Shane M. Sherlund; Christopher B. Barrett; Akinwumi A. Adesina
Type of Document:
Journal of Development Economics (forthcoming)
Date of Publication:
Place of Publication:
Abstract: Smallholder agricultural production depends heavily on environmental production conditions that are largely exogenously determined. Yet, few data sets collect necessary, detailed information on environmental production conditions. This oversight raises the spectre of likely omitted variables bias because farmers’ input choices typically respond in part to environmental conditions. Moreover, because environmental production conditions are rarely symmetrically distributed, the omission also Generally leads to upward bias in estimated technical inefficiency and to biased estimates of the correlates of estimated technical inefficiency as well. Using panel data from 464 traditional rice plots in Cote d’Ivoire, we show that controlling for heterogeneous environmental production conditions significantly changes inferences, perhaps especially with respect to smallholder rice farmers’ estimated technical inefficiency.