ABSTRACT
The Basel II Framework allows banks to develop internal frameworks for measurement of operational risk, known as Advanced Measurement Approaches (AMA). Models are subject to regulatory approval, rendering model validation a critical activity, where the validation process shall provide assurance that the measurement methodology credibly reflects the operational risk profile of the bank. A challenge concerning validation of AMA models is the required combination of observational and subjective information. For such models statistical validation alone is not sufficient and other techniques, such as knowledge based validation should be explored.
This paper examines the basic concept of validation in general and of an AMA model in particular. It addresses what it means to validate, and discusses sound principles for validation of an AMA model. The validation should be an integral part of the model development process ensuring transparency and sound use of available resources.
