CECL Q&A Quick Reference Guide


Ask the Regulators: CECL Q&A Session (July 30, 2018)

Banking professionals across the country often ask us the same questions about how CECL will be implemented and our approach to helping uncomplicated banks and credit unions comply with this new accounting standard.

On July 30, 2018 all the relevant banking agencies presented one of the most informative and straight forward reviews that include answers to these questions we hear from bankers every day. If you had to choose one webinar to listen to on CECL we would recommend this one. We found that the answers to the questions further validated the approach we chose to take with our CECL calculator, and we hope you might choose us as your preferred CECL solution provider.

You can find a link to the recorded webinar and the PDF slide hand-out immediately on the right of this page. Also, we wanted to provide a "CliffsNotes" summary of the answers to the questions (as we heard them) as well as the minute mark for where each question is discussed during the 90 minute webinar.

1 Small and Less Complex

Q: Is there a definition of "small and less complex" or a set of factors to consider in determining whether an institution fits that description? Would most institutions under the FDIC's $10 billion threshold for "large and highly complex" qualify?

A: Minutes 4-8 The main take away would be that CECL is scalable and so are the implementation methods (from spreadsheets to econometric models). The bank should continue the "type" of methodology (loss rate analysis) that they have currently been doing so long as they change their inputs and assumptions to comply with CECL. They likely won’t need to leap frog to a different type of methodology.

2 Supervisory expectations

A: What can community institutions expect during 2018 examinations relating to CECL? Do examiners have a standard set of expectations for community institutions?

Q: Minutes 6-9 Regulators are evaluating for CECL ‘readiness’ not CECL compliance at this time. The most important expectations are to determine when implementation will take place (is your bank a PBE?), and evaluate the financial impact CECL will have on your balance sheet.

3 Supervisory expectations

Q: How will the agencies evaluate the institution’s process to determine allowance for credit losses (ACL) under CECL? Will examiners challenge institutions if their method results in a lower ACL under CECL than under the incurred loss model?

A: Minutes 9-11 If the overall allowance under CECL goes down the institution should be prepared to offer support. The examiner would challenge the bank if the allowance was unreasonably low if the assumptions of the calculation are adequately supported. The agencies will not establish benchmarks for allowances. The goal for institutions to maintain allowances that are appropriate for their specific portfolios under CECL.

4 Third-party vendors

Q: Do the agencies have a specific expectation regarding the use or purchase of third party vendor services to implement CECL?

A: Minutes 11:15-13:00 The regulators are okay with banks using a vendor OR not using a vendor. If an institution uses a third party you are still responsible for understanding the inputs, assumptions, and methodologies that go into your CECL allowance. This cannot be outsourced since that responsibility belongs to management. 

5 Charge-offs and recoveries

Q: When determining historical loss rates to use in the calculation for ACL, how should recoveries be considered in the charge-offs (i.e., net or gross of recoveries)?

A: Minutes 13:30-18:00 Yes recoveries should be considered when estimating future losses. This is a complicated and lengthy answer so if you have specific interest in this subject it would be worth listening. In the Curinos BankTrends model we consider historical losses net of recoveries, and use those loss rates to estimate future losses. Consequently our future projections consider recoveries since we are forecasting net charge offs as our allowance calculation.

6 Peer data

Q: Is it acceptable to use data from various regulatory reports (e.g., FFIEC Call Reports)? What types of peer data are available as a reference for historical loss experience?

A: Minutes 18:00-20:00 It is acceptable to use external data such as peer data in the Call Report or internal data from third party providers. Provided that the data is a relevant peer group it can be beneficial when an institution doesn’t have sufficient historical loss data to estimate credit losses for a group of loans with similar risk characteristics. 

7 Historical data

Q: What is the minimum number of years of historical data required to calculate the ACL?

A: Minutes 20:00-20:30 Institutions will need historical loss data to cover the life of the loan, or use external loss data, or utilize qualitative factors if sufficient historical data is not available.

8 Low historical loss experience

Q: What guidance is available for institutions with zero to extremely low historical loss experience? To what extent may institutions rely on qualitative adjustments to determine the appropriateness of the ACL?

A: Minutes 20:30-24:30 Judgment is really important here. Allowances must be supported based on qualitative factors and economic factor changes between current and future conditions. Consider external factors when establishing qualitative factors such as peer data if internal historical loss data is not available.

