CECL: Key Model Risk Practices
2/25/2020
With extra time to prepare, there will be a higher expectation from the regulatory environment as it relates to your CECL methodology(s). Model Risk Management practices will be critical to your organizations success. It is important to remember Model Risk Management encompasses both a quantitative and qualitative approach. We're highlighting some key Model Risk Management practices that we recommend to help you successfully prepare for a validation.
Develop Policies and Procedures: Ensure clearly defined policies are approved by the Management Body. We recommend a clearly defined policy for the following:
- Roles and Responsibilities
- Data Maintenance and Access Permissions
- Model Documentation
- Model Validation Processes
- Updates or Changes to Model
Assign a Project Manager: Select a key leader in your organization who will serve as the single point of contact for validation needs. For example, If Valuant is your partner in the validation of your model(s), this will be Valuant's main point of contact for any questions or issues as they arise over the course of the project. It is also important to note, if Valuant is performing validation on a Vendor model(s), the Bank's Project Manager may act as a liaison between Valuant and the Vendor for any questions related to the project. It is a best practice to notify your Vendor of any validations being performed to ensure there is an open line of communication as needs arise.
Establish Ongoing Monitoring: While your model(s) may be brand new, now is the time to establish the policy for monitoring in the future. Evaluate process and policies of continuous monitoring of model(s), we recommend at least annual frequency. This monitoring may include:
- Review of assumptions and risk factors of the model
- Assessment of calculations and all components of the model for accuracy
- Evaluation of back-testing strategy, ensure the model is functionally correctly as economic scenarios evolve
No Model is Perfect. Don’t be afraid to document model limitations. It is better to point out and document model limitations and address how such limitations are mitigated.
It's never too late to get started with Model Validation! Learn More