Social networks and efficiency in dairy farming: The case of the Program for the Development of Dairy Farming in Minas Gerais, Brazil.

• This paper set out to evaluate to what extent efficiency in production and the position of dairy farmers in the social network, formed by the Program for the Development of Dairy Farming (PDPL), affect the diffusion of technical information and technologies in the production chain. • The efficiency analysis revealed benefits when inefficient dairy farmers sought best practices and the optimal production scale; • Through the model created, it was found that the selection of technically efficient dairy farmers for the diffusion process is a good policy, but when the adoption pay-off is low and resources are scarce, selecting central farmers would be the more appropriate policy. • It was evidenced that the choice of the initial adopters, adoption and non-adoption pay-off values, and the relationship pattern all influence the diffusion process. Farmers' poor levels of knowledge are related to low rates of efficiency and productivity in milk production in Brazil due to problems of information diffusion and innovation. This paper set out to evaluate to what extent efficiency in production and the position of dairy farmers in the social network, formed by the Program for the Development of Dairy Farming (PDPL), affect the diffusion of technical information and technologies in the production chain. The data used in the study were made available by the PDPL technical staff and refer to 31 dairy farmers, located in municipalities in the Zona da Mata micro region of the state of Minas Gerais, who participated in the Program in 2016. Through Social Network Analysis, Data Envelopment Analysis and Agent-Based Modeling, a diffusion model was structured into which information on the technical efficiency of the dairy farmers was incorporated, as well as information about their relationship pattern. The efficiency analysis showed benefits when inefficient farmers sought best practices and the optimal production scale. Through the model created, it was found that the selection of technically efficient dairy farmers for the diffusion process is a good policy, but when the adoption pay-off is low and resources are scarce, selecting central farmers would be the more appropriate policy. Additionally, it was shown that the choice of the initial adopters, adoption and non-adoption pay-off values, and the relationship pattern all influence the diffusion process. The study, thus, suggests a new way of choosing policies to accelerate the diffusion of information and technology in typical dairy systems in Brazil and other countries. [ABSTRACT FROM AUTHOR]