The implementation dates for the new accounting standard known as CECL are fast approaching. As with any new regulatory or compliance change, there is often a great deal of uncertainty and ambiguity surrounding meeting the requirements. In this post, we attempt to cut through the noise and provide a concise summary of the fundamentals and what they may mean to your institution.
On June 16, 2016, the Financial Accounting Standards Board (FASB) issued the final current expected credit loss (CECL) standard. CECL standards apply to any institution issuing credit and filing under GAAP accounting standards. CECL is described by banking regulators as the largest accounting change in banking history. It requires that institutions account for the expected credit loss over the life of a loan at origination/acquisition and set aside capital for expected future losses.
CECL replaces the incurred loss standard, where financial institutions recognize credit loss when a financial asset reaches a probable threshold of loss. This standard requires that institutions account for credit losses expected in the next 12 months.
In practice this means that institutions often calculate the annual charge-off rates using historical data for a specific loan pool and adjust this rate for the current credit environment. Multiplying the adjusted rate by the balance of the loan pool yields the 12-month, credit-loss amount used in allowed loan and lease loss reserves (ALLL). This amount is often referred to as the ASC 450 or FAS 5 ALLL. The figure below summarizes the incurred loss standard.
In contrast to current methods, CECL requires that institutions recognize each asset’s expected lifetime credit loss immediately. Whereas the existing standard for loss accounting measures the current loss on a portfolio using backward-looking methodologies, the new standard measures the current risk of the portfolio using forward-looking methodologies.
Note the changes in the figure below. Rather than using historic annual charge-off rates, CECL requires banks to consider historic lifetime charge-off rates. Furthermore, lifetime charge-off rates must be adjusted for the current and future credit environments. Multiplying this life-time adjusted rate by the loan pool balance yields the total expected credit loss over the lifetime of a bank’s financial assets as the new ALLL.
Adjusting historic lifetime charge-off experience for the future credit environment requires forecasting. First, institutions need to determine which economic variables correlate highly with their loan loss history. Then, institutions can use proprietary forecasts or use publicly available data from the various government agencies.
Remember, George Box’s famous quote: “All models are wrong, but some are useful.” For this reason, institutions may be wise to consider forecasting loan losses under multiple scenarios. Regardless of the complexity and quantity of forecasts used, the selection of “reasonable and supportable” forecasts needs to be well documented and defensible to auditors. For assets whose life exceeds reasonable and supportable forecasts, historical loan loss rates can be used to bridge the gap in estimating future loan losses.
There is still a great deal of uncertainty surrounding this question, meaning there is no clear answer at this point. The new CECL requirements could increase an institution’s operating cost. This would be in the form of personnel, technological and governance costs of implementation. It is also possible the amount of capital set aside in loan loss reserves could increase. Again, however, much remains to be seen.
According to a survey by Sageworks, financial institutions expect changes to capital requirements to be the most significant impact of CECL. Because CECL requires an earlier recognition of credit losses than the existing standard, loan loss allowances are expected to increase and retained earnings are expected to decrease initially.
Although it is possible that ALLL balances in some banks will decrease, early estimates in 2011 suggested that CECL implementation would increase the ALLL by 30-50 percent. More recent estimates, however, are significantly lower. Nevertheless, ALLL balances are expected to increase on average, which reduces capital.
Institutions should be considering adjustments to pricing to account for the opportunity cost of setting aside additional capital. Variations in expected credit loss rates by product type create variations in the necessary additions to loan loss allowances by product type. Thus, the relative prices of products may adjust accordingly.
The same survey by Sageworks shows that financial institutions are also concerned about changes to financial statements. CECL requires that financial statements explain the credit risk inherent in the portfolio, how management monitors credit risk, management’s estimate of expected credit loss and changes to this estimate during the period.
Respondents also expressed concern about the necessity of cooperating across departments to calculate the ALLL. Accounting, credit, ALCO and other departments must work together to incorporate expected credit losses into the ALLL calculation.
This highlights an important and overlooked challenge presented by CECL – governance. Holistic approaches built on transparency and coordination will facilitate auditability and repeatability. It is suggested that banks form cross-departmental teams to optimize the production workflow for CECL implementation.
The viability of the above approach, however, is significantly impacted by the institution’s size and complexity. Although a cooperative approach may be appealing, it can easily over-complicate things for a smaller institution. The focus should be on developing a sustainable methodology and one that should improve over time. It should also be designed to be flexible as the requirements and regulatory expectations are still largely to be determined.
