Journal of Development Economics, 2014, 108, pp. 106-118
Misallocations of human capital across sectors are potentially a source of low measured aggregate productivity. Using individual-level data I explore whether there are significant labor market misallocation over ten sectors for a set of 14 developing countries. In each country, there are statistically significant differences between sectors in log wages, even when controlling for education and occupation of the individuals as well as a variety of other factors influencing human capital. Despite the statistical significance, though, the analysis shows that reallocating individuals across sectors until wages converge would yield gains in output of less than 5% for most countries, and this represents an upper bound on the gains due to unmeasured human capital. The microdata allows for inference to be done regarding reallocation, and for eight of the countries I cannot reject the hypothesis that apparent misallocations are simply the result of sampling noise. Overall, frictions in labor markets have a minimal effect on aggregate productivity.
****If you'd like to replicate the results in this paper, please e-mail me. I can send you Stata do-files that will replicate all the figures and tables. You technically need authorization to use the RIGA database, which can be obtained by signing up online with them.***