Measuring the Technical Efficiency of Cotton Farms in Turkey Using Stochastic Frontier and Data Envelopment Analysis

Abstract

The determination of technical efficiency for cotton farms can be invaluable in estimating optimum farming practices and in identifying strategic options for the industry. This paper investigates the productive efficiency of a sample of farmers in Turkey's Aegean region by estimating a stochastic frontier production function (SFA), constant returns to scale (CRS) and variable returns to scale (VRS) using output-oriented data envelopment analysis (DEA). Data were obtained from 198 cotton farms using structured questionnaire interviews. The estimates of technical efficiency based on these two frontier methods were compared. While efficiency scores for cotton farms differed between the SFA and the DEA models, the mean efficiency scores are quite low for the CRS DEA model compared with the VRS DEA and SFA approaches. The mean efficiency measure (0.91) obtained from the stochastic frontier was higher than that calculated from the VRS DEA (0.77) and CRS DEA (0.25). This study suggests that more efficient political instruments need to be adopted to review current subsidies because of increasing outlays for diesel oil used in cotton farming.