Using internationally comparable data from the World Agricultural Census, we document a factor of 36 difference in average farm size between rich and poor countries. Small farms of less than 2 hectares represent more than 70% of farms in poor countries but only 15% in rich countries, whereas large farms of more than 20 hectares represent none of the farms in poor countries and almost 40% in rich countries. Two questions emerge. First, what explains the striking differences in farm size across countries? Second, are farm-size differences important in understanding agricultural and aggregate productivity gaps across countries? We develop a two sector model with agriculture and non-agriculture that features a non-degenerate size distribution of farms. The theory embeds a Lucas (1978) span-of-control model of farm size into a standard sectoral model with non-homothetic preferences. In the model calibrated to the United States, a reduction in economy-wide productivity from 1 to 1/4 produces an increase in the share of employment in agriculture from 2.5% to 53%, a 21-fold reduction in average farm size, and a 25-fold reduction in agricultural labor productivity. These results are broadly consistent with data on the sectoral allocation of labor and the size distribution of farms across countries.JEL classification: O11, O14, O4.

We use household-level panel data from China and a quantitative framework to document the extent and consequences of factor misallocation in agriculture. We find that there are substantial frictions in both the land and capital markets linked to land institutions in rural China that disproportionately constrain the more productive farmers. These frictions reduce aggregate agricultural productivity in China by affecting two key margins: (1) the allocation of resources across farmers (misallocation) and (2) the allocation of workers across sectors, in particular the type of farmers who operate in agriculture (selection). We show that selection can substantially amplify the static misallocation effect of distortionary policies by affecting occupational choices that worsen the distribution of productive units in agriculture.

There are large differences in transportation infrastructure across nations. Constructing a measure of transportation infrastructure density for a large set of countries, I show that the disparity in this measure between the 5% income rich and the 5% income poor countries is a factor of 28. Are these differences a source of productivity differences across nations? Using a three-sector, two-region, general equilibrium model, I show that high transport costs can distort the allocation of resources not only across geographically dispersed production units within sectors but also between agriculture and non-agriculture. Taking as given the observed differences in transportation infrastructure densities, I quantify the role of transportation for cross-country income differences. The calibrated model produces an income disparity of 10.9 between the 5% rich and 5% poor countries. This corresponds to an improvement of 35% relative to the disparity predicted by a two sector model of agriculture and non-agriculture. Furthermore, the effects of advancements in transportation are non-linear: the elasticity of aggregate labor productivity with respect to the stock of transportation infrastructure in the poorest nations is 15 times higher than in the richest ones.

We use household‐level panel data from China and a quantitative framework to document the extent and consequences of factor misallocation in agriculture. We find that there are substantial within‐village frictions in both the land and capital markets linked to land institutions in rural China that disproportionately constrain the more productive farmers. These frictions reduce aggregate agricultural productivity by affecting two key margins: (1) the allocation of resources across farmers (misallocation) and (2) the allocation of workers across sectors, in particular the type of farmers who operate in agriculture (selection). Selection substantially amplifies the productivity effect of distortionary policies by affecting occupational choices that worsen average ability in agriculture.

All remaining errors are our own. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.