Become a Pool Member!
The validation of internal models is an integral part of operating rating models. Under what we call the Pool Approach, several institutions co-operate to develop a common rating model using the same data and rating philosophy. The advantages of such an approach are obvious: The joint analysis and modelling enable a joint handling of changes – which need to be implemented only once. This creates economies of scale and thus cost savings for the participating institutions. On the other hand – and this is the more important part for the considerations below – the combined and significantly larger dataset available leads to a permanently more reliable identification and quantification of the relevant risk factors than would be possible for an individual institution with its naturally smaller and more limited database.
Unfortunately, these advantages come at the price of new challenges: How does an institution’s data fit with the rest of the data pool, how well does the model work for an institution’s specific portfolio? These questions suddenly become the focus of validation from the institution’s point of view, especially when the actual model validation initially takes place at pool level. In addition, the institution remains responsible for validating the system for its own portfolio according to all relevant regulatory requirements. A clear division of tasks and responsibilities in this context is imperative.