IDENTIFYING POTENTIAL REGIONS FOR A PRECIPITATION INDEX INSURANCE PRODUCT IN PARANÁ – BRAZIL: A HIERARCHICAL CLUSTERING APPROACH

In this article the availability and quality of public databases for soybean yields and daily rainfall in the state of Paraná in Brazil is assessed in order to verify the feasibility of an index insurance product. The multiple imputation by chained equations (MICE) method is utilized to fill missing values in the rainfall dataset and study the existence of spatial and temporal patterns in the data by means of hierarchical clustering. The results indicate that Paraná fulfills data requirements for a scalable weather index insurance with MICE and hierarchical clustering being effective tools in the pre-processing of precipitation data