Membership in farmers’ organizations and intention to innovate: A mixed random utility and behavioral approach

Despite a growing body of literature on the adoption of agricultural innovations, their uptake by smallholder farmers in developing countries is often slow. The processes underpinning farmers’ decision-making in these countries are yet to be fully understood, and the existing literature remains contradictory regarding the factors explaining adoption. Conflicting conclusions emerge from studies on different countries, which may result from specific social, cultural, and institutional environments. We develop a theoretical model that combines random utility and behavioral approaches to assess how membership in farmers’ organizations affects the intention to innovate. Using a two-step framework, we assume that membership in such organizations acts as a mediator variable between background factors and the intention to innovate. Three distal behavioral factors – openness to new ways of production, attitude toward risk, and trust in organizations promoting innovations – are considered additional intermediate drivers. We test our framework using primary survey data from five African countries, covering a total of 4,529 farmers and more than twenty farmers’ organizations. We find that generic organizational membership has limited mediating power and a marginally positive impact on the intention to innovate, which becomes non-significant when accounting for the behavioral attitudes of individual farmers. These attitudes do work as mediator variables between the background factors classically included in random utility models (i.e., farm, household, and farmer’s characteristics) and the intention to innovate. In turn, membership in some specific organizations proves to be a significant predictor of the intention to innovate, although the direction is not univocal. A specific institutional approach is thus needed to evaluate which characteristics of a farmers’ organization impact its members’ intention to innovate. We provide some hypotheses based on local knowledge.