Vulnerability, Climate-Smart Agriculture and Food Security: Evidences from Ethiopia and Uganda

Smallholder agriculture in Sub-Saharan Africa is increasingly shaped by interacting stressors, including climate variability, institutional constraints, and, in some contexts, armed conflict. While existing literature has examined climate vulnerability and adaptation, much of it remains fragmented, often focusing on single outcomes, isolated mechanisms, or relatively homogeneous contexts. This dissertation addresses these gaps through an integrated analysis of livelihood vulnerability, climate information, local institutions, and the interlinkages among climate-smart agriculture (CSA), food security, and mitigation of climate-induced shocks across heterogeneous agrarian settings in Sub-Saharan Africa. The study specifically aimed to: (i) assess livelihood vulnerability across environments characterized by varying levels of conflict exposure and climatic stress; (ii) examine the determinants of access to and use of climate information and assess their effects on the intensity of adoption of climate-smart agriculture (CSA) and other climate change adaptation practices; (iii) analyze the joint determinants of participation in local agricultural groups and CSA adoption among farm decision-makers; (iv) examine the role of community-based institutions in facilitating collective action and adaptation; and (v) investigate the interdependence among CSA adoption, household food security, and mitigation of climate-induced shocks. The dissertation used quantitative, qualitative, and mixed-methods approaches, drawing on evidence from Ethiopia and Uganda. Findings from Eastern Tigray, Ethiopia, show that livelihood vulnerability in fragile settings is driven not only by climatic exposure but also by conflict-induced shocks that erode adaptive capacity and heighten sensitivity. Sensitivity and adaptive capacity contribute more to overall vulnerability than exposure alone, underscoring the importance of incorporating conflict dynamics into vulnerability assessments. Evidence from Uganda analyzed using double hurdle model demonstrates that climate information affects adaptation outcomes through both access and effective use. Education, extension services, labor availability, and participation in community groups significantly enhance climate information use and increase the intensity of CSA adoption. The study also investigated the joint determinants of participation in local agricultural groups and CSA adoption in Uganda using a bivariate probit model. Results indicate that gender, education, land ownership, extension access, climate information use, and drought exposure