Abstract:
If a health insurance product has only few insured, its claims experience becomes very volatile and is therefore not reliable enough as the only source for repricing the product. Traditionally, a similar product with many insured is used as a reference. However, legislative changes and market forces have led to a fragmentation of products. As a result, such a reference product with many insured is often no longer available. Here we propose a statistical model that combines the data of several products with few insured to derive a common relative claim inflation as well as the expected claims of these products in the future, thus enabling stable pricing for these products. The model was designed so that the usual premium adjustment process is changed as little as possible, making it easy to use in practice.