This article focuses on accounting for Allowance for Credit Losses (ACL) under the Current Expected Credit Losses (CECL) Standard (ASC Topic 326), including the recent FASB developments as well as considerations for implementing the standard.
Most organizations have commenced with their implementation initiatives for the Financial Accounting Standards Board (FASB) Accounting Standard Update (ASU) 2016-13, Financial Instruments – Credit Losses
, which changes the way organizations account for credit losses. Currently, an incurred loss approach is used under ASC 450 which requires lenders to reserve for losses that are probable and reasonably estimable, whereas CECL requires a forward-looking approach that encompasses the entire remaining lifetime of every asset that could experience a credit loss. With an effective date as early as Jan. 1, 2020 for large public companies, the looming deadline has proven to be more challenging than many initially believed. Based on feedback provided to the FASB, preparers identified numerous challenges in implementing the requirements of the ASU, with the vast majority of them citing aggregating and properly analyzing data as primary concerns.
Key Provision of Proposed ASU
On Aug. 15, 2019, the FASB issued a proposed ASU that defers the effective dates of the CECL ASU for SEC filers that are smaller reporting companies (SRCs), non-SEC filers, and all other companies (see figure below).
The dates and provisions described above are subject to change pending the FASB’s finalization of the proposed ASU. A company’s determination about whether it is a SRC would be based on its most recent filing status prior to the date the proposed ASU becomes final.
Although a potential deferral may relieve some of the immediate time pressure, it should not be viewed as an opportunity to postpone the work as lenders will need to undertake robust implementation efforts to comply with the CECL ASU. Rather, smaller SEC and private companies can view the deferral as an opportunity to learn from the knowledge and experience gained from implementation issues encountered by larger public companies, as well as closely monitor interpretations and implementation of practices. A brief overview of the CECL ASU follows.
The CECL ASU applies to lenders with financial assets, leases, and loans not reported at fair-value (see figure below).
The CECL ASU does not prescribe specific methodologies, but lenders are expected to measure allowances based on historical loss experience adjusted for differences in loan attributes. Credit losses should be measured first on a collective basis for financial assets with shared risk characteristics, and then on an individual asset basis for assets without shared risk characteristics. The estimates of expected credit losses should consider all information relevant to assessing collectability of cash flows including internal and external information relating to past events, as well as current economic conditions along with reasonable and supportable forecasts. Other considerations include:
- The treatment of expected prepayments as well as open loan commitments
- Only contractual term is considered, unless a troubled debt restructuring is reasonably expected
- Beyond a reasonable and supportable forecast period (e.g., 5 years), revert to historical experience
Data and Modeling
Lenders that have started the implementation process point to the need for significant data mining and the retention of historical credit quality. Data is the real driver; the more that is available, the better the options are for selecting the various models available under CECL. Additionally, more robust data improves the quality of the output. Data governance is also critical for ensuring quality, reliability and consistency. There are a number of models available for implementing CECL, including:
- Loss-Rate – a simpler model based on a historical rate of loss and utilized by lenders with less complex portfolios (e.g., community financial institutions)
- Vintage Analysis – essentially a variant of the loss-rate model, where loans are grouped by similar risk profiles and origination period and utilized by lenders with smaller ticket homogeneous pools
- Discounted Cash Flow (DCF) – a scenario-based loan-level model that is utilized by lenders with larger ticket shorter duration loans
- Probability of Default (PD) / Loss Given Default (LGD) – a scenario-based loan-level model where two estimates are made for each pool of loans: the PD—the likelihood that a default event will occur to an asset/pool—and the LGD—the magnitude of a loss given a default
While there are benefits, challenges and limitations with each of the models, the key is the documentation underlying the model inputs, calculations, forecast assumptions, scenarios, etc., including the organization’s reasoning for why the chosen model best represents the risk of its portfolio.
For vendor finance lenders with more significant and complex portfolios that have the necessary data and analytical power to execute, the PD/LGD model is the preferred model due to the utilization of loan-level data, along with the PD and LGD estimates which allow for more granular analysis. The formula for the PD/LGD model can be expressed as Estimated Credit Loss (ECL) = PD x LGD x EAD
(Exposure at Default).
Key considerations for implementation include: having a clear plan, involving the correct personnel, making timely and well-considered decisions regarding leveraging existing systems or acquiring new ones, and documenting company efforts along the way.
Lenders will also need to undertake extensive efforts to validate their CECL models, including:
- Properly challenging the model, including back-testing following the model risk governance framework
- Validating each component within the model (e.g., PDs, LGDs, prepayments, etc.), including third party models
A recent Moss Adams CECL survey of financial institutions indicates that lenders believe the most difficult part of their implementation will be developing the loan loss models and related loan loss methodology, followed closely by the challenges in developing reasonable and supportable forecasts.
The impact of the CECL ASU will be recognized on day one as a retroactive adjustment to capital, based on the anticipated effective dates noted above, with quarterly adjustments to the forecast thereafter for every open exposure. Key considerations for implementation include: having a clear plan, involving the correct personnel, making timely and well-considered decisions regarding leveraging existing systems or acquiring new ones, and documenting company efforts along the way. The recent survey noted above determined that roughly 57% of lenders expect their existing allowance for credit losses to increase between 0–50% upon implementation, with another 13% expecting increases of over 50%. Based on those survey results, it appears that the impact to reserves will be significant for most lenders and a timely, and well executed implementation effort that provides for successfully managing the impact is well merited. In light of the recent development, smaller SEC and private company preparers should continue to work on their CECL implementation efforts and engage in discussions with their external auditors to obtain insight into the large public company implementation interpretations and best practices.