A neural network for evaluating environmental impact of decoupling in rural systems

摘要

This paper aims to assess the environmental impact of a specific disposition of the 2003 Common Agricultural Policy reform of the European Union aimed at decoupling governmental payments from production in the agricultural economy of an Italian rural region during the 2005-2007 period. Effects from the removal of policy income support are also evaluated. For this purpose, a Multilayer Feedforward Neural Network is applied to model farmers' behaviour over the period 2003-2007. Results indicate that the choice of decoupling direct payments from production, compared with the hypothesis of maintaining the previous policy regime, allowed a reduction in the use of chemicals in the regional case analysed. Positive environmental effects were relatively higher in farms that introduced crops only for obtaining short-term gains. Moreover, results show that the total removal of direct payments would have brought about stronger environmental benefits than decoupling, by pushing the most penalised farmers to reorient their production towards alternative activities.