Using Stochastic Frontier Analysis (SFA), this paper centered on analysing the impact of Crop Diversification (CD) on farm level technical efficiency in Afghanistan. Data from a household level survey conducted in 2013-2004 by the Central Statistic Organization (CSO) is used in the analysis. The results revealed that adoption of a diversified portfolio of crops by the farmers significantly improves technical efficiency. In addition, access to extension services, farm size, cattle, oxen and tractor ownership by the farm households, and regional variables were other important factors that significantly affect technical efficiency. It is evident from the results that the estimated technical efficiency indices from the preferred truncated normal distribution range from 1.5% to 99.29%, with a sample mean of 71.9%. The basic SFA model was investigated for potential endogeneity in crop diversification. Instrumental Variable (IV) method was employed to correct for endogeneity in crop diversification. The results of the IV estimation reveal that failing to account for endogeneity in the basic model leads to a downward bias which is consistent with attenuation bias (measurement error in CD implies that OLS coefficients are biased towards zero, so one would predict IV coefficients greater in absolute size). The results of crop diversification index showed the presence of a relatively low level of crop diversification. Maximum likelihood estimation of translog stochastic frontier model shows that land, labour, and other purchased inputs (fertilizer, seeds, pesticides usage) have positive impact on farm revenues. The results show an evidence of constant returns‐to‐scale.
Download the paper in PDF format
View all CREDIT discussion papers | View all School of Economics featured discussion papers
Sir Clive Granger BuildingUniversity of Nottingham
Nottingham, NG7 2RD
+44 (0)115 951 5620
Connect with the University of Nottingham through social media and our blogs.
Campus maps | More contact information | Jobs
Browser does not support script.