Looking at third-party options. Most non-public business entities aren't required to implement the current expected credit loss (CECL) model until fiscal years starting after Dec. 15, 2020. However, many credit unions are heeding the advice of the NCUA and advancing steps to ensure effective implementation of this major change in estimating losses. The model will require more inputs, assumptions, analysis and documentation, making the option to automate and modernize the process significantly more attractive than under existing standards. Credit unions considering software to comply with the regulations may choose to build their own software solution or work with a third-party vendor. In either case, credit unions have several considerations to help narrow the playing field.
Data
Data is the foundation of historical loss experience calculations. In fact, a credit union's ability to calculate lifetime loss rates as of the implementation date is predicated on the adequacy of yesterday's data. In many instances, however, the historical data required to properly evaluate the performance of various loan types and risk attributes is not available or is riddled with inconsistent or incorrect information. Credit unions will benefit from comparing internal and vendor options for ensuring data adequacy and data retention. Clarify what third-party providers offer for data archives, data architecture and data adequacy services, and compare that to in-house capabilities.
Three areas of consideration related to data adequacy are:
- Identification of loan-level data gaps;
- Assessment of the impacts of identified gaps on methodology selection, input calculation and forecasting; and
- Remediation for loan-level historical data – both for data correction and proper go-forward data collection.
Areas of consideration related to data retention include storage capacity and fees associated with data storage. If a third-party vendor is chosen, identify the process and costs for handling data in a separation event. Also understand the mechanism(s) by which loan-level data will be transferred and how much human interaction vs. automation is involved in the process. Clarify the controls for manual adjustments to loan-level data and the timeline for successful data transfer.
Methodologies
The new accounting standard does not specify a single method for measuring expected credit losses. Indeed, regulators have said that "allowances for credit losses may be determined using various methods that reasonably estimate the expected collectability of financial assets and are applied consistently over time."
In order to decide what changes to current methodology are best, most institutions will run scenarios parallel to the institution's current ALLL to see what is most accurate and what potential impacts may be. Having various methodologies to choose from will provide flexibility as data issues become apparent during execution.
Before deciding whether to build or buy a CECL solution, credit unions should compare which of the following methodologies will be available through each option: Vintage, probability of loss and loss given default (PD & LGD), migration, cumulative loss rate and discounted cash flow (DCF). Consider costs for developing and maintaining these options. How automated and flexible are the options? What, specifically, is the process for calculating, evaluating and changing methodologies? Consider what in-house or vendor-provided advisory assistance and support is available for each of the methodologies.
Forecasting and Adjustments
Institutions cannot rely solely on past events to estimate losses and must attempt to create a reasonable and supportable forecast. For periods beyond an established forecast, reversion to average historical experience is required. Qualitative adjustments to historical loss experience will remain under the new standard as a viable approach to accommodating the impact of forecasts. Models should allow for swift inclusion and exclusion of all observable analysis periods and provide forecasting intelligence, support and application. Ideally, management should be able to evaluate all observed loss rates, make any documented exclusion/inclusion decisions, include forecasted conditions and see those decisions reflected in the estimated reserve level.
Model Risk and Auditability
Documentation and support is often a most time-consuming exercise in today's ALLL processes. The new standard will require more inputs, assumptions and analysis at the pool-level. Consolidating and displaying all information necessary to review, support and recalculate will be a critical function of any solution.
Consider what level of transparency and auditability a solution offers, and take into account associated costs and time spent preparing documentation and support when utilizing a solution developed by the institution. Credit unions might be tempted to use an analytical tool for an allowance model. However, analytical tools may lack the transparency and rigor that auditors and regulators might expect when results will end up on your financial statements.
Neekis Hammond is a Senior Consultant at Sageworks
984-242-2736 or [email protected]
© 2025 ALM Global, LLC, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to [email protected]. For more information visit Asset & Logo Licensing.
