Model validation is sometimes perceived as the less glamourous and perhaps even obstructive bit of the model risk management (MRM) workflow. Certainly, done wrong, it can feel that way to the model development team. The validation team are the ones who keep finding problems with the models and telling everyone about it, including management and sometimes even the regulator.
But the model development team needs to ask themselves, who are being the obstructive ones? With the right institutional, cultural, and technical framework in model development, and across the entire model lifecycle, model validation can have an enormous added value.
Software tools and workflows reinforce a culture of collaboration and partnership between development and validation, while keeping the discrete separation of responsibilities intact. The model development team knows what the validation team expects to receive, and how they expect to receive it. This facilitates an organized, iterative, and efficient validation process.
- Model templates help developers build models in a consistent style using wellstructured code.
- Documentation templates ensure that every aspect of the model is well documented.
- Structured, scripted, executable models allow for an iterative, agile development process.
- Full model lineage through model inventory eliminates confusion about the versioning of code and data.
- A secure, controlled environment allows appropriate access to sensitive data.