Maximizing CECL data for strategic insights

Rate studies

Understanding how the output described above can be helpful aside from CECL, let’s look at uses for certain input data beyond calculating the allowance. After all, building relevant CECL models using an institution’s own data is generally preferred to relying on peer-based studies. Having the loan-level data necessary for a CECL calculation all in one place opens opportunities for financial institution managers to glean statistically relevant observations of portfolio performance over time. For example, the loan files typically give us the data we need to produce prepayment rate studies: observations of balance changes period over period. These studies often impact the CECL results themselves, but they can be useful in other areas like pricing and ALM as well.

Economic stress testing

It’s important to fully understand a CECL model and be able to predict the model performance in varying economic conditions to allow managers to anticipate needed accounting entries. Economic stress testing is made much simpler when a CECL calculation exists in the same system to leverage common inputs. Whether it is for adherence to DFAST or an institution’s desire to gather deeper understanding of their model’s potential, having CECL and stress testing available together to play off each other is an efficiency directly resulting from the data gathering required for CECL models.

Business intelligence and insights

The same files that come into play for a meaningful CECL calculation can be used to produce insightful business intelligence for a financial institution.  Advanced banking intelligence tools like Abrigo’s Connect can produce, for example, past due trend analyses that give visibility into portfolio performance that wouldn’t be possible without the data sets and time series data needed for CECL. There is power in generative AI tools that can analyze data and give access to the results financial institutions otherwise would have likely missed.

Another example of intelligence is heat mapping data that has been generated related to both collateral and borrower files (all part of a CECL data stream). Financial institutions serving areas hit by the California wildfires of this year have been able to leverage these data files within Connect to identify areas impacted within their portfolio. They have been able to cross reference the heat maps to the government databases of wildfire impact to respond more quickly. The data can help them offer better service to their borrowers and begin anticipating the negative impact of the fires on their unique loan portfolios.

What-if and trend analysis

This is a bit of a subcategory of point #4, but a quality CECL tool with the datasets required for a calculation can provide insight that can’t easily be replicated outside of a platform of its kind. For example, probability of default trend analyses are produced as part of certain methodologies used in creating a CECL calculation. But they also offer insights to credit teams who are generally not even involved in CECL calculations. The observations of historical trends created for CECL allow credit risk managers to identify what may happen given projections in real time. As a result, the financial institution enhances its ability to anticipate events rather than simply react to them. Coupled with additional analysis, the influence of these trends can guide credit teams in pricing loans and other related activities.