Technology adoption and farmer beliefs: Experimental evidence from Tanzania

Adoption of yield-enhancing technologies in Africa is disappointingly low. Many reasons have been proposed for the low observed adoption rates, including credit, insurance and/or information constraints, high transactions costs often associated with poor infrastructure, and imperfect land and labor markets. Furthermore, the variability in returns to technologies can be high and likely contributes to lower adoption rates. For example, soil heterogeneity makes the expected returns to fertilizer application vary. An important but often overlooked consideration in the technology adoption literature is farmers’ beliefs about their soil quality, how closely it aligns with their true soil quality, and how this influences farmer decisions to apply fertilizers. We exploit a randomized controlled trial designed to test the effect of information provision and subsidies on fertilizer use, yields and subjective soil beliefs with three treatment arms: farmers receive a voucher only, fertilizer recommendations only, or both a voucher and recommendations. We find that farmer's subjective soil beliefs and true soil quality do matter to their decision to apply fertilizers. Farmers with underlying objectively good quality soils who received vouchers and recommendations apply more fertilizers six months after treatment, suggesting that farmers are aware that fertilizers are more responsive on better quality soils. In particular, farmers who have an initial low subjective belief of their soil quality and learn that their soil quality is good, are more likely to apply fertilizers and are more likely to apply a greater amount.