Exploring the effects, mechanisms, and heterogeneity of cluster agriculture on wheat production technical efficiency: evidence from Ethiopia

BACKGROUND The Ethiopian government has promoted cluster farming to improve market access, commercialization, and the adoption of technology. Evidence regarding its impact on the technical efficiency (TE) of wheat production, the mechanisms involved, and the heterogeneous nature of these effects is limited. This study examined how cluster farming affects wheat production TE, the mediating roles of yield per hectare, input access, and selling price, and the heterogeneity of these effects across locations, cluster groups, and the efficiency distribution. RESULTS Survey data from 445 wheat farmers were analyzed. Technical efficiency was estimated using a translog stochastic frontier model and selection bias was addressed through an endogenous switching regression model, supported by robustness checks using a treatment effect model and propensity score matching. Mechanisms were assessed with a generalized structural equation model, and heterogeneity was examined using quantile regression. Cluster‐farming participants achieved higher mean TE (0.842) than non‐participants (0.738). The endogenous switching regression model indicated a positive causal effect (ATT = 0.255), which was consistent across estimation methods. Mediation analysis showed that cluster farming enhanced TE primarily through improved yield per hectare, better access to inputs, and higher farm‐gate selling prices. CONCLUSION The benefits of cluster farming can be heterogeneous, with effects being strongest among low‐efficiency farmers and varying across districts and cluster groups. The aims of policy should be to expand cluster farming participation, strengthen the identified mechanisms of input supply chains, enhance knowledge transfer for yield improvement, and promote collective marketing. Interventions should also be tailored to specific districts while prioritizing low‐efficiency farmers.