A subcommittee of the AICPA task force focused on the current expected credit loss standard (CECL) has provided its first look at how auditors will approach the changes with clients, noting that CECL represents a “fresh start” from the incurred loss model.
“CECL model estimates will be evaluated against ASC 326, not anchored to incurred loss model estimates,” according to published takeaways from an AICPA CECL Task Force Auditing Subgroup meeting. “Management may find it useful in the context of validating their CECL model to understand what drove changes from ALLL levels today to ACL estimates under ASC 326. However, management should be aware of potential anchoring, confirmation, availability bias that might occur when implementing the new standard.”
“CECL is a ‘fresh start’ from the incurred loss model,” the subgroup said in the April 29 meeting notes, which members approved at another meeting on June 27.
Garver Moore, Managing Director of Abrigo Advisory Services Group and one of three Abrigo advisors invited to provide input at both meetings, said the commentary from the Auditing Subgroup will be helpful to financial institutions working to implement CECL under FASB’s effective dates.
“What they’re making clear is that so long as your estimate is made in good faith and you’re following the new standard, which is different than using the incurred loss method, auditors are not just going to say, ‘Hey, your previous number for the allowance was this, so your new number should be at least that much,’ ” Moore said.
Moore said the Auditing Subgroup has been meeting to align expectations of financial institutions, auditors and other interested parties as they relate to CECL. SEC-registered financial institutions are expected to implement the new standard in Q1 2020; all other banks or credit unions have until the following year or later.
Additional overarching themes related to CECL released by the Auditing Subgroup in the meeting notes included:
The Auditing Subgroup said that appropriate governance processes and controls by management are critical when it comes to CECL. “Methodologies and models should be well documented and supported,” it said. “Management must have an appropriate model selection/validation process and cannot solely rely on vendors.”
The Subgroup also noted that management “may outsource model development activities, but not overall responsibility for the appropriateness of the model.” An auditor will expect management to have a full understanding of their models and be able to explain inputs, assumptions and methodology. The Subgroup also noted that transparency into third-party models is important, both for internal controls and for auditor purposes.
“Vendors will likely play an important role in support (including training management in both the standard and workings of the model), but management cannot rely solely on the vendor,” the Subgroup said. “Management should own the model and have a clear understanding of how the model works and the data utilized in their model.”
The Auditing Subgroup emphasized that financial institutions will need to evaluate whether a CECL model selected is appropriate for the specific portfolio for which is it being used. This should be part of an ongoing, robust evaluation of an institution’s CECL model, according to the AICPA’s subcommittee.
Another aspect of that evaluation is related to qualitative adjustments. “Although CECL models may make greater use of quantitative data than incurred-loss ALLL models today, qualitative judgments and adjustments are foundational to arriving at a reasonable estimate of expected credit losses,” the Subgroup noted. Like the allowance for credit losses itself, qualitative adjustments “should be independently generated and not anchored to, or grounded in,” those used in the existing incurred loss model, according to the meeting notes. Auditors will be looking for management to support not only directional consistency in quarter-over-quarter changes in qualitative factors, but also amounts of the adjustment itself.
Retrospective reviews of CECL models will be helpful in refining methodology and inputs for future forecasts, the Subcommittee noted. However, the group acknowledged that due to the complexity of CECL and the many variables involved, “an outcome that is different than previously estimated does not necessarily mean that the previous estimate was incorrect or that there were errors in the estimation process.”
“This Auditing Subgroup is taking a reasonable approach to many of these aspects of CECL,” Abrigo’s Moore said. “Members understand that financial institutions and other entities implementing CECL are seeking clarity on how their external auditors are going to view certain aspects of the standard. The themes they have outlined thus far will provide useful information as we go forward.”
Whitepaper: A Practical CECL Action Plan
Webinar: CECL Data Quality