This paper examines the role of unobserved ability in explaining inter-industry wage differentials, drawing on data on brothers. Such data allow us to account for unmeasured abilities due to common family and community factors shared by siblings. Important advantages of this approach are that we do not rely on assumptions of exogenous job mobility and that estimates reflect long-run wage differentials rather than short-run differences following switch of industry. The data sets come from four of the Nordic countries and the United States. We find that, in the Nordic countries, only a moderate proportion of the variability in industry wages is due to unobserved ability, while unmeasured factors explain as much as half of the industry wage variation observed in the United States. Accounting for such differences, we show that the U.S. inter-industry wage dispersion is similar to those of the Nordic countries.