Technical efficiency in crop production across agro-ecological zones in Ethiopia: A meta-analysis of frontier studies

Abstract

This article investigates whether study-specific attributes account for systematic variations in reported technical efficiency (TE) scores in crop production. We conducted a meta-regression analysis using mean TE (MTE) estimates from primary frontier studies of Ethiopian crop sub-sector over the period 1991–2015. The estimated MTE of 66% indicates a capacity to increase efficiency in crop production. Results from a fractional outcome regression model revealed that reported efficiency estimates vary substantially across studies and agro-ecologies. We found that reported efficiency estimates are influenced by the frontier methodology used, the functional form assumed, assumptions about technology representation, the estimated dimension of the model, output aggregation and the publication outlet. We show that reported efficiency estimates are sensitive to variations in agro-ecologies. We argue for the need to consider differences in agro-ecologies when estimating TE because failure to account for this may bias efficiency estimates.