Multi-dimensional impact assessment for priority setting of agricultural technologies: An application of TOPSIS for the drylands of sub-Saharan Africa and South Asia.

The importance for multi-dimensional priority-setting of agricultural innovations is growing, given that agricultural technologies usually play multiple roles for smallholder farmers. This study assesses agricultural technologies based on their multi-dimensional impacts in the drylands of sub-Saharan Africa and South Asia. The study applies the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) to a set of promising agricultural technologies and uses three outcome criteria: the benefit-cost ratio, poverty reduction, and nutrition security. The technologies are related to important cereals and grain legumes grown in these regions: sorghum, pearl millet, and finger millet; groundnut, cowpea, chickpea, lentil, pigeon pea, and soybean. The results show that the top technologies based on individual criteria can differ from the top technologies identified using a combination of criteria. For example, in semi-arid southern Africa, a promising technology which involves integrated pest management for cowpea ranks among the top five technologies which can reduce poverty. However, the analysis involving TOPSIS shows that nutrition security is more important in that region compared to poverty. As such, the top 5 technologies with the highest multi-dimensional impact for semi-arid southern Africa do not involve a cowpea technology; rather, they are all related to pigeon pea, a nutritious grain legume which is currently more consumed in that region compared to cowpea. One limitation of this study is that it did not consider all the roles of agricultural technologies in the drylands of sub-Saharan Africa and South Asia; this should be considered in future studies involving TOPSIS or other MCDM techniques. Nevertheless, the study shows that TOPSIS can successfully be used for multi-dimensional ex-ante impact assessment of agricultural technologies, and thus can support the prioritization of investments targeting agricultural research for development.