Comparing experiments for modelling farm risk management decisions with a focus on extreme weather losses

Extreme weather events pose an economic threat to farms. The risk management behaviour against such events is often studied using prospect theory as a framework, but empirically deriving corresponding parameters in the field involving farmers is challenging. To address this issue, we compare three methods of eliciting prospect theory parameters using a multiple price list design in Germany: a framed field experiment, a framed student experiment and an artefactual field experiment. The results show that these experiments generate different prospect theory parameters. The lower the probability the higher the differences, which is particularly important for managing risk from low-probability shocks. Despite these differences, the mean coefficients of the three experiments reveal a low willingness to pay for crop insurance. We find evidence that individual responses to the artefactual and student experiments correlate with the risk attitude self-assessment, whereas responses to the framed field experiment correlate with the purchase of crop insurance.