Résumé
In this investigation, we question the consequences of a top down decentralization of irrigation system in Tunisia on the cooperative behavior of farmers. We examine whether the voluntary adhesion to the provision of a collective does affect the cooperative behavior. This is accomplished by comparing the provision of a public good and the provision of a club good, both of them with a provision point. The investigation is performed in the lab and in the field. Our experiment reveals that voluntary adhesion does increase group contributions, success of provision and welfare. It also increases the number of contributors and moderates cheap riding. A possible explanation to our results is that voluntary adhesion reduces strategic uncertainty. Our investigation shows also that voluntary adhesion is more effective when it deals with an assurance problem. When we rule out the assurance problem with a Money Back Guarantee mechanism, voluntary adhesion does no longer increase group contributions, success of provision and welfare. However, it decreases variance of group contributions mainly in the high threshold. Voluntary adhesion with a refund mechanism does also increase the number of contributors. The investigation of the external validity of our laboratory results shows that our findings hold with respect to the origin of subjects. Subtle differences exist in an experiment between subjects in Tunis and in Montpellier with lower variance of group contributions and a higher number of contributors but no dramatic change are observed. In contrast, the field experiment reveals less effective findings in the voluntary adhesion treatments with Tunisian’s farmers. Success of provision and welfare increases only in one sample of farmers out of three. Nonetheless, our field experiment reveals that voluntary adhesion increases the number of contributors in all the treatments performed. It also shows a high level of cooperation among farmers that sustain over time