A version of this article previously appeared in the California Bankers Association Regulatory Compliance Bulletin on August 15, 2016.
In a move that will change the way financial institutions determine their allowance for credit losses, the Financial Accounting Standards Board (FASB) released its Credit Impairment Update on June 16, making the much anticipated Current Expected Credit Loss (CECL) standard official.
This will impact entire institutions—not only the accountants. Specifically, loan officers, internal auditors, chief credit officers, and IT personnel can expect an increased workload as a result of the new standard. While the changes can’t be adopted before 2019—the more likely adoption dates will be in 2020 for SEC Registrants and 2021 for other entities—institutions should start preparing now to address changes. Some of the new standard’s key conceptual changes include:
- Removing the “probable” and “incurred” loss recognition thresholds used to estimate the allowance for credit losses today, effectively doing away with existing practices for determining the allowance for credit losses
- Basing loss estimates on lifetime “expected” credit losses
- Requiring that determination of lifetime credit loss estimates using past and current events be supplemented with “reasonable and supportable” expectations about the future
These changes fundamentally alter the way financial institutions will account for estimated credit losses on not only their loans but also debt securities.
The issuance of the new standard concludes an exploration that began in the wake of the global economic crisis. During that time, the delayed recognition of credit losses associated with loans was identified as a weakness in the application of existing accounting standards, which contributed significantly to the financial crisis.
More specifically, the problem was that the existing “incurred” loss model delayed recognition until the probability of a credit loss was established. As a result, the FASB began exploring alternatives that led to the use of more forward-looking information, a requirement under the new standard.
Although the CECL standard applies to multiple types of financial assets, we’ll provide an overview of how this affects loans, loan commitments, and their related allowances.
Why Does this Matter to You?
The allowance for credit losses is the most significant estimate at virtually every financial institution. Adding complexity to an already-challenging estimate is certain to create strain on the financial institution. That’s likely to spill over to employees who historically have had limited to no involvement in the allowance estimation process.
Loan and Credit Officers
If the CECL methodology adopted is heavily based on cash flows, those closest to the customer—the loan officers—will inevitably be involved in forecasting loan-level cash flows. This applies particularly to credits that aren’t performing—or aren’t expected to perform—within the loan’s original contractual terms.
The portfolio monitoring and management process and the traditional risk rating change that triggers additional reporting is likely to be accompanied with an updated cash flow forecast provided to those in charge of the allowance process. Collecting and documenting guarantor information in the loan accounting system is likely to become more important in the allowance process as well.
There will also be additional considerations for institutions if the CECL methodology adopted is along the lines of a probability of default (PD) and loss given default (LGD) methodology. In that case, dual risk ratings and regularly updated collateral values will be necessary to appropriately evaluate the LGD, making it potentially necessary to add data fields to current loan accounting systems. Existing data fields also will need to be better utilized.
Internal Auditors & Data Accuracy Expectations
Auditors, along with regulators, are likely to push for increased rigor with regard to model validation as complexities arise within the CECL models. Institutions also may stratify, segment, or group loan portfolios by loan type as well as origination year, maturity date, and numerous other factors. Segmenting the portfolio will be heavily reliant on system data, making data accuracy and integrity more important. Having an inaccurate origination date, maturity date, interest rate, or collateral value in the system today probably won’t significantly impact your allowance estimate, if at all. However, calculations of future allowance estimates will likely utilize this data and become more complex once the standard takes effect.
Controls to ensure accuracy, proper updating, and security of the data will take on increased importance, as well as a renewed focus on validating the data itself. Budgeting and forecasting will inevitably become more complex and “reasonable and supportable” expectations will be subject to greater scrutiny (and audit).
Why Segmenting Matters
A commercial loan with a 4 percent interest rate and a 40 percent loan-to-value (LTV) ratio that matures in 14 months may have a different credit loss estimate than a loan with an 8 percent interest rate, 85 percent LTV, and 55 months to maturity. Under many allowance methodologies that exist today, the calculated loss rate is probably the same on these two very different loans as long as their risk ratings are the same. Collateral values and interest rate tiers will likely be used to further segment portfolios when estimating the allowance for credit losses under CECL.
Reasonable and Supportable Forecasts
The larger the institution, the greater the expectation will be that all the underlying loan data is appropriately maintained to determine correlation to predictive internal data or external economic data.