9 Qualitative factors

Q: What qualitative factors would be considered reasonable when using a loss rate method to calculate the ACL?

A: Minutes 24:30-28:30 Previously the agencies outlined qualitative areas that institutions may consider under the incurred loss methodology. Examiners will not require institutions make a qualitative adjustment for every one of these factors. Under the current methodology the qualitative factor analysis examines factors up to the reporting date, but under CECL future conditions need to be considered. In the Curinos BankTrends solution we provide an economic forecast tool to help evaluate future economic changes and their impact on loss rates.

10 Reasonable and supportable forecast period

Q: Is there a minimum preferred range for the reasonable and supportable forecast period? How can institutions estimate losses if the reasonable and supportable forecast does not cover the entire contractual life of the loan?

A: Minutes 28:30-30:00 Institutions MUST create a forecast. Not every institution will be able to create a forecast period that covers the full life of their financial assets, however, it would be inappropriate for an institution to say they are unable to create a forecast of any length. Each bank should periodically review its forecast period and make appropriate changes.

11 Reasonable and supportable forecasts

Q: What are the agencies’ expectations regarding the use of economic forecasts? Do the agencies expect institutions to use multiple scenarios when developing reasonable and supportable forecasts?

A: Minutes 30:15-31:00 Agencies will not ‘require’ multiple scenarios. This is an area that agencies are working to provide additional guidance- look for that shortly particularly from the AICPA.

12 Segmentation

Q: What is the appropriate level of loan pool segmentation? How granular should it be? Would it be acceptable for a community bank to pool loan segments based on FFIEC Call Report categories?

A: Minutes 31:00-32:45Group your loans with sufficient granularity so that it breaks out different risks but not so granular that the pool is so small that you can’t adequately predict any trends or potential risks. If your pool of loans is too small it won’t be statistically significant. For small banks you might need to segment at a higher level and the Call Report codes might be sufficient. However for larger banks you might need to have more segmentation beyond the Call Report codes.

13 Segmentation and life of loan

Q: Is it appropriate to pool loans with different maturities into one segment? For example, can a seven-year term commercial real estate (CRE) loan be pooled with a five-year CRE loan if the loan risk characteristics are similar? If yes, how is the average life of loan calculated for such pool?

A: Minutes 32:45-37:00 It depends, but there is no requirement in the standard that all loans within the pool have the same term. WARM is an appropriate pool. This is a very technical answer so consider listening to it further for more detail.

14 Life of loan

Q: What factors should be considered when determining an average life for a pool of loans? How are prepayments considered in calculating the average life for a pool of loans?

A: Minutes 37:15-39:30 Weighted Average Remaining Maturity (WARM) is different than Weighted Average Life (WAL). Prepayment studies using the individual institution’s portfolio are ideal, but peer data studies can be acceptable too.

15 Life of loan

Q: How is life of loan determined for lines of credit with a one year maturity?

A: Minutes 39:30-40:30 Use twelve months as the contractual life.

16 Renewals

Q: How should renewed loans be considered in the calculations of ACL? For example, if a loan pool has an average life of five years and a loan is renewed at the end of five years, does the renewed loan start a new five year period or should it remain in the original five year pool?

A: Minutes 40:30-41:45 Do not extend the contractual term unless you have a reasonable expectation of entering a TDR at the reporting date. Nothing under CECL changes when a loan is considered a ‘new’ or renewed loan.

17 Credit cards

Q: How are historical losses on open-ended credits (e.g., credit cards) determined under CECL?

A: Minutes 41:45- 48:00 This is a technical answer and worth listening too since they reference many previous TRG meetings. In short credit card borrows can be basically broken into two groups- borrowers who pay off their balance every month- ‘transactors’, and borrowers who maintain a balance that they add to and pay off over time- ‘revolvers’. Large credit card portfolios ought to be segmented into these two groups since they have different risk characteristics.

18 Methods

Q: Some have suggested that the vintage method will be the minimum standard required to implement CECL (i.e., other types of loss rate methods will not be acceptable). Is this accurate?

A: Minutes 48:00-50:45  No, the vintage method is only one possible method of estimating credit losses at institutions. The agencies are not mandating any specific method for estimating credit losses. The FASB had often demonstrated the vintage method simply because it lends it self well to showing people the math in a relatively simple way. In fact the vintage method likely wouldn’t be appropriate for smaller portfolios with limited losses.