The effective date for CECL varies by the institution type. The effective date for public business entities (PBEs) that are SEC filers is the fiscal year beginning after December 15, 2019. The effective date for other PBEs and non-PBEs is one year later. See the table below for a summary of the effective dates.
| Type of Institution |
U.S. GAAP Effective Date |
| PBEs That Are SEC Filers |
Fiscal years beginning after 12/15/2019, including interim periods within those fiscal years |
| Other PBEs (Non-SEC Filers) |
Fiscal years beginning after 12/15/2020, including interim periods within those fiscal years |
| Non-PBEs |
Fiscal years beginning after 12/15/2020, and interim periods for fiscal years beginning after 12/15/2021 |
| Early Application |
Early application permitted for fiscal years beginning after 12/15/2018, including interim periods within those fiscal years |
FASB defines PBEs as institutions that meet one of the following conditions. [1]
The most urgent task required for banks to prepare for CECL is data collection. Insufficient and incomplete data limits the credit-loss estimation methods available to the bank. Additionally, the quality of loan loss estimations and the extent to which it can be relied upon for pricing and other banking functions depend on the comprehensiveness and quality of historic loan loss data.
Therefore, banks need to consider the readiness and quality of their data immediately if they have not done so already. What data is available? What data is needed? How can my institution obtain and use the necessary data to not only comply with CECL, but also to improve the profitability of my bank?
A robust implementation of CECL requires more granular and complete data than what many banks are currently using in loan loss calculations. Under the current standards for loss accounting, institutions often rely on aggregate, pool-level data. Banks typically need data for charge-offs, recoveries, aggregate pool balance, beginning pool balance and ending pool balance.
More complex CECL methodologies require detailed loan-level data including, but not limited to, measurements of loan duration, customer and book balances, segment identifiers, credit quality, origination dates and amounts, renewal dates and amounts, charge-offs and recoveries. For banks that currently do not collect detailed data as described above, this requirement is one of the most immediate hurdles of CECL implementation.
As we will discuss in future posts, data requirements also represent one the greatest opportunities of CECL implementation. Loan-level data can be used to improve pricing, capital planning and risk management. Accurate modeling of credit risk over the life of the loans in a bank’s portfolio provides the bank’s decision makers with the tools to improve profitability.
For many mid-size and smaller banks, gathering the data necessary to meet more complex CECL requirements may seem like a challenge. Indeed, community banks may take the path of least resistance and simply adjust their current loan loss calculations to meet the new standards. However, more robust CECL implementation represents an opportunity to improve the quantity and quality of data that institutions collect. In so doing, banks can accelerate the adoption of analytical tools to improve their decision making and their bottom line.
CECL is not prescriptive of the specific estimation method for modeling credit loss. Methodological choices span a wide range of quantitative complexity and data intensity. Banks will aggregate loans into portfolios based on shared characteristics and select the appropriate methodology for each loan pool. Many institutions will elect to run multiple methodologies before selecting the best fit for each pool. Regulators and auditors expect methodological complexity to vary with bank size.
In a webinar targeting community banks and conducted by members of the Fed, the FDIC, the CSBS, the FASB and the SEC in February 2018, three methods were discussed, including
Other commonly discussed methodologies include [2]
Institutions will select the appropriate model for each portfolio given the availability of data and the appropriateness of each model for the loan characteristics in each portfolio. For example, DCF estimation is data-intensive but requires little historical data. Also, revolving credit creates complications for classifying loans in a specific vintage.
In our estimation, for institutions wishing to improve their forecasts, the PD/LGD model will likely emerge as the best fit for most institutions due to its intuitive appeal, usefulness, and versatility. Each methodology, however, may present different opportunities for leveraging results for use in different areas of the bank. We will discuss this in more detail in future posts, but the selection process for CECL methodologies should consider the benefits of each methodology for other banking functions.
[1] Source: Pages 5-6 of the December 2013 FASB Accounting Standard’s Update.
[2] For more information regarding methodologies and the data required for each, see the Sageworks webinar .How to cite this blog post (APA Style):
Premier Insights. (2018, July 5). The Essentials of CECL: What Your Institution Needs to Know [Blog post]. Retrieved from https://www.premierinsights.com/blog/the-essentials-of-cecl-what-your-institution-needs-to-know.