Assuming that immigrants select destinations according to absolute returns to their observable and unobservable human capital, I present a human capital model of migration accounting for taxes, transfers and limited portability of skills. The model predicts both segmented sorting of migrants to countries with a compressed income distribution, with negative sorting increasing with lower portability and positive sorting increasing with portability. Sorting to countries with greater income dispersion increases unambiguously with host-country relevant skills. Migrants to countries with compressed incomes will hence be more likely to be either out of work or overqualified and low-paid compared to natives with similar observable skills, and compared to migrants to countries with greater income dispersion. Regressions results on data for 16 OECD countries from the OECD Survey of Adult Skills are in line with the model. Controlling for observable skills and characteristics, including a literacy test score, immigrants from countries that are less wealthy or further away in geographical and cultural distance are significantly more likely to be either out of work or overqualified and low-paid in high-benefit countries. Wage compression, generous transfers and high taxes, typical traits of the so-called “Nordic” or “Flexicurity” model, may therefore contribute to making immigrant integration more challenging.