Learning agro-environmental measures: a sequential model.

Summary The negative environmental externalities of agricultural activities (pollution, reduction of biodiversity) are now better analyzed. They justify agri-environmental policies aimed at reducing their negative consequences, or even generating positive externalities. In the search for a new coherence between these activities and the environment, contractual measures should play a major role, mainly in areas where preservation has a high social value, but requires a considerable change in the operating system of the farmers concerned. This paper attempts to model and simulate contractual incentives in a bounded area, characterized by a given type of externality and heterogeneous farmers who have different interests in the contract. To determine the best contract terms, it is therefore necessary to define models of rational behavior for the actors, the principal (the government) and the agents (the farmers), and to simulate the results relative to the evolution of the principal's control variables (the various incentives). The complexity of this problem requires sequential modeling, which is essential to capture the dynamic interactions between the agents' decisions, the multiple hazards, and to integrate the agents' learning obtained in the first period of the contract.