This paper uses labor market evidence to quantify the importance of quality-adjusted schooling differences in accounting for cross-country income differences. I model labor markets that are consistent with cross-country data on schooling attainment, education quality, and the average returns to schooling of a country’s emigrants and its non-migrants. The model suggests that the Mincerian returns to schooling of immigrants to the United States measure the education qualities of their source countries. Measured this way, quality differences across countries are large, and the calibrated model shows that schooling accounts for a factor of 5 of the income difference between the U.S. and the poorest countries. The evidence suggests that immigrants to the U.S. are positively selected members of their source country, and that immigrants from developing countries are more selected than those from developed countries. Then the low education quality measured in the sample actually overestimates the education quality of the average non-migrant, particularly for developing countries. Two methods of controlling for selection among immigrants thus predict a moderately larger role for schooling, between a factor of 6.5 and 7.9.