Tom Van Lenten, CPA, CFO Consulting Partners LLC
In mid-2016 the FASB issued its long-awaited standard on accounting for credit losses (ASU 2016-13). The new standard adds to US GAAP an impairment model known as the current expected credit loss model or CECL. In a shift from current practice the CECL model is based on expected losses rather than incurred losses. Under the new CECL model an entity recognizes its forward looking estimate of the expected credit losses in the Allowance for Loan and Lease Losses, or ALL when an asset is booked.
The FASB believes that the new standard will result in more timely recognition of credit losses. There is a long transition timeline and implementation is required at January 1, 2020 for calendar-year reporters that are SEC filers, and January 1. 2021 for other entities.
The CECL approach is to be applied to most debt instruments that are not measured at fair value. This includes loan commitments, financial guarantees, lease receivables, trade receivables and held-to-maturity securities. Available for sale debt securities are specifically excluded and will continue to be evaluated for impairment under the AFS securities impairment model.
The Standard does not require use of any one method to compute expected credit losses. Management is expected to maintain complete data sets, documentation, and analysis to support how the selected method fits with the standard and produces realistic estimates of future losses. Examples of some of the acceptable methodologies are:
- Loss rate: Historical loss rates can still be used as a starting point for determining expected credit losses. When using historical loss rates, the Bank would have to consider how conditions that existed during the historical charge-off period differ from current projections and adjust the future loss rates accordingly.
- Discounted cash flow: Projections of principal and interest flows over the life of the asset.
- Probability of default method estimates the probability loans in certain risk-stratified segments will default. This approach computes the percentage of loans that have defaulted in a pool over a look-back period.
- Roll-rate method. The roll-rate method is often referred to as “migration analysis”. It is based on determining a prediction of credit losses based on segmentation (by delinquency or risk rating, for example) of a portfolio of financial assets. No standard roll-rate model is used throughout the financial institutions industry, but most of the models used are based on similar underlying principles.
- Aging analysis: Banks use aging analysis in determining the effectiveness of their credit and collections functions and for estimating potential loan losses. The information from this analysis is used by Banks as a component of their estimate the allowance for loan & lease losses.
- Collateral-value: For loans that are collateral dependent, evaluation of collateral and costs of disposal can be used to estimate expected credit losses.
Institutions with multiple loan types will likely need several methodologies to properly analyze the loss content of each portfolio segment and then aggregate the results to understand the expected losses for entire loan portfolio.
Expected credit impairment will be recognized as an allowance, instead of as a direct write-down of the amortized cost of the financial asset. Changes to the ALLL are recorded immediately through credit loss expense. The carrying amount of a financial asset that is considered to be uncollectible will be written-off in a manner similar to current US GAAP.
While there is significant time until required implementation, it is imperative that management begin the implementation process as soon as possible. The most problematic aspects of CECL implementation will likely be: a) developing the support systems and processes, b) gathering and organizing the historical data to support new methodologies, and c) developing going forward data capture and analysis processes.
- With respect to a) above, existing modeling approaches currently used for Basel, and stress testing may be leveraged and adapted for CECL. However, new credit loss models will likely need to be implemented. It is anticipated that the development or purchase of credit models that estimate future credit losses may be the most time consuming phase of CECL implementation.
- For b) the historical data requirements for the new models will often not be resident in current financial and credit systems. This data will have to be developed from the bank’s core operating system and organized into databases which the models can access.
- For data capture and analysis c) the starting point for developing a future loss credit loss data base is current loan and credit loss data. Companies implementing CECL should undertake a robust review and analysis of its current loan and credit data so that the process of developing a future credit loss data base starts with historic data that has a high level of accuracy and integrity.
Many of the disclosures required under the new standard are similar to those required under current GAAP. For example disclosures on credit quality, ALLL roll-forward, policies tor determining credit losses, past due and non-accrual status, and collateral dependent assets are the same. There are, however, significant additional disclosures of credit quality indicators, disaggregated by year of origination for a five-year period.