Institutions should validate all EVE and NII models according to the principles articulated in the MMS and in particular related to model life cycle management.
6.6.2
The validation of EVE and NII models should be based upon the principles articulated for both deterministic and statistical models. The validation exercise should ensure that modelling decisions are justified and documented and cover all the model components presented in the previous sections. In particular, the appropriate use of data input should also be reviewed by the validator.
(i)
The validator should ensure that the mechanistic construction of these models is sound. This should be tested with partial replication and internal consistency checks.
(ii)
The validator should ensure that the financial inputs are correctly flowing into these models. This step may require the join work between several teams including the risk and finance teams.
(iii)
The validator should ensure that the results produced by these models are coherent. For that purpose sensitivity analysis can be performed.
(iv)
Finally, some of the inputs are derived from statistical models, including the behavioural patterns observed for non-maturity deposits, fixed rate loans and term deposits. Consequently, the validation should consider the robustness, stability and accuracy of the ancillary statistical models employed to derived inputs to EVE and NII models.
6.6.3
Overall, the validation process of EVE and NII models should focus on the economic meaning and business intuition of the model outputs. The development and validation processes should not be dominated by the mechanistic aspect of these models, but also ensure that those are suitably designed to support robust decision making and the appropriate management of interest rate risk in the banking book.