Articles & E-Books


Summary and analysis of the Fed’s Scaled CECL Allowance Estimator

Jul 20, 2021

On July 15, 2021, the Federal Reserve hosted a webinar on its new tool, the Scaled CECL Allowance for Losses Estimated (SCALE) method.

Designed for smaller, less complex institutions, the SCALE method is described by regulators as one of many acceptable methods for applying CECL, but may not be used by banks over $1 billion in assets. The SCALE model attempts to simplify the process of evaluating the allowance under CECL by allowing banks to leverage peer data from publicly available regulatory reports instead of calculating a lifetime loss rate based on individual bank experience.

Components of the SCALE method

The SCALE method includes 4 key components:

1. Lifetime loss rate

The lifetime loss rate is the key input in the SCALE method. Whereas most CECL models require calculating an estimated lifetime loss rate by considering the institution’s historical loss data, the SCALE method instead uses aggregated call report data. All institutions that have adopted CECL and are over $1 billion in total assets are required to disclose the amortized cost and related allowance balance in part II of schedule RI-C on the Call Report. The average loss percentages (calculated by taking the allowance balance divided by amortized cost) of a selected peer group are used as the lifetime loss rate.

The regulators have suggested that it is up to the institution to determine the appropriate peer group for computing average loss percentages. In some cases, a national average of all reporting banks may be appropriate, while in other circumstances, a more personalized peer group may be beneficial. It should be emphasized that the loss percentages computed from the peer group already incorporate the peer group’s evaluation of qualitative adjustments, reasonable and supportable economic forecasts, and estimated prepayments. As a result, institutions using the SCALE method will potentially require less work in evaluating those components of the calculation.

2. Individually assessed loan losses

The CECL standard requires loans with dissimilar risk characteristics to be evaluated individually. The SCALE model provides a worksheet specific for tracking such loans. Loans evaluated individually are removed from the pool, and impairment amounts are calculated individually. In the SCALE method, individually assessed loans are accounted for very similar to impaired loans in incurred loss allowance calculations.

3. Qualitative adjustments

The SCALE method has institutions make qualitative adjustments for information not already captured in the peer group expected lifetime loss rate, including any forward-looking adjustments unique to the institution. The webinar pointed to the Interagency Policy Statement on the Allowance for Credit Losses for a list of potential factors for management to consider, which is similar to current practice for many banks under the incurred loss method.

While the SCALE model includes inputs for qualitative adjustments, it’s expected that the support for the qualitative adjustment is determined and documented outside of the model. As noted above, the peer group rate calculated in component 1 already includes the qualitative adjustments by the peer group. It is left to the discretion of the institution as to how that should impact its evaluation of additional qualitative adjustments

4. Adjustment to peer group lifetime loss rate to reflect the institution’s insight on portfolio performance

The SCALE method prompts for input of the institution’s net loss to average loan rate from 2007-2020 and also the UBPR peer groups net loss to average loan rate over the same years. These loss rates are compared, and an adjustment is calculated based on the differences (for example if an institution’s loss rate is less than UBPR peer, a negative adjustment is calculated). This calculation is prepared on the portfolio as a whole; however, the regulators indicated it may be appropriate to expand this analysis and make specific adjustments for each individual loan segment. Regulators also indicated that it may be appropriate to consider a different peer group than those included on the UBPR.

Acceptability of the SCALE Model

The Fed has described the SCALE method as “one of many acceptable methods of applying CECL.” Although the model has been rolled out by the Fed, they indicated that other regulatory agencies have also reviewed the SCALE model, and it would appear to have applicability regardless of an institution’s primary regulator. FASB was also represented on the Fed webinar. If applied properly, it would appear that the SCALE method would be consistent with generally accepted accounting principles.

Like other CECL methods, each institution must individually determine whether the SCALE method is appropriate for its loan portfolio. The SCALE method is not a “safe harbor” from regulatory scrutiny over the allowance process, and using the SCALE method does not ensure compliance with generally accepted accounting principles and regulatory expectations. Each institution will need to calculate appropriate qualitative factors within the SCALE model framework to appropriately adjust the peer group loss rates to a rate that is appropriate in relation to the institution’s individual credit quality.

Wipfli’s thoughts

The SCALE method was clearly designed to provide a simple methodology for smaller institutions with limited loan complexity. One of the primary challenges institutions face in developing a CECL model is accumulating the necessary data to calculate a lifetime loss rate. The SCALE model substantially eliminates this challenge by replacing institution-specific loss rates with peer-group loss percentages from Call Report data.

Another substantial challenge in adopting CECL was the development of reasonable and supportable forecasts of future losses. Again, the SCALE model mitigates this challenge by using peer loss rates that have built-in forecasting assumptions.

Despite that, the SCALE method has some limitations, and as such, may not be an ideal choice for all institutions. Most obviously, the SCALE method is designed to use peer loss rates as its starting point. In many cases, peer data may be less precise than loss rates based on an institution’s specific loss experience. This will likely elevate the importance of thoughtful qualitative adjustments to ensure peer loss rates are appropriately adjusted for the unique risks of the institution.

In addition, the SCALE method is not permitted for banks in excess of $1 billion. Banks approaching that threshold will need to develop an alternative CECL approach to comply with requirements.

For institutions under $1 billion with less complex loan portfolios, we believe the SCALE method warrants consideration. Institutions struggling to accumulate the necessary data to complete other CECL methodologies, or that lack the resources to effectively develop a model, will likely appreciate the SCALE model’s simple approach using readily available peer data. However, institutions with more complex portfolios or with greater resources available to devote to CECL may find that other methodologies using institution-specific data will provide a more precise and more easily supportable loss estimate.

If you have any questions about the SCALE method or need assistance applying it, contact Wipfli.

Sign up to receive additional financial institutions content and information in your inbox, or continue reading on:

Podcast: CECL’s impact on financial institutions

Measuring the impact of CECL adoption

CECL adopters vs. non-adopters: Comparing ALLL to loans


Nick G. Ansley, CPA
View Profile