Measured variation in factor ratios across farm-households is often interpreted as inefficient input allocation, generally reflecting market frictions and leading to cross-country income differences. I evaluate these conclusions using detailed micro data from Ethiopia and Malawi to examine resource allocation both within and between households. I find that about 70% of the overall variation in factor ratios can be attributed to within-household variation rather than between-household variation even after controlling for several observable characteristics of households and plots as well as measurement error. Controlling for these features decreases total inferred misallocation, but equally changes within- and between-household misallocation, leaving the ratio unchanged. I argue that market-level distortions theoretically generate within-household misallocation. In particular, the lack of land markets leads to households with spatially segregated plots, which are hard to optimize across. This accounts for half of the total share of within-household variations. A critical implication of agricultural market distortions is therefore their ability to distort within-household decision making.
Distortions or Sorting in Agriculture Intermediates? Theory and Evidence from Malawi Fertilizer Subsidies
In developing countries intermediate inputs are used less intensively by the poor than the rich, contributing to vast agricultural productivity differences across countries. To what extent is this empirical finding driven by distortions, as opposed to efficient sorting based on complimentary skill or inputs? We develop a dynamic general equilibrium model of agricultural decision making, and show that these two channels can be distinguished by considering farmers responses to an intermediate subsidy. We then utilize the roll out of a large-scale fertilizer subsidy program in Malawi to assess the relative importance of these two channels. We calibrate the model to the year before the subsidies begin, then compute the transition during the introduction of a subsidy program. Utilizing a five year panel of Malawi farmers and variation in program roll-out, we show that the efficient sorting model is inconsistent with the evidence under any reasonable parameterization.
Models of structural change in an open economy predict that countries should specialize in producing in their comparative advantage sector. Therefore, in an open economy (unlike a closed economy), we expect to see higher shares of employment in the most productive sector. This study evaluates the predictions of an open economy model of structural change using US local data, where trade costs are significantly lower than across countries. The analysis starts across the states within the US and gets more disaggregated to the level of counties and metro-areas. As data gets more disaggregated the distances between locations decline, implying lower trade costs and therefore higher degrees of specialization. In this study, I do not find evidence from US local data to support the predictions of the open-economy structural change model.