We quantify the barriers to the economic integration of immigrants using an occupational choice model with natives and immigrants of multiple types subject to wedges that distort their allocations. We show that key parameters, including wedges, can be estimated to match the distribution of employment and earnings across individuals and occupations. We find sizable output gains from removing immigrant wedges in the U.S., accounting for 7 percent of immigrants' overall economic contribution. These gains arise from increased labor force participation and from reallocation from manual toward cognitive jobs. We show that the model-implied elasticities are consistent with empirical estimates and that immigrant wedges affect the impact of alternative immigration policies. Finally, we use harmonized microdata across 19 economies and find substantial cross-country differences in the estimated immigrant wedges. Differences in immigrant labor force participation and the correlation between wedges and productivities account for the heterogeneous gains from removing the wedges.