Weather Index-Based Insurance in Agricultural Risk Management

The increase in the frequency and severity of extreme weather events associated with climate changes has relevant impacts for farmers. Weather parametric insurance schemes are a possible option in managing agricultural risks because they refer to objective and immediate data to assess the payouts (as e.g. weather station data) accelerating time of reimbursement and reducing disputes with respect to conventional crop insurance coverages. These insurance can be considered an attractive opportunity, due to their advantages: low costs, no information asymmetry, abundant data, wide spectrum of activities covered, flexibility. The goal of the study is to investigate the potential benefits that the improvement in the design of insurance solutions (and the predictive analytics techniques) could offer in this area. Specifically, with reference to grape production in two Italian regions, we study which meteorological indices are most suitable as predictors of agricultural production and the predictive efficacy of different models: GLM, Neural Network, Random Forest.