10 Crucial Truths about CECL
There’s plenty of uncertainty for financial institutions in the transition to the new Current Expected Credit Loss (CECL) accounting standard. In May 2017, at the annual National ALLL Conference, professionals Mike Lundberg, auditor and partner with RSM US, and Graham Dyer, auditor and partner with Grant Thornton, aimed to help clear up some of that uncertainty by separating CECL truth from fiction. A summary of that discussion follows:
Truth 1: CECL Does Not Mean ‘Clearly Excessive Calculation of Losses’
Some people dealing with the CECL transition have mistranslated the acronym to convey their concern that the new standard requires institutions to over-estimate their potential credit losses.
Lundberg set the record straight on what CECL aims to achieve, “The goal really is moving to a life-of-asset loss measurement, so we’re going beyond just the incurred – what’s happened up through our balance sheet as of our reporting date – and for the first time, really looking forward and projecting future credit events, and measuring and capturing future loss events.”
Dyer added that financial professionals need to understand that the CECL standard does not require them to peer into the future and predict specific outcomes — and to be held accountable if their predictions fall short.
People shouldn’t worry that they’re making educated guesses that are bound to be mistaken, rendering the whole thing pointless.
“It’s not actually asking you to get it right,” Dyer said. “It’s asking you to say, ‘What are your expectations today, and let’s measure that.’ “
Truth 2: CECL is Not So Complex that You’ll Have to Outsource if You Don’t Have a Whole Team of Quants
Lundberg noted that regulators have repeatedly said that they expect CECL to be implemented in a scalable way based on the size and complexity of the loan book to which it’s being applied. That means financial institutions should be able to scale their CECL compliance operations according to their available resources. That could require companies to outsource if they lack the talent to build the statistical models, but everything depends on each institution’s individual needs.
The standard acknowledges that each institution is unique.
“We’ve had some clients who are mono-line consumer lenders, like auto lenders, who have looked at the standard and shrugged and said, ‘This doesn’t seem so hard. This is kind of what we do anyway’,” Lundberg said. “There are others who do a lot of commercial lending, especially commercial real estate lending, who look at this and say, ‘I don’t even know how to take the first step.’ It’s very dependent on your institution’s portfolio.”
Truth 3: CECL is Not Something You Just Hand Off to Your Controller
CECL is a significant undertaking across several business departments — accounting, audit, credit, lending, IT and more — so every stakeholder needs to be involved in its implementation.
“You’re going to need economic inputs in the data. You’re going to need a lot of involvement from IT,” Lundberg said. Just one example Lundberg cited: Companies that go through acquisitions often destroy historical data that will have to be preserved to build CECL estimation models.
“It really is a fundamental change in operations that’s going to require a broad, multi-functional, cross-functional team throughout the organization and perhaps beyond.”
Dyer noted that a lot of financial professionals will be exposed to rigorous internal audit controls that they may not have encountered before.
“You may have estimates and forward-looking data that you use, say, in board reporting, but that probably hasn’t come under the scrutiny of your auditors before — and now it very well may,” Dyer said. “The expansion of internal controls beyond just the core finance function into the credit function, into the risks function, and if you’ve got an economics group, into their function, is going to be new for not just a lot of banks and credit unions, but a lot of people in those institutions.”
Truth 4: CECL Allows a Variety of Methodologies to Estimate Losses
There’s no one-size-fits-all model for CECL estimation, Dyer noted.
“One of the key takeaways when you start to think about the standard is you go, ‘I’ve got different portfolios. Which methodology do I use that covers all of them?’ The answer may be you don’t need to use one to cover all of them. You may use different methodologies that are appropriate for different portfolios.”
Institutions will have to figure out the best models appropriate to their circumstances. What works for the institution down the street might be a poor fit for your institution, if you have more real estate or commercial loans, for example.
“I sometimes refer to this as the blessing and the curse,” Lundberg said. “The blessing is that this accounting standard is very, very principles-based and doesn’t require or dictate a set methodology. The curse is the accounting standard doesn’t give you a set methodology. That sword cuts both ways.”
Truth 5: You Won’t Base Forecasts on the Longest Contractual Period in Your Portfolio
The key to the reasonable-and-supportable philosophy of CECL is that forecasts can use multiple loan durations to provide realistic loss estimates based on what you know at the time of the estimate.
“You don’t necessarily have to have a standard forecast period,” Dyer said. “That may be practically what happens, but you may be in situations where you’ve got a fairly stable economic environment, so you feel more confident looking further out in the future, or you may have a very volatile economic environment, where you may have less clarity about the future. For instance, if it’s Monday morning and Lehman Brothers collapsed on Friday afternoon, you may have a very short forecastable future period.”
Lundberg encouraged the audience to think of CECL modeling from an auditor’s perspective to get an idea of the kinds of challenges they will face. Say, for example, an auditor has two clients — one that is forecasting unemployment rates over the next 18 months and the other for 36 months. The forecasts predict improvements over 18 months but a steady decline over 36 months.
“How, as an audit partner, assigning both of those opinions that same day, or that same week, do we reconcile those different approaches?” Lundberg asked.
“In many cases, these different answers are both going to be appropriate — they’ll both be reasonable and supportable. They’re both acceptable to use. It will create some interesting challenges as you diverge from your peer institutions.”
