A robust Bayesian analysis of the impact of policy decisions on crop rotations.

We analyse the impact of a policy decision on crop rotations, using the imprecise land use model that was developed by the authors in earlier work. A specific challenge in crop rotation models is that farmer’s crop choices are driven by both policy changes and external non-stationary factors, such as rainfall, temperature and agricultural input and output prices. Such dynamics can be modelled by a non-stationary stochastic process, where crop transition probabilities are multinomial logistic functions of such external factors. We use a robust Bayesian approach to estimate the parameters of our model, and validate it by comparing the model response with a non-parametric estimate, as well as by cross validation. Finally, we use the resulting predictions to solve a hypothetical yet realistic policy problem