Modelling farm production risk with copulae instead of correlations

The optimisation of production plans is an important topic in agriculture, often related to diversificationand specialisation as the classical instruments of coping with production risk. Althoughthe measurement of embedded risk is often inaccurate, it is nevertheless necessary for decisionmaking to describe the common behaviour of different variables in a model. Impreciselydefined relationships influence the “right” choice, why it is important to find a good approximationof the real circumstances. In financial science, copula functions are frequently used insteadof correlation coefficients to model joint price behaviour, because of the possibility to link themarginal distributions on multifarious ways. By now, agricultural science makes less use of thismethod. This research uses the concept of “partly nested Archimedean copula” to model therelationship between different crop yields and compares it with a correlation based approach.The analysis focuses the differences of the approaches in the context of production planningand the use of weather derivatives.