Truth 6: CECL Calculations Are Not the Same as Fair Value
Financial professionals need to understand the distinction between fair value and CECL.
“CECL has some similarities to fair value. It’s forward-looking, like fair value is forward looking, and it considers future economic events or conditions that are expected, just like fair value would, but it’s different than fair value in a number of important ways,” Dyer said.
“One is that CECL has this concept of a mandatory reversion to historical experience. Fair value doesn’t have that sort of notion built into it.”
Perhaps most important, Dyer said, is that CECL has no market-participant standard. “In other words, an institution may have a view, but as an auditor, they say, ‘We don’t think that’s consistent with a market participant view for the following reasons.’ It’s the market participant view that rules the day in a fair-value measurement — but for CECL, there is no objective market participant standard.“
Truth 7: A CECL Reserve Will Generally Be Larger than an Incurred-Loss Reserve
Why should CECL reserves be bigger than their incurred-loss counterparts?
“There are a couple of conceptual differences here,” Lundberg said. “One is the duration. In an incurred loss model, we are cutting off loss events as of a balance sheet date, whereas under CECL, we are forward looking. We have a longer period of time to consider loss events in our reserve.
“The other key theoretical change is removing the threshold. Currently under the FASB 5 model, we have the probable, possible, and remote categories, and a loss has to be probable to be recorded, so our reserve under a current methodology should reflect probable losses. Under CECL, that probable threshold gets pooled, so all the losses, even remote losses, are included.”
Truth 8: You Can Properly run a CECL-Compliant Methodology Even if Your Portfolio Has a Very Low Loss History
Having a long history of no losses does not mean you have no loss history, Dyer said. “I don’t think just because you have a history of low losses that means you just throw your hands up and say, ‘Ah, forget it. Give me a number and I’ll book it.” Institutions may have to consider supplementing internal data with external data.
The key is for institutions to find a model that suits their unique circumstances. “There may be methodologies that are better suited for portfolios with infrequent losses,” Dyer said.
Truth 9: There’s No Definitive List of all the Data Elements You Need to Capture
“There’s no way we or anyone else can tell you exactly what you’ll need,” Lundberg said. “It’s so dependent upon the methodology that you choose, how you decide to model, and what data you have available.”
“We sort of have a cart-horse problem here,” Lundberg added “Do you pick the ideal methodology and then try to find or recreate the data you need, or do you figure out what data you have and find a methodology that will work with that data?”
The abundance of data doesn’t have to traumatize those who are just delving into CECL.
“People sort of get deer-in-the-headlights when they start thinking about data” because they aren’t sure where to begin, Dyer said. To find a starting point, ask basic questions like “what are the drivers of credit risk in your portfolio?”
“Even if you don’t have the math behind that, everyone has an intuitive sense. I think those are the places to start investigating,” Dyer said.
Lundberg added: “If you don’t know where to start, pick a methodology and try to use it. See if you have the data to make it run. If you don’t, then you’re going to need to figure out where to go from there. That might mean a different methodology. It might mean going into the archives to try to recreate some data.”
“What I’m recommending to our clients is just to get started,” Lundberg said.
(Some commonly tracked data for CECL models. Click here.)
Truth 10: CECL Requires Pooling of Assets with Similar Risks
“CECL has a requirement that loans share common risk characteristics,” Dyer said. “Then you must pool those loans and predict on that pool basis.”
“Loans can move,” Dyer added. “The only time you would evaluate a loan individually is if it does not share common risk characteristics with any other loans.”
The regulators, auditors and other experts have not figured out everything about asset pooling, particularly in cases of extremely small pools of assets, Dyer said. That’s just one more of the uncertainties that will have to be worked out in the transition to CECL.
For more information about setting up for CECL compliance, read this whitepaper.
About the Presenters
Partner, National Director of Financial Institution Services | RSM US LLP
Mike is the National Director of Financial Institution Services for RSM US LLP (formerly McGladrey LLP). In this role, Mike has responsibility for audit, accounting and risk containment matters across the firm’s financial institution practice, which includes community banks, credit unions, finance and leasing companies, and other specialty lenders.
Prior to joining RSM’s National Professional Standards Group, Mike served as a financial institutions specialist and assurance partner in the firm’s Des Moines office. He has worked with a variety of financial institutions, with a primary focus on large community banks. In addition, Mike has worked with a variety of not-for-profit organizations, primarily private colleges.
Mike holds a Bachelors in Accounting from the University of Northern Iowa.
Partner, Accounting Principles Group | Grant Thornton
Graham currently consults with financial institutions clients of all sizes and audit teams regarding technical accounting and auditing matters, with a focus on regulatory capital and compliance impact.
His background includes the Accounting Principles Group at Grant Thornton and serving as a Professional Accounting Fellow in the Office of the Chief Accountant at the OCC. Those experiences provide him with a unique perspective on complex accounting transactions, accounting policies, and internal control matters concerning financial institutions. Graham also serves on the FASB’s CECL Transition Resource Group (TRG), and previously served in the same capacity on the IASB’s expected credit loss standard, IFRS 9.
Some specific areas of expertise include accounting for the allowance for loan losses, accounting for business combinations, accounting for purchased credit-impaired debt, fair value, modeling impact on regulatory capital, and accounting for mortgage banking and servicing activities.
Graham holds an MPA in Accounting from The University of Texas at Austin and is a CPA.