Answering 5 top CECL model validation questions
- The current expected credit loss (CECL) is an accounting model used by financial institutions to estimate expected losses from outstanding loans and reported on in quarterly call reports.
- Institutions need regular CECL model validation to find flaws or errors before they can be spotted by regulators and show up in an examination or audit report, triggering questions from shareholders.
- CECL model validations help you ensure you’re applying your methodology correctly, using the right model inputs and producing correct outputs.
- An effective CECL validation framework involves three essential pieces: evaluating your CECL model’s design and construction, conducting ongoing monitoring and assessing outcomes.
- Financial institutions should work with a third-party audit and accounting firm that has a deep understanding of how to validate and implement different CECL methodologies and do a validation every 1-3 years.
Now that the CECL implementation date has passed and financial institutions are beginning to report under the CECL framework in their quarterly call reports, considerations around validating the CECL model are becoming increasingly important. But since this is still so new, many are asking questions around CECL model validations before they get started.
Below are answers to these common CECL questions:
- Why do you need a CECL model validation?
- What are the benefits of a CECL model validation?
- What procedures are required in a CECL model validation?
- How often do you need a CECL model validation?
- What should you look for when choosing a third-party CECL validation firm?
Why do you need a CECL validation?
As with other significant models used by financial institutions, the primary reason for obtaining a validation is to manage model risk and to satisfy regulatory expectations. Because the allowance for credit losses is a key financial reporting estimate, most institutions will likely identify the CECL model they are using as a key model.
Many CECL models are complex, and regulatory guidance suggests that model complexity generally increases model risk. Given the importance of CECL models as well as their potential complexity, we anticipate that regulators will generally expect management to perform periodic valuations to effectively manage model risk.
The last thing you want is for regulators or auditors to come in and identify flaws in your CECL calculation that would need to be reported in your examination or audit report, triggering questions from shareholders and making it appear that management hasn’t been proactive in managing risk.
You’ll want to know now — far before you get to your year-end reporting — whether there are meaningful flaws in your CECL process.
What are the benefits of a CECL validation?
Effective model validation helps reduce model risk by identifying model errors and verifying the reliability and proper application of a given model. By having a CECL model validation done, you can verify that the methodology you’ve selected is being applied appropriately, the model inputs are coming from appropriate sources, and the calculations and resulting outputs are correct. And if everything isn’t functioning as it should, you can then identify the areas where you need to implement fixes and build additional documentation and processes.
The sooner you have a CECL validation done, the more secure you can feel going forward that your model is working correctly and won’t cause problems in your year-end audit or next regulatory examination.
What are the CECL model validation requirements?
Regulators have issued guidance on model risk management, which provides detailed considerations related to model validation. The guidance on validations details that an effective validation framework should contain three core elements.
- Evaluation of conceptual soundness: This covers the quality of the model design and construction. Validation procedures should ensure that judgment was exercised in model design and that model application is consistent with industry and accounting practices. This may also involve sensitivity analysis of key inputs. For an institution performing its first CECL model validation, this element will likely be the most significant.
- Ongoing monitoring: The validation should confirm that the model is properly implemented and being utilized as intended. Ongoing monitoring is key to evaluating changes in conditions that may require adjustment to the model and to verify that any extension of the model beyond its original scope is valid. Ongoing monitoring should also include analysis of overrides to the model to ensure appropriate documentation.
- Outcomes analysis: This element of the validation process involves a comparison of model outputs to actual outcomes. Back-testing is one form of outcomes analysis performed in many model validations. For many CECL models, there may be limitations on the precision of the comparison given CECL models are evaluating estimates of lifetime loan losses; however, comparison of loss estimates computed by the model to actual loss history may provide useful information in assessing the model.
Regulatory guidance suggests that the extent of an institution’s validation testing needs to be commensurate with the risk level of your model. Along with feedback from regulators and auditors, this will also impact the frequency in which CECL model validations should occur. We anticipate most CECL models will ultimately be validated every 1-3 years, meaning less complex models will likely not require annual validations, absent other risk factors.
What should you look for when choosing a CECL validation firm?
When choosing a third party to perform your CECL validation, you want to look for the right expertise and experience.
The firm you choose should understand the accounting standard, the validation guidelines and the different CECL methodologies out there on a deep level. CECL is complex, and a firm that has experience implementing or validating different methodologies knows what to look for and can more easily identify issues.
This firm should have a flexible validation program built to fit your CECL model and the associated risks. The validation of a simple weighted-average remaining maturity model should be structured differently than a model utilizing a probability-of-default calculation capturing peer data. You don’t want a firm that takes a one-size-fits-all solution approach.
How Wipfli can help
Discover whether your CECL model and methodology is functioning as expected with a CECL validation from Wipfli. Our CECL specialists have performed dozens of validations already and have experience with different methodologies and models used by financial institutions. Learn more about our validation service.
Sign up to receive additional financial institutions content and information in your inbox, or continue reading on:
