Eco-efficiency offers a promising approach for the sustainable intensification of production systems in Sub-Saharan Africa. Data Envelopment Analysis (DEA), which is widely used for eco-efficiency analyses, is however sensitive to outliers and the analysis of the influence of external factors in the second stage requires the separability assumption to hold. Order-m estimators are proposed to overcome those disadvantages, but have been rarely applied in eco-efficiency analysis.
This paper assesses the eco-efficiency of smallholder perennial cash crop production in Ghana and Kenya. It examines factors influencing eco-efficiency scores and in doing so, tests the application of order-m frontiers as a promising method for eco-efficiency analysis in the agricultural context.
The analysis is performed for four selected perennial crop cases, namely cocoa, coffee, macadamia, and mango, applying DEA as well as the order-m approach to a comprehensive empirical dataset. Seven relevant environmental pressures as well as determining factors around capacity development, farm and farmer features, and crop production environment are considered.
The distribution of eco-efficiency estimates among coffee farms showed the widest spread, which indicates the greatest potential to increase eco-efficiency. However, also the dispersion of scores within the other crop cases suggests room for improvements of eco-efficiency within the current production context. The subsequent analysis of determinants based on the order-m scores revealed that eco-efficiency scores were strongly influenced by variables, which measure capacity development, and resource endowments, such as labor and land, whereas the crop production environment had some influence, but results were unspecific. Generally, a positive effect is highly context-specific. The results underline the importance of designing effective training modalities and policies that allow knowledge to be put into practice, which involves the creation of marketing opportunities, the provision of targeted and regular advisory services, as well as region-wide measures to build and maintain soil fertility in a sustainable manner.
To our knowledge, this study presents the first attempt to apply inputoriented order-m frontiers to assess eco-efficiency in the agricultural context, comparing its eco-efficiency rankings to those estimated with the widely applied DEA approach. This can inform the discussion on robust eco-efficiency assessments.