I study how the relative efficiency of high- and low-skill labor varies across countries. Using microdata for countries at different stages of development, I document that differences in relative quantities and wages are consistent with high-skill workers being relatively more productive in rich countries. I exploit variation in the skill premia of foreign-educated migrants to discriminate between two possible drivers of this pattern: cross-country differences in the skill bias of technology and in the relative human capital of skilled labor. I find that the former is quantitatively more important, and discuss the implications of this result for development accounting. (JEL I26, J24, J31, J61, L16, O15)