AbstractWe use data on Turkish immigrants in two European welfare states, Denmark and Germany, and data on Turks at home. Unlike in most studies of immigrant poverty, we thus control for the differences in immigrant composition. Denmark and Germany have different welfare state types, labour market structure and institutions. We find that in both countries Turkish immigrants have much higher poverty rates than natives. We perform Fairlie decompositions to find that in Denmark, compared to Germany, a larger part of the native‐immigrant poverty difference is explained by market valuation of characteristics and by unobservables. Finally, we decompose poverty by subgroups and find that certain immigrant subgroups (such as families with children and the elderly) are especially vulnerable in both countries and that not much has changed in the two countries between 2008 and 2013 in terms of the vulnerability of these sub‐groups to poverty risk.