In the simplest example, an institution could have a CECL methodology that’s analyzed historical national unemployment rates and correlated losses in a particular portfolio segment to those rates. As a result, the institution would be able to make a reasonable prediction about loss rates in the future for that segment based on the losses experienced in the past and forecasted unemployment. Practically speaking though, an institution would really need to correlate more than just the single factor used in this example for statistical accuracy and also make adjustments for borrower-specific considerations.
The expected portfolio segmentation enhancements for determining the allowance will also inevitably impact loan production. If a specific loan segment is performing poorly, then it likely will have a higher forecasted expected loss rate, absent specifically supported adjustments. That means you’ll need additional incremental reserves relative to the rest of the portfolio every time you book a loan in that segment.
As a result, the next loan you take to loan committee might not get approved simply because a specific loan type requires an abnormally high level of reserves when booked on day one. Strategically, someone in the approval process may suggest the institution would be better served to not book a new loan requiring a 15 percent reserve when other loan segments may only have a requirement of 0.50 percent.
Following are several of the key areas to be aware of when considering the impact of CECL on loans and commitments as well as new methodologies for assessing allowances:
- Prepayments. Your expected losses are to be measured over the contractual term. Consideration is given for prepayments but not for expected extensions or renewals unless a troubled-debt restructuring is likely.
- Loan pools. This is required for loans that share similar risk characteristics. However, keep in mind methodologies and models can vary by loan pool. Further, what constitutes a “similar” risk characteristic is highly subjective, giving institutions the opportunity to document and support loan-by-loan accounting if deemed appropriate.
- Determining methodologies. Widely expected methodologies to be utilized are discounted cash flows, historical loss rates, provision matrices, vintage analysis, and regression modeling. However, there’s no prescribed methodology.
- Credit loss determination. All available information relevant to collectability shall be considered, including historical credit loss information and qualitative factors specific to the borrower and entity’s operating environment.
- Forward-looking information. Estimated losses shall consider relevant internal and external information. This includes using “reasonable and supportable” forward-looking information to inform credit loss estimates.
- Reversion methodology. An entity will revert to its historical credit loss information—adjusted for borrower-specific considerations—for future periods where forecasts aren’t adequately supported. For example, if an entity can reasonably support a forecast for only the next two years and the contractual life of the loan is five years (with consideration given to prepayments), the last three years should be based on this reversion methodology. Further adding to the complexity of the estimate, the current economic climate could result in a change to the forecast period each reporting period as well.
- Commitments. Off-balance sheet credit exposure will need to consider both the likelihood and amount expected to be funded over the estimated life of the commitment.
Entities of different sizes and capabilities should expect differing degrees of model sophistication. Smaller entities, for example, aren’t expected to be required to implement complex or costly models when adjusting existing allowance methodologies.
If you aren’t responsible for loan-level data, you’re probably at least utilizing it. Open a dialogue with those typically charged with adopting new accounting standards, and consider organizing an implementation team to deal with the far-reaching impact of CECL.
The team would communicate the importance and need to focus on data—internal and external. There’s also an opportunity for you to get out in front of this process with effective communication between the right parties.
Here are some practical first steps for the implementation team:
- Preserve your loan data
- Develop a formal loan information management process
- Identify what data can be recovered quickly and economically
- Determine missing data and the cost of acquiring it
- Enhance understanding of collateral values and credit scores data as well as your ability to archive and update it in your system
- Improve the quality of guarantor data
- Understand which systems your data interfaces with
- Accumulate historical and forecasted national economic data (unemployment rate, treasury rates, or consumer price index, for example) to correlate to historical losses for forecasting purposes
Adjusting to the CECL standard will require collaboration. Accountants will have a good understanding of when the standard will need to be implemented as well as the new accounting and disclosure requirements. Credit risk management should be thinking about modeling options and portfolio risk management. Both of those groups of people will likely benefit from collaborating with individuals who having intimate knowledge of:
- Customer operations supporting repayment of the loan
- Loan data —inside and outside of the system
Personnel with critical loan and system knowledge who get involved early can provide key insights to help financial institutions navigate one of the most significant accounting changes in recent history.
We're Here to Help
If you’d like more information on how to prepare for the anticipated changes coming with the new CECL standard, contact your Moss Adams professional.