Effectiveness of the lead farmer approach in agricultural extension service provision: Nationally representative panel data analysis in Malawi uri icon


  • Our results point to two major conclusions. First, LFs support and assist AEDOs in their work, especially in organizing community meetings and farm demonstrations, and are also an important bridge between farmers and AEDOs. But LFs complement AEDOs' work rather than substitute for it. In communities without strong AEDOs and community leaders to work with and monitor them, LFs were not active or performed at a substandard level. Second, results show limited coverage and weak implementation and effectiveness of the LF approach at the national level. Only 13 percent of farmers reported receiving agricultural advice from an LF in the last two years, and only 20 percent reported having interacted with an LF. Our econometric models also consistently show neither the farmers' exposure or interaction with LFs nor farmers' access to LFs' advice had an effect on awareness and adoption of the major agricultural management practices being promoted. When heterogeneity and types of LFs are unpacked, results show that quality of LFs, adoption behavior of LFs, and regular training of LFs have strong and consistent effect on the awareness and adoption of most agricultural practices promoted.
  • The lead farmer (LF) approach has been implemented and heavily promoted nationwide in Malawi since 2009 to support government extension workers and accelerate technology dissemination. Earlier reports have shown that donor-funded projects in Malawi widely adopted the LF approach, indicating positive roles and contributions of LFs. However, national data show persistently low rates of adoption of management practices being promoted by the LFs, prompting this study to look closely at the nationwide implementation and effectiveness of the LF approach. Specifically, we model the effects of farmers' interaction with and exposure to LFs, and farmers' access to LFs' advice, on farmers' awareness and adoption of several promoted technologies and management practices. We use data from 531 randomly selected LFs linked to panel data from 2,880 farming households and, using correlated random effects, model the effectiveness of the LF approach on technology awareness and adoption. This modeling is complemented by 52 focus group discussions and in-depth interviews with agricultural extension development officers (AEDOs) and service providers.

publication date

  • 2020
  • 2020