This paper analyzes the impact of the Millennium Village Project (MVP) in Sauri, Kenya. We found significant higher
agricultural productivity, self-consumption, production margins, and total (surrogate) income, but an insignificant impact on cash income
among targeted households, when compared to similar ones in neighboring villages. These outcomes are explained by a large allocation
of productivity gains to self-consumption, instead of to the market. Despite showing sizeable effects of the MVP, the results call
for revising of the assumptions about the relationship between agricultural productivity and cash income on which the MVP and other
rural development interventions rely on