Abstract
Climate-smart agriculture (CSA) is increasingly important for advancing rural development and environmental sustainability goals in developing countries. Over the past decade, the international community has committed billions of dollars to support various practices under the banner of CSA. Despite this effort, however, CSA adoption remains low in many contexts. Lack of conceptual clarity about the range of potential farm-level CSA practices across contexts impedes understanding of CSA adoption in developing countries. Here we review relevant literature to develop a typology of farm-level CSA practices to facilitate analyses of CSA adoption. The typology consists of six categories, organized from least to most resource intensive: (1) residue addition, (2) non-woody plant cultivation, (3) assisted regeneration, (4) woody plant cultivation, (5) physical infrastructure, and (6) mixed measures. We use the typology to generate and test hypotheses about CSA adoption using primary household survey data from a large aidfunded CSA intervention area in southern Malawi. We then use recursive bivariate probit regression (controlling for endogeneity and selection bias) to estimate the effect of program participation on adoption across CSA categories. We find positive and statistically significant effects of program participation on adoption of CSA practices generally with the strongest effects on resource-intensive CSA categories. Results demonstrate the potential for wider application of the typology to build knowledge of the effectiveness of CSA promotion efforts across different social and environmental contexts. Our findings also suggest the importance of external support for the adoption of more resource-intensive CSA practices among rural households and communities in Malawi and elsewhere in the developing world. (C) 2019 Elsevier Ltd. All rights reserved.