Neural Network-based Design of Index Insurance

Weather risk affects agricultural production. Index insurance has been proposed to hedge against severe weather risk, but large basis risk and low demand accompany current piecewise-linear index insurance contracts. We propose embedding a neural network-based optimization scheme into an expected utility maximization problem to design the index insurance. Neural networks capture the highly nonlinear relationship between the high-dimensional weather variables and production losses. We test this index insurance using corn production loss and weather data in Illinois. We show that this index insurance reduces basis risk, increases farmers' utility, is less reliant on government subsidies, and improves social welfare