Heterogeneity, Measurement Error, and Misallocation: Evidence from African Agriculture

How important is misallocation in explaining the income differences across countries? The dispersion of productivity across production units seems to be higher in poor countries than in rich countries, which might account for some of the differences across countries in income. We examine this hypothesis in agriculture. There is enormous dispersion in output per factor input and in factor input ratios across farms in Africa. However, measurement error, unobserved heterogeneity in input quality and transitory productivity shocks as well as misallocation can generate dispersion in measured average productivity. We take advantage of rich data on inputs and outputs of the plot level in three African countries to disentangle the productivity dispersion that arises from misallocation from that arising from measurement error, unobserved heterogeneity or transitory shocks. We show that since farmers face no market imperfections in allocating resources across their own plots within the growing season, the dispersion of measured average productivity across the plots of a given farmer and the correlations between that measured productivity and the farmer’s input decisions identifies the magnitudes of measurement error, unobserved heterogeneity and transitory shocks. When these sources of observed variation in average productivity are taken into account, the overall importance of misallocation drops substantially.

Written with Douglas Gollin, University of Oxford