We estimate the economic gains from the capability in the host country's language using administrative data on the universe of all African migrants that applied for asylum in Switzerland between 1998 and 2009. To isolate the causal effect of language capability on a variety of labor market outcomes, we exploit the random placement of refugees to cantons and the existence of a sharp language border dividing Swiss German and French-speaking cantons. We apply a difference-in-difference estimator using French-speaking African migrants assigned to the French-speaking parts of Switzerland as treatment, and English-speaking African migrants as control group. Our main results suggest an increase in the probability of finding a job of 160\% in the first year after arrival, and more than 100% for the following years.