19 Methods

Q: Is it acceptable to use different loss rate methods for different pools of loans? Can institutions select a method after seeing the results of using several methods? How often can institutions change methods used to estimate the ACL?

A: Minutes 50:45-52:30  Yes, it is acceptable to use different methods for different portfolios- say residential vs. commercial. However it is not appropriate to run various methods against the same pool and then "cherry pick" the method that results in the lowest loss rates. Changing between methods should be something that is "well documented and infrequent."

20 Transitioning methods over time

Q: Is it appropriate to use one loss rate method (e.g., open pool or WARM) and then transition to another loss rate method (e.g., vintage) at a later time when the institution has collected a sufficient amount of data?

A: Minutes 52:30-54:45  Yes it is okay to start implementing CECL using one method and then transition to another method later once you have more data. It would not be acceptable to change your method just because it gives you a more favorable result. Any changes should be well documented as to why the change is being made.

21 Supervisory expectations

Q: Will agencies object to institutions’ use of the Weighted Average Remaining Maturity Methods, commonly referred to as WARM?

A: Minutes 54:45-56:00 No, agencies will not object. The WARM method is okay to use. It may be only appropriate to use for smaller less complex institutions and for certain types of loan portfolios. Check with your public accounting professionals or examiners with any specific questions.

22 Public company financial statement disclosure

Q: Is there a requirement to disclose the quantitative impact of the adoption of CECL on capital in the financial statements (e.g., Form 10-K)?

A: Minutes 56:30-59:00 Assuming the impact on capital is significant and reasonably estimatable, then yes, it should be disclosed. At a minimum disclosures should include an explanation of the current accounting principles being used and a comparison to the new principles which are going to be adopted.

23 Individual impairment

Q: What type of loans are required to be evaluated individually under CECL? Does CECL eliminate the need to identify and measure impaired loans?

A: Minutes 59:00-1:01:50 The notion of an ‘impaired loan’ has been removed. However, there may be instances where a loan within the pool demonstrates characteristics that are different than the risk characteristics of the other loans in the pool; in which case it would be appropriate to evaluate this loan separately or along with other loans with similar risk characteristics. Also a discounted cash flow is NOT required unless it is the only way to capture the effect of a concession. 

24 Troubled debt restructuring (TDR)

Q: Does determination and measurement of expected losses on TDRs remain the same under CECL?

A: Minutes 1:01:50-1:03:00 In terms of identification of a TDR nothing has changed under CECL. However under CECL you can evaluate TDRS individually or within a pool. If a concession has been granted- specifically an interest rate concession- then you would need to use a discounted cash flow in order to estimate the impact of that concession.

25 Risk ratings

Q: How do loan risk ratings impact the calculation of the ACL under CECL?

A: Minutes 1:03:00-1:05:10 Risk ratings are an example of a risk characteristic that CAN be used in order to segment loans so it can be one way of segmenting loans but not the only way. However, it can be important to evaluate the weighted average risk ratings within pools and how that has changed over time and how the historically observed portfolio might be different than the current portfolio of loans.

26 Revenue recognition 

Q: How does CECL affect the revenue recognition and expense matching principle?

A: Minutes 1:05:10-1:08:45 This is a lengthy and technical explanation and should be listened to. CECL doesn’t change the economics of lending. CECL simply changes the timing of when the loss is recognized.

27 Non-public business entity effective date

Q: Will there be a change in the effective date for non-public business entities?

A: Minutes 1:08:45-1:08:45 Yes, the initial guidance required PBEs and non-PBEs negated a benefit that was intended to be provided to non-PBEs. Section 326.10.65.1 has been amended so that non-PBEs must adopted for fiscal years beginning after 12/15/2021 (listen to minute 1:10:15). The practical take away would be that a non-PBE would need to adjust retained earnings on 1/1/2022 and in order to do that the institution ought to have vetted and implemented a solution much earlier in order to compute the impact as of 1/1/2021. The consistent message from the regulators was "do not slow down" as you prepare for CECL and compute the gap between your current ALLL and your ALLL under CECL. Examiners want to see a "good faith effort" toward answering the questions about CECL preparedness.