The (perceived) quality of agricultural technology and its adoption: Experimental evidence from Uganda

Recently, issues related to the (perceived) quality of inputs and technologies have been proposed as an important constraint to their adoption by smallholder farmers in low income countries. Taking maize seed embodying genetic gain as a case, we train random agro-dealers to test whether under-adoption by farmers is caused by low quality due to sellers' lack of knowledge about proper storage and handling. In a second hypothesis, we randomly introduce an information clearinghouse similar to popular crowd-sourced review platforms such as yelp.com or trustpilot.com to test whether information asymmetries crowd out quality seed. We find that the information clearinghouse treatment improves outcomes for both agro-dealers and farmers, with agro-dealers receiving more customers and reporting higher revenues from maize seed sales, and farmers reporting significantly higher use of improved maize seed varieties obtained from agro-dealers, leading to higher maize productivity after two seasons. The primary mechanisms behind this impact appear to be an increased effort to signal quality by agro-dealers and a general restoration of trust in the market for improved seed. The agro-dealer training does not have a clear impact on agro-dealers, nor on farmers in associated catchment areas. However, we do find that the information clearinghouse increases agro-dealer knowledge about proper seed storage and handling. Upon exploring interaction effects between the training and the clearinghouse treatment, we also find that the training becomes effective for agro-dealers that are also in the clearinghouse treatment group. This underscores the importance of incentives to make supply side interventions such as trainings effective.