The Good, The Bad, And The Complicated: How Atomic Credit Union Learned To Live With CECL

Jerod Wiley has seen a thing or two during his 20 years as CFO at Atomic Credit Union ($576.4M, Piketon, OH), but perhaps nothing quite like CECL.

After several years of review and revision, the new current expected credit loss (CECL) standard from the Financial Standards Accounting Board is now in place for credit unions with more than $10 million in assets. The rule changes how financial cooperatives project and reserve for lending losses, with an eye toward securing their ledgers against convulsive losses such as those experienced by many financial institutions during the Great Recession.

There are some concerns that complying with CECL will require credit unions to boost their allowance for loan and lease losses (ALLL) to the point it affects member service, especially the ability to lend.

Wiley, whose 15-branch cooperative serves a wide swath of rural southern Ohio, doesn’t see that happening at Atomic. Despite the challenges of implementation, he actually sees process and strategic improvements coming out of CECL.

Here he explains.

What have been the biggest process challenges when it comes to implementing CECL? How have you managed them?

Jerod Wiley: Historical data collection and the quality of current data. Like most financial institutions, we encountered issues with the amount of data necessary to effectively calculate a true reserve amount that adheres to the FASB standard. You need that to implement a viable model that accurately projects future losses in a loan portfolio. Plus, we converted our core processor in April 2021 and had to combine two sets of historical data into one format so our model could effectively process it.

We want at least five years of data for our historical models to give us an accurate representation to effectively calculate reserve amounts assigned to our loan pools. There are several methods to do that, including vintage or weighted-average remaining maturity (WARM), which uses historical components, and others that use current data to model probability of default or discounted cash flow methods to project future losses.

We’ve found the challenge with these two methods is getting your data quality above a certain threshold. Atomic has internally designated data quality to not fall below 80% while using probability of default or discounted cash flow methods. However, management is collectively working to get each loan pool’s data quality to be at the 95% mark, which will take time to achieve.

“I don’t believe FASB will significantly alter its rule on reserving for future losses, but it will be interesting to see how the dynamic among the examiners, CPA practitioners, and credit unions change during that time span.

What CECL challenges have you turned into opportunities?

JW: Evaluating all data associated with implementing an effective CECL model requires significant time and effort across numerous credit union departments. That allowed the credit union to evaluate what data was associated with each loan type and whether any revisions needed to occur at a branch level concerning underwriting standards.

For example, during our review of data quality associated with the probability of default method, we realized inconsistencies in certain loan pools, such as commercial real estate. This was a product of merging two sets of data from two different cores.

This data issue gave us an opportunity to find and correct incomplete data as well as the opportunity to evaluate the interview process at the branch level and modify what information was required moving forward. It also gave us the opportunity to drill down into the underwriting process and create some efficiencies within our loan origination workflow.

Many credit unions are concerned increasing loan loss reserves might affect their ability to lend. How much have you raised your reserves? Has it impacted your lending practices?

JW: Atomic experienced a slight increase once CECL was adopted. The reserve amount at year-end 2022 was $2.3 million. That was 0.51% of the loan portfolio. We adopted CECL on Jan. 1, 2023, and closed that month’s books with a reserve amount of $2.35 million, or 0.53% of the outstanding loan portfolio.

That said, CECL has not altered our lending strategies or underwriting standards. Management believes our underwriting and collection efforts are strategically aligned to give our loan personnel the ability to intelligently lend to our membership.

Atomic has experienced year-over-year loan growth of 16.30% and 21.40% in 2021 and 2022, respectively, while maintaining delinquency ratios of 0.11% and 0.10% for the same years. We also maintained a charge-off ratio below peer average for that same time. Atomic’s charge-off ratio was 0.15% and 0.20% for year-end 2021 and 2022, respectively. This is compared to a peer average of 0.24% as of Dec. 31, 2022.

Want more inspiration for CECL best practices? CreditUnions.com has got it. Read “ After 7 Years CECL Is Here. Quest FCU Was Here”  and “Look Beyond CECL Model Selection.”

What are some other benefits that have emerged from CECL implementation?

JW: One benefit is our new team-oriented approach in evaluating how the organization goes about calculating CECL. Atomic’s risk committee now gets more involved with the credit union’s delinquency and attempts to project loss factors based on trends occurring within the portfolio. It allows numerous perspectives on how a certain loan portfolio is progressing and how the credit union can pivot strategies based on management’s risk appetite.

It’s also a much more efficient process. The majority of data is uploaded into our software model at month-end and then calculated automatically. We’re eliminating a lot of data entry.

What’s another major upside of CECL implementation?

JW: We increased our used vehicle portfolio — direct and indirect — by 27.0% and 23.3% in 2021 and 2022, respectively. Our new CECL model allows management to view loan-to-value migration across several time periods.

During the pandemic, Atomic experienced increased automobile collateral values while underwriting new loans. Used car values spiked as demand remained consistent with previous years and car inventory levels decreased due to supply chain issues with new vehicles, thus pushing up demand for used inventory.

Today, collateral values are decreasing due to inventory levels beginning to come back to pre-pandemic levels. This is one of the examples of how our risk committee might approach evaluating potential future losses based on trending data. It simply brings more data and opinions into the conversation, which I believe vastly improves the process.

How did you train and educate your staff to effectively implement and comply with CECL?

JW: We had to move quickly to meet certain milestones to hit our target implementation date. That limited our implementation process to a few key employees that had the requisite knowledge of the data and concepts that were involved to do this effectively.

Our current strategy consists of training the board of directors and certain staff on the concepts of CECL in its current state. We’ll further train some employees to analyze the data associated with CECL and how it affects our reserve requirements. The time frame for the strategy will be the next six to nine months.

How has CECL impacted Atomic’s financial reporting, including income statements and balance sheets?

JW: That impact also has been minimal. We adjusted our allowance for loan loss with an entry of slightly less than $13,000. We also incurred some additional expenses by partnering with a third-party software platform to help analyze loan data and perform the CECL calculation. We also will incur an annual expense of a CECL validation by our CPA firm. This is not a current requirement, but Atomic wants to stay proactive regarding CECL and stay compliant with GAAP standards.

Will compliance continue to be a moving target, part of a learning curve for everyone involved?

JW: I believe there will be a learning curve for all parties involved. We’ve not faced an examination from the federal or state regulatory agencies at this point, so it’s hard to say what their approach will be toward CECL. I believe it will take 12 to 24 months for examiners, CPA practitioners, and credit unions to develop a solid, working arrangement where all involved parties are on the same page and have a firm understanding of the dynamics surrounding CECL moving forward.

It will be interesting to see how the examiners approach the new method. For example, will the examiners come up with one methodology to determine a baseline calculation and then reconcile their approach to the credit union’s internal methodology? Will the examiners defer to the CPA’s determination to ensure the credit union’s calculation adheres to GAAP?

The next year or two will be interesting in the credit union space. I don’t believe FASB will significantly alter its rule on reserving for future losses, but it will be interesting to see how the dynamic among the examiners, CPA practitioners, and credit unions changes during that time span.

Is CECL On Your Mind?

Learn how to better project and reserve for lending losses at your institution. Callahan’s Peer Benchmarking Suite makes it easy for credit union leaders in any role to measure performance and support strategic plans.
Request A Demo Today