Takeaways from the July 30, 2018 Ask the Regulators Webinar – Part 4
The July 30 installment of the “Ask the Regulators” series included speakers from the Board of Governors of the Federal Reserve System (FRB), the Federal Deposit Insurance Corporation (FDIC), the Office of the Comptroller of the Currency (OCC), the U.S. Securities and Exchange Commission (SEC), the Conference of State Bank Supervisors (CSBS), and the Financial Accounting Standards Board (FASB). The goals of the session were to:
- Share the agencies’ perspectives on 27 questions submitted by community institutions
- Demonstrate a consistency of views across agencies
- Share common challenges faced by community institutions about implementing CECL
Question 13. Segmentation and life of loan for CECL.
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?
As with many things related to CECL, how an institution pools its loans depends on its unique circumstances. In the case of pooling, the determinants are specific to your portfolio and the CECL methodology or methodologies you choose to apply to your pools. There is no requirement in the standard for all loans in a pool to have the same term. The requirement is that loans in a pool share similar risk characteristics.
The regulators noted that a weighted average remaining maturity (WARM) is an acceptable way to measure the average life of a loan for a CECL “life of loan” calculation. WARM measures the remaining term of a loan from when you run your CECL estimate calculation to its contractual maturity. WARM differs from a strict WAM, which measures a loan from the date of its origination to its date of maturity.
The overall goal of “life of loan” estimating is to capture losses over the life of the loan considering prepayments.
Question 14. Life of loan.
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?
The standard requires expected credit losses to be measured over the contractual term of the financial assets in a pool. Prepayments should be included, but not extensions, renewals, or modifications; when there is a modification of a loan, it is treated as a new loan and its new origination date is its renewal or modification date. A loan also becomes a new loan when the terms are changed, unless those changes were written into the original loan terms. The requirements do not apply to TDRs, which can be expected to extend beyond their original terms. Typically TDRs will be estimated separately.
To estimate prepayments, an institution should look to its historical data to determine the extent to which prepayments reduce the contractual terms. For example, if 50 of the 100 loans in a group of five-year loans are prepaid by year three, then the prepayment factor for that group is 50 percent by year three.
Methodologies other than a discounted cash flow (DCF) can be used to estimate loans that include prepayments. In a DCF model, you need to insert prepayment expectations; in other models, it’s inherent in the calculation.
Question 15. Life of loan.
How is life of loan determined for lines of credit with a one-year maturity?
The Allowance for Credit Losses (ACL) should be based on 12 months, unless a TDR is reasonably expected. Lines of credit with portions undrawn will require an expected loss for both drawn and undrawn portions. The unused portion will require an expected loss number that results in an “other liability” on the balance sheet and not recorded as a contra to the loan balance, since there is no loan for the undrawn portion.
Question 17. Credit cards.
How are historical losses on open-ended credits (e.g., credit cards) determined under CECL?
As with many things under the principles-based standard, it depends. There was potentially two approaches discussed by the transition resource group (TRG) during the June 2017 meeting. View A is all payments from the borrower expect to be collected, View B is that only some of the payments are expected to be paid by the borrower. The staff believes either view to be appropriate and it will rely on management judgement for the estimate.
One of the ways to approach tracking credit card losses is to break them down as loans in groups of “transactors,” people who pay the full the balance at the end of every billing cycle, and “revolvers,” borrowers who carry balances longer, and therefore are more likely to have higher balances and for whom there will be credit losses.
Transactors won’t have anything on books so they most likely will not need sub-segmentation for CECL. Revolvers typically have more losses and, depending on the portfolio, will most likely need to be pooled. Some examples are as follows:
- By historical payment rate
- By utilization rates
- By credit score
- By delinquency status
Again, the approach your institution takes depends on the nature of this segment of your portfolio. Your approach will depend in great part of the volume of your credit card business.
Question 19. Methods.
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?
The standard does not require adoption of a specific methodology for calculating the allowance under CECL. As well, you can employ different methods for different pools, though you can’t use different methods on the same pool and just “cherry-pick.” Justifying a separate method for a particular pool is usually based on portfolio size. For example, institutions with more than one large portfolio type, say a large commercial portfolio and a large consumer portfolio, might find it useful to use a different methodology for each. However, a cost-benefit analysis usually rules against building a different methodology for an immaterial pool.
Question 23. Individual impairment.
What types of loans are required to be evaluated individually under CECL? Does CECL eliminate the need to identify and measure impaired loans?
The concept of impairment goes away under CECL. However, there may be instances where a loan within a pool exhibits different risk characteristics from the rest of the loans in the pool. In this instance, the institution should determine whether that loan should be removed from the current pool and pooled with other similar loans to be analyzed individually. There could be instances where a reserve is determined for an individual loan because it exhibits unique risk characteristics.
Read part 1 of the series covering supervisory expectations, third-party vendors, historical data, segmentation, and the non-PBE effective date.
Read part 2 of the series covering peer data, low historical loss data and the ACL, methodologies for CECL, and reasonable and supportable forecast periods.
Read part 3 of the series covering supervisory expectations, qualitative factors, more about reasonable and supportable forecasts and TDRs.