Amid widespread concern that Generally Accepted Accounting Principles (GAAP) are inadequate when it comes to advising investors on deteriorating credit quality, the Financial Accounting Standards Board (FASB) has issued a new methodology. The new standard, known as Current Expected Credit Loss, or CECL, uses data analytics to forecast expected losses based on internal and external trends, as well as borrower-specific information. In its simplest form, CECL replaces the old standard of actual or “incurred” loss with a forward-looking estimate of “expected loss” over the foreseeable future. (See our analysis of its anticipated impact.)
The standard was originally scheduled to become effective for public companies in December 2018, but that deadline has been pushed back to December 2020, with private companies to follow a year later.
CECL represents a significant change with far-reaching implications for loss reserves. And yet, just one in ten affected companies has made any significant effort to assess the potential impact and prepare for the change.
Protiviti conducted a webinar recently aimed at internal auditors trying to get the ball rolling at their organizations. As is often the case, the webinar generated more questions than we were able to address during the live session. We want to address some of the additional questions here.
Q: Isn’t the “foreseeable future” loss prediction based on “historical losses” as well? It’s hard to see how CECL offers any real improvement if the underlying data is essentially the same.
A: The forecast into the foreseeable future could be based on historical experiences (losses) and management judgment based on the most updated information.
For the forecasting based on historical losses, data is essential, and that is why CECL implementation will require companies to retain a variety of historical data over a much longer time horizon and analyze it against external information, such as FICO scores, loan-to-value and debt-to-income ratios, and debt service coverage. Internal audit will need to provide assurance on data completeness. With a longer time horizon and more variety of historical data, the CECL model should be able to better estimate the loss under different foreseeable future scenarios. Most companies already have such data saved. Even those who don’t, if they start saving data now, will have four years of historical data to work with by 2020.
For the forecasting based on management judgment, unlike the incurred loss model, the CECL model explicitly requires management to take into account the current information and identify the future scenarios for loss estimation.
Q: With the implementation of CECL, will there also be a corresponding allowance for loan and lease losses (ALLL) requirement on the lending institution?
A: Yes. Regulators published a Joint Statement on CECL on June 17, 2016. Expect more on ALLL in the future, but the June 17 statement is already out there.
Q: Isn’t stress modeling sometimes subjective even when using a third party?
A: Not necessarily. Third-party vendors typically use industry-level data to develop their models, and these models then serve as objective benchmarks against which institutional assets can be evaluated.
Q: What is going to be expected of internal auditors under CECL? Will we be expected to audit the ALLL process and controls over the model, or will we be expected to perform full model validation as well?
A: Both would be expected. Right now, internal auditors should be talking to management to ensure there is transparency into the portfolio and the credit quality evaluation process. There should be clear lines of reporting and communication to the board, and internal audit must remain close to the process throughout to ensure that the model is being applied, and that the model itself is valid as a predictor of credit losses in the foreseeable future.
As we discussed during the webinar, and at the highest level, processes, data sources and accounting will be changing under the CECL guidance. Whenever processes change, internal controls must be reassessed to make sure that no new critical risks have been created and that all critical risk areas have adequate controls in place.
Once in place, the controls must be tested by internal audit. For example, here are some critical concerns:
- Data, process and judgments – Internal audit must collect and test company loss experience and other past events. Some of the processes will require judgment; those judgements must be articulated and supported by evidence. Forecasts on factors that affect collectability, either internal or third-party, must be validated and back-tested.
- Other models – For some institutions, Asset Liability Management (ALM) and DFAST/CCAR models, because they incorporate effective lifetime and credit risk assessment, may be utilized (or modified) for CECL estimates as well. However, these models are used for regulatory and management purposes, not as a source of disclosures in financial statements.
- Documenting processes and controls – Documenting processes and controls will be a major undertaking. Ideally, areas of control weakness in the new processes should be identified as the processes are being developed, not after the fact.
- New skill sets – Many internal audit departments may require skills in data and modelling. Adequate budget must be provided for staff and training.
Q: Do you advise firms to develop benchmarking CECL models?
A: It may not be necessary to develop a complete benchmarking model. Nevertheless, during the development process, it is reasonable to assume that after considering a variety of alternative approaches, data and assumptions, a benchmarking model may emerge as a side product of verifying the performance of the primary model.
The bottom line is that the time for the internal audit function to develop key CECL-related objectives is now. What auditors have to audit has changed significantly. Data has a certain subjectivity, and auditors must ensure that subjectivity is reduced. In addition, auditors have to increase their skill competency – they have to increase their understanding of modeling and data analytics. To provide assurance, auditors must become confident of their skills and ability to analyze credit risk. The archived webinar is a good first step.