Revisiting the Efficiency–Poverty Nexus: Evidence from Kenyan Smallholder Maize Farmers

Abstract

The relationship between farm productivity and poverty reduction remains contested in settings characterised by persistent spatial inequality. This study examines whether technical efficiency is associated with multidimensional poverty among 8,434 smallholder maize farmers in Kenya, using Tegemeo Institute survey data collected under the Kenya Climate-Smart Agriculture Project during the 2020-2021 agricultural season across nine counties. Technical efficiency is estimated using a stochastic frontier approach while its effect on multidimensional poverty is analysed using an instrumental variable Tobit model with county fixed effects. Results indicate a mean technical efficiency level of 35.5 percent. In pooled specifications, technical efficiency is associated with lower multidimensional poverty; however, this relationship disappears once county-level heterogeneity is accounted for, suggesting that baseline results are driven by between-county variation rather than within-county differences. Further evidence from the first-stage analysis indicates that the identifying variation in the instrument is largely cross-county, implying weak within-county identification under fixed effects. Overall, multidimensional poverty is strongly shaped by spatial and structural county-level conditions, highlighting the importance of context-specific constraints in determining multidimensional poverty outcomes. The findings suggest that efficiency-enhancing interventions alone are insufficient for poverty reduction unless complemented by spatially targeted policies addressing infrastructure, market access, and institutional disparities.