This article examines the functioning of a Community-Driven Development (CDD) project that, unlike most CDDs, was set in a homogeneous community, consisted of simple tasks, had a year-long pre-intervention planning period, and maintained records of its decisions. To identify the impacts of the intervention on agricultural outcomes, and on strategies to manage risk, we use a semi-parametric difference-in-difference model that collected retrospective panel data on programme participants and non-participants over multiple seasons. Our findings suggest that the CDD project led to greater exposure to risk and no gains in productivity for the community.

Journal Name: The Journal of Development Studies