Internal risk models are widely used by insurance companies and banks in measuring risk, calculating overall capital requirements, and allocating capital for business decisions. In addition, the valuation and pricing of complex financial and insurance products is regularly carried out using sophisticated mathematical models because of the unavailability of market prices for these products. The use of models for these purposes introduces risk through the potential for model error. This is especially true in situations where the financial framework for the operation of a business relies substantially on the results of a model.
The intent of model validation is to limit the risk that use of a model could mislead management into making poor decisions. That risk can be referred to as “model risk.” The first part of this paper identifies and describes five distinct elements of model risk and then outlines a model validation process with specific steps to address each element.
Model validation often takes place in a dynamic environment where changes in reporting relationships and changes in models are taking place continuously. The second part of the paper discusses the challenges faced when performing model validation in a dynamic environment, with a focus on the risk management control cycle.