When capital and labor are not allocated to the more productive firms, aggregate total factor productivity (TFP) suffers. Can this explain observed productivity differences across countries? We estimate manufacturing TFP levels for 52 developing countries and decompose it into a part due to misallocation and a part due to (residual) technology differences. The results show that removing misallocation would increase TFP by an average of 60 percent, but productivity gaps relative to the US remain large. The degree of misallocation is uncorrelated with observed productivity.