, asylum-seekers were centrally assigned to counties and could not choose their residence for at least three years. This policy aimed to ensure a fair distribution of refugees across counties, minimize tensions between natives and refugees, ease fiscal pressures, and reduce ethnic enclaves, but did not explicitly match asylum-seekers to local integration capacity.
The authors exploit the centralized allocation of asylum seekers across counties, which limits the self-selection of refugees into high-demand labor markets or existing ethnic enclaves. The analysis draws on data from the IAB-BAMF-SOEP Survey of Refugees, a representative sample of asylum seekers who arrived in Germany between 2013 to 2016. The survey provides detailed residence histories, socio-demographic characteristics and integration outcomes. Data on county-level characteristics come from the German Federal Statistical Office. Sentiments of German residents towards immigrants are measured using geo-coded Twitter data as well as vote shares for far-right parties. Refugee integration is measured across economic, linguistic, navigational, political, psychological, and social dimensions, following Harder et al. (2018).
Consistent with improving integration over time, the share of refugees not in employment or training declines with years of residence in Germany. However, after three years of residence in Germany, only about 30 percent of refugees are employed.
Main empirical results:
- Central allocation is unrelated to local socio-economic conditions. County population size is the only statistically significant predictor of the number of asylum seekers assigned to county. There is no statistically significant association between local socio-economic characteristics (unemployment rate, population density, GDP per capita, average age, and housing space per capita) and the number of asylum seekers at the county level. This supports the claim that placement did not target labor demand or housing conditions and helps identify causal impacts of initial local context on later outcomes.
- Attitudes and unemployment capture distinct local factors. Local attitudes towards immigrants are only weakly correlated with unemployment.
- High local unemployment rates and unfavorable attitudes towards immigrants reduce refugee employment rates. A one standard deviation increase in either the unemployment rate or the negative sentiment index is linked to roughly a five percentage point reduction in the likelihood of being employed.
- Adverse local conditions depress earnings and overall integration. Both high unemployment and negative sentiment are linked to lower net monthly wages and weaker scores on the Multi-dimensional Integration Index, with attitudes appearing somewhat more important than unemployment for multi-dimensional integration. This suggests that local factors affect not only job access but also the quality of economic participation and social integration.
- Far-right vote share corroborates the sentiment mechanism. A one standard deviation higher far-right vote share (about a 1.07 percentage point increase) lowers the probability of being in employment or education by 3.2 percentage points and of being in full- or part-time work by 2.2 percentage points, and reduces net monthly wages by 14.4 percent on average.
- Effects concentrate in social and economic dimensions. Unemployment and negative sentiment explain variation in social and economic components of the integration index but do not significantly affect psychological, linguistic, political, or navigational outcomes. This suggests that local conditions operate primarily through labor market access and social inclusion rather than language acquisition or civic engagement.
- Gender patterns differ: unemployment matters more for men; attitudes matter more for women. For men, a one standard deviation higher local unemployment reduces the likelihood of full- or part-time employment by about seven percentage points and lowers wages and overall integration. For women, the earnings penalty associated with more negative local sentiment is larger than that for higher unemployment.
Taken together, the findings show that unfavorable initial local conditions—both weak labor markets and hostile attitudes—can have negative consequences for the economic integration of refugees, even in the absence of explicit restrictions on employment. Unfavorable initial conditions also affect refugees’ future earnings and social integration. The authors conclude that attitudes towards immigrants are as important as local unemployment rates in shaping refugees’ integration outcomes. Moreover, the centralized allocation of refugees into economically weaker or less welcoming counties imposes substantial integration costs, particularly in employment and earnings. The results provide evidence for aligning placement policies with local conditions to improve refugees’ early trajectories and reduce longer-term social and economic costs.