This paper studies a natural experiment to identify the causal effect of exposure to refugees in the neighborhood on the support for far-right, nationalist, anti-immigration parties. In the state elections in an Austrian state in September 2015 the far-right Freedom Party of Austria (FPOE) doubled its vote share with a fierce anti-asylum campaign. Since only 42 percent of communities hosted refugees at the time of the election, exposure to refugees varied at the local level. To account for the potential endogeneity in the distribution of refugees, I use pre-existing group accommodations as instrumental variable. To cope with the sudden inflow of large numbers of refugees, these buildings were used to accommodate refugees and thus strongly increase the probability of refugee presence in a community. In line with the contact hypothesis I find that hosting refugees dampens the overall trend and decreases FPOE support by 4.42 percentage points. Further analysis using exit poll data reveals a positive effect on the optimism in the population that the integration of refugees can be managed. Placebo tests show that there were no effects in elections prior to 2015.