7 steps to help your CECL adoption succeed

Does the institution have one source of truth for its data? If not, now is the time to create one. A data warehouse, with proper data intake controls, can store all data elements needed across the institution, which helps improve data quality and consistency. The benefits of this go far beyond CECL adoption – some banks have used CECL as a springboard to clean up their data and get more value out of data in all areas. Ultimately, instead of viewing this as a compliance exercise, institutions can consider looking at it as a catalyst to make better business decisions with data.

On-track tip: Data quality and completeness must be taken seriously for a successful transition.

The allowance for credit losses is the most important estimate in an institution’s financial statements, and many ways exist to arrive at management’s best estimate for expected credit losses. As an institution develops the calculation to create this estimate, it should make sure it has picked the most appropriate method. Running the calculation multiple times can give the institution comfort with how the calculation will react to different inputs and assumptions. Running the calculation using different time periods and economic scenarios to see how it aligns with the actual outcome also can be a helpful step. These alternative runs can help identify or quantify qualitative factors.

On-track tip: Look for a modeling approach that strikes the right balance of complexity for the institution and allows efficient implementation to provide more time for parallel runs. Having the agility to support robust, on-demand analysis and sensitivity testing is invaluable. Use stressed scenarios to determine calculation limits and develop contingency plans.

While CECL calculations can vary, no allowance estimate is complete without an evaluation of qualitative factors to adjust for how current or forecasted conditions differ from historical experience. Best practices for creating a Q-factor framework include:

• Reduce, or even eliminate, the use of arbitrary adjustment mechanisms that are difficult to defend or support. If used, these likely should be in place for a short time period and generally should not be material to the overall estimate.
• Evaluate which qualitative factors listed in the standard are sufficiently captured in the baseline calculation and which might be redundant. Clearly document how each one is addressed without duplication.
• Consider using multiple forecast scenarios to evaluate sensitivity to key assumptions over the forecast period. If the institution has determined that the model underperforms in times of economic distress, consider incorporating a more severe forecast as part of the qualitative framework to address this model limitation.
• Create a clear, transparent, and succinct narrative supporting management’s Q-factor assessment, avoiding unnecessary documentation that distracts from the key narrative or that indicates there is double counting or redundancy embedded in the framework.
• Avoid double counting. Under the incurred loss methodology, many institutions had a Q-factor framework that included an allocation for each of the nine factors in the “Interagency Policy Statement on the Allowance for Loan and Lease Losses.” If more robust models are being used for CECL, many of these factors already might be incorporated. For example, if a risk rating or delinquency migration is considered in the base calculation, an additional adjustment for those items might not be needed.

The key is to build a framework that allows for a systematic approach to establishing qualitative factors that can be supported, be repeated, and use available and relevant information.

On-track tip: Q-factors are still important, and it’s vital to identify, and adjust for, what is missing or different from the base calculation and avoid double counting.