Farmer Organizations, Spatial Effects, and Farm Household Performances : Econometric Evidence from Senegal
In sub-Saharan Africa, agriculture is a significant source of economic growth and the sector has the largest contribution to poverty reduction. But its development is challenged by the need for institutional innovations to solve problems such as market failures or access to improved technologies. Meanwhile, for decades, collective action groups were considered as policy institutional tools to address these challenges and improve agricultural performance. However, despite the growing interest in these organizations in recent years, impact evaluations of the contribution of farmer organizations are still limited. This study, therefore, attempts to fill in the gap by providing more comprehensive insights on the role of farmer organizations, neighbourhood, and spatial heterogeneity in farm performances. Several methodological approaches were applied, and the main data used for empirical analyses come from a survey conducted in 2017 in Senegal which randomly sampled 4480 rain-fed cereal producing households. The dissertation is a collection of five essays. The first essay examines the empirical causal relationship between membership in farmer organizations and food availability. It applied a generalized spatial two-stage least squares method to control for selection biases and spatial heterogeneity. The results showed a positive and significant association between membership in farmer organizations and households' levels of food availability. The second essay analysed the impact of membership in farmer organizations on household land productivity and income. It applied the Endogenous Switching Regression model to derive treatment effects of membership in farmer organizations. The results showed positive, significant and heterogeneous effects of membership in farmer organizations. The third essay analyses the impact of membership in farmer organizations on rice farms technical efficiency. The essay combined the propensity score matching method with the sample selection stochastic frontier model and the stochastic meta-frontier approach, to mitigate selection biases in the sample and to account for technology heterogeneity. Findings mainly showed that members of farmer organizations do not perform better than non-members. The fourth essay explored the roles and complementarity of neighbourhood and membership in farmer organizations on the adoption of two productivity-enhancing technologies. After applying a Bayesian Spatial Durbin Probit model, the results reveal that close neighbouring farmers show similar choice behaviour regarding productivity-enhancing technologies, and membership in farmer organizations affects significantly and positively the choice of farmers and of their neighbours. The last essay aimed to provide empirical evidence on the Senegalese farmers' technical efficiency in the context of climate variability and spatial heterogeneity. Using simulated data, the paper first evaluated the newly developed spatial stochastic frontier estimation technique based on skew-normal distributions. Moreover, empirical findings reveal that farm technical efficiency appears to be significantly affected by unobserved spatial features. The findings of this dissertation induced some implications for policy and future research. First, support for farmer organizations in Senegal should take into account the spatial distribution of farmers. Second, policymakers when designing programs for rural areas should consider the social links created by both farmer organizations and farmers neighbourhoods. Third, policymakers should encourage more the design and dissemination of agricultural technologies that are very adaptable to specific spatial conditions of farmers. Finally, in the field of spatial stochastic frontier modelling, future studies should continue investigating the performances of the skew-normal approach.