Quantity Recommendation as a Solution to Imbalanced Fertilizer Use? A Field Experiment in Bangladesh

We use a field experiment in Bangladesh to test if two types of variety-specific fertilizer recommendations---government provided, community-level recommendations and plot-specific recommendations based on individual soil tests---affect fertilizer use, yield, and profits. Treatment farmers do not change their fertilizer application behavior in response to either recommendation. One exception is the case of TSP (phosphate fertilizer), which is recommended for all crops but recommended quantities are significantly lower than the baseline average usage. Treatment farmers over-react to the recommendations by stopping TSP use after the intervention. In the soil-testing treatment arm farmers also shift their seed choice to varieties for which their baseline fertilizer consumption aligns with the recommendation. Opting out of using TSP, an essential fertilizer, ultimately hurts productivity as farmers in the community-based recommendation arm experience a 6.6% yield reduction. Attention costs may explain the extensive margin but no intensive margin effect in response to the treatments