Resilience is generally understood as the ability of systems to function in the face of disturbance (Holling, 1973). There has been a substantial push to operationalise this concept to reduce the vulnerability of marginalised communities. While development actors across the world recognise the potential of resilience thinking, operationalising the concept presents a number of conundrums. 

Funded by the Department for International Development (DFID), Building Resilience and Adaptation to Climate Extremes and Disasters (BRACED) is one of the largest resilience programmes globally. This initiative aims to help people become more resilient to climate-induced shocks and stresses in South and Southeast Asia, East Africa and the Sahel. Grants have been awarded to 15 consortia, with projects covering a wide range of issues, from securing, servicing and promoting transborder livestock mobility across the Sahel, to sharing skills and technology to improve uptake of climate information in Ethiopia, to supporting smallholder farmers in Nepal to take advantage of economic opportunities and investments in climate-smart technologies (Harvey 2015, forthcoming). 

There is now a growing body of literature that recognises that resilience is highly contextual and pathways to enhancing it vary greatly from one location to the next (Carpenter et al., 2001). The projects funded under BRACED will produce a diverse set of outcomes based on the varied hazards, vulnerabilities and socioeconomic characteristics of the locales in which they are unfolding. 

This paper presents an explanatory conceptual framework for measuring resilient outcomes that embraces and makes sense of this diversity. Outcomes from BRACED projects are understood to be a set of interrelated resilience capacities – the capacity to adapt to, anticipate and absorb climate extremes and disasters (the 3As). The 3As framework can organise practical actions or processes, but which of the 3As they fall into can vary depending on the context, as actions and processes can overlap and interact.