Model Validation: An Industry Perspective

Ben Carr

Insurance companies use models in pricing, underwriting, reserving, claims management, asset/liability management, marketing and distribution, as well as in risk management and economic capital modelling. The use of models will continue to grow over the coming years as a result of the increase in computing power available, the growing importance of online marketing and distribution channels, and the burgeoning use of – and interest in – predictive analytics.

Wherever models are used, it is important that they are validated to ensure that:

• • the models are fit for the purposes for which they are being used;

• • the controls surrounding the operation of the models are adequate;

• • changes to the models are subject to appropriate governance and testing; and

• • the users of the models understand their weaknesses and limitations.

If models are not validated to ensure that these requirements are met, insurance companies will be exposed to undue model risk. If this risk crystallises it could not only have an adverse impact on an insurance company’s short-term financial performance but, in extreme cases, could undermine aspects