In the agricultural insurance, the presence of adverse events can compromise the financial health of the agricultural insurance companies. Drought, for example, can affect not only one producer, but thousand of producers in a huge territorial extension. To skirt this problem, insurance companies diversify geographically their operations. This paper applies the cluster analysis in a corn yield data set for the regions of Paraná state, considering the period from 1990 to 2005. In total, 39 groups were formed according to similar characteristics of expected yield and relative risk. Based on one of insurability criteria, some high-risk regions were excluded, and groups were classified according to the degree of relative risk.
crop insurance; cluster analysis; risk diversification
