Seed Beliefs and Fertilizer Use: On-farm Misallocation due to Seed Misclassification

• Development Workshop

Date and Location

Tuesday, May 16, 2023, 4:10 PM - 5:30 PM
ARE Library Conference Room, 4101 Social Sciences and Humanities

Abstract

Optimal input allocation in agriculture leverages production complementarities. Improved seeds are generally more responsive to fertilizer than landraces. Inaccurate beliefs about whether sown seed is improved may therefore result in sub-optimal fertilizer application. We document precisely this pattern using unique data from Ethiopia that includes farmer beliefs and genotyping tests of maize seeds. Farmers purchase and apply fertilizers based on their beliefs about their seeds, which often deviate from their true genetic identity and associated production characteristics. These results suggest that asymmetric information about agricultural input quality can trigger input distortions relative to farmers’ intent, raising on-farm production costs and potentially hampering learning and subsequent technology adoption. Using a machine learning approach to extrapolate the extent of seed misclassification nationwide, we estimate that a third of the maize area was sown by farmers with false seed beliefs and discuss implications for national-level nitrogen use.