Dear Congress: Don’t toss CECL out, work with FASB to amend it
Many industry leaders, influencers and stakeholders continue to question and raise doubts around the new accounting standard for Current Expected Credit Losses before it takes effect in 2020.
These concerns range from CECL’s accuracy and volatility; its impact on loan-loss reserves and loan pricing; and the impact it could have on availability of funds to borrows. In fact, Rep. Blaine Luetkemeyer, R-Mo., who serves as the ranking member of the House Financial Services Subcommittee on Consumer Protection and Financial Institutions, recently wrote a BankThink article to that effect.
While some of these apprehensions have merit, there is no need to throw out CECL entirely as many have called for. Instead, lawmakers should look for ways to improve what’s already on the table — and quickly — so that the industry can reap the many potential benefits CECL can provide, while better protecting institutions and consumers.
A common misconception argued from those making the case against CECL is that it would not improve accuracy in loss predictions. CECL provides a broad range of acceptable methods for determining reserve requirements.
Sophisticated CECL solutions will look at historical loss rates, client risk migration patterns and the average life of a loan’s portfolio (among many other risk factors) to determine the risk associated with each loan, and the associated reserves. CECL results from good models will eliminate the need for the extensive subjective quality-factor (Q-factor) adjustments in the incurred loss model, making the results more accurate.
Even before the Great Recession, some large banks were starting to use CECL-like models but did not believe the warning signs provided. So they failed to incorporate it into Q-factor adjustments, which could have made a difference in what was to come. These failures from the Great Recession proved that incurred loss models with Q-factors did not adequately protect banks and consumers.
Another assertion is that the economic forecasts key to CECL are inaccurate. This statement has some validity in that creating long-term economic forecasts is a difficult and imprecise science.
Recent research has shown that a perfect economic scenario and an average economic scenario have only a one- to two-quarter difference in the peak loss point. Although not negligible in the midst of a crisis, the peak timing may be less important than the multiyear trend toward the peak. Even flawed economic forecasts are still more valuable than assuming the past continues unchanged.
The more important factor in a good CECL model is measuring changes in credit quality and understanding the credit cycle. Good CECL models that understand and identify changes in the credit cycle can provide institutions with years of advanced warning prior to an economic downturn, allowing them to adjust their CECL reserves and their lending policies to avoid future failures.
CECL’s cost to the industry and impact to loan rates has also been cited as a significant negative. Undoubtedly, CECL will require additional tracking, calculation and documentation. However, the real concern should not be about CECL costs but around helping institutions effectively price transactions based on the risk for that transaction in the current economic environment.
Mispricing was a core cause of failures in the last crisis. The information that lenders should be gathering to comply with CECL is information that they should have already been gathering to measure and price for the risks that they are taking on.
One of the most valid CECL concerns involves its impact to the mortgage industry. The CECL life-of-loan component will have a significant effect on the reserve for all long-term loans, especially mortgages. This could cause mortgage loan reserves to increase by six to eight times, likely impacting interest rates.
To counter this concern, regulators should either provide some capital relief for long-term loans or ask the Financial Accounting Standards Board for an adjustment to the life-of-loan concept for longer term loans, such as a maximum reserving horizon to help lessen the impact. This approach could also resolve uncertainties around line-of-credit reserves.
A final valid concern is the lack of quantitative impact assessment. Given the magnitude of this change, an assessment should have been performed by regulators shortly after CECL’s publication. With that opportunity past, regulators should examine what impact the International Financial Reporting Standard called IFRS 9 has had on other industrialized countries that adopted similar principles, and consider the differences between the two.
The single biggest variance is the treatment of longer term instruments. With IFRS 9, over 80% of a loan portfolio is allocated at a “single year” reserve amount, which is far different from that of CECL for a 30-year mortgage. The bounding of reserve horizon in IFRS 9 is the reason no impact study was necessary for a change.
While we do share some of the industry’s commonly cited CECL concerns — especially around the impact to the mortgage portfolio and long-term loans — we believe that with some adjustments for these loans, CECL is a far better solution in accounting for loan-loss reserves than the prior “incurred loss” model.
With the risk of a recession in 2020 increasing, this would be the wrong time to abandon greater portfolio monitoring. Recent history has shown that the costs of bank bailouts greatly exceed the costs of better portfolio intelligence.
Rather than an all-or-nothing conversation, relatively simple changes to CECL like limiting the reserving horizon could solve almost all of the major concerns.
With adoption slated for SEC filers in 2020, Congress, regulators and FASB must act quickly if they are going to make any adjustments to CECL prior to adoption. The significant potential benefit of CECL should not be disregarded because of a few concerns, as CECL concepts are key to safeguarding institutions and avoiding future credit bailouts.