Skip to main content

Looking for Valuant? You are in the right place!

Valuant is now Abrigo, giving you a single source to Manage Risk and Drive Growth

Make yourself at home – we hope you enjoy your new web experience.

Looking for DiCOM? You are in the right place!

DiCOM Software is now part of Abrigo, giving you a single source to Manage Risk and Drive Growth. Make yourself at home – we hope you enjoy your new web experience.

CECL Data for Credit Unions: Navigating Uncertainties

Mary Ellen Biery
March 19, 2021
Read Time: 0 min

How credit unions can manage CECL data challenges

As credit unions prepare for the Current Expected Credit Loss standard, they'll uncover several data issues they'll need to address.

You might also like this webinar: CECL in 2023 - Steps to Take This Year


Related Subhead

Different model, different data needs

Credit unions’ CECL preparation efforts can reveal challenges that are quite different from those facing banks, and data-related concerns are a prime example. However, understanding these data issues and sound planning can ensure a smooth credit union transition for the Current Expected Credit Loss model, or CECL, by the 2023 deadline. Abrigo advisors recently discussed CECL data and credit unions during a webinar, “CECL in 2023: Steps to Take This Year.”

The current expected credit loss model (CECL) doesn’t prescribe specific approaches for credit unions to develop estimates for the allowance for credit losses. However, for credit unions especially, data is a key consideration in CECL methodology selection.

“The standard gives you a lot of options on how to estimate future credit losses,” Garver Moore, Managing Director of Abrigo Advisory Services, said during a recent webinar hosted by Abrigo. “And those different options have different data requirements, including the amount of data you need, how far back it goes, and what that data covers.”

What data is needed for CECL

In its Frequently Asked Questions on CECL, the National Association of Federal Credit Unions (NAFCU) noted that the specific data used in CECL models will vary among institutions. Nevertheless, a survey NAFCU performed found respondents anticipated collecting 22% more data points than they did presently. Common data pieces for CECL, according to the NAFCU, may include:

  • origination dates and balances
  • maturity dates
  • changes to delinquency status
  • loss history
  • member information, including risk indicators
  • other segmentation data

Collecting vital loan-level information for credit unions can be challenging when the data reflects a high number of smaller-dollar, risk bearing loans that are stored in multiple systems, Moore said. Going from data source to data source to capture data adds time and complexity. Keep in mind that such an added burden related to the CECL transition will be coming at a time the credit union is continuing to report the ALLL under the existing incurred-loss model.


How much data is enough?

Some credit unions find that partnering with a third party can streamline data gathering and identification of critical inputs to support a CECL methodology suitable for the institution.

"“If you're thinking, ‘Oh, I need eight years of data to do this,’ or whatever, full stop. You don't.”

It’s also important to remember that CECL doesn’t require having enough quality data to vet every available methodology before selecting one, Moore noted.

“If you're thinking, ‘Oh, I need eight years of data to do this,’ or whatever, full stop,” he said. “You don't.”

Credit Unions’ CECL forecasts can utilize external data

Furthermore, simply having copious data won’t necessarily provide useful insight for CECL. A credit union might have 15 years of data they could access with a lot of effort. But without a material loss during that period, the data won’t be especially useful to forecasting losses under CECL.

“If we do have 15 years of data, we have things to measure,” said Abrigo Senior Consultant Jared Mills, CPA, during the webinar. “But we work with credit unions that a lot of times, that data doesn’t just live in one place. That could mean having to aggregate from multiple sources. We really start to think: Are we making the best use of our time? Would we better served to work on the current data set?”

External sources

If a credit union has insufficient data, has inaccurate data, or has found gaps in its own data, it will likely need to make assumptions based on peer or industry data. For example, a credit union with insufficient data on commercial real estate losses, can supplement it with forecasted probability of default (PD) and loss given default (LGD) rates from CRE benchmarking data.

“You’re using available information to construct this estimate,” Moore said. “That means internal, external. That means top-down. That means bottom-up." Even so, he cautioned financial institutions from buying into the myth that depth and breadth of data is required for all CECL estimates.

Get help preparing for CECL.  Learn more

Reasonable & Supportable

Credit union concerns about accurate forecasting

Another area where data comes into focus with CECL is related to the requirement that credit unions incorporate “reasonable and supportable forecasts” in developing the allowance for credit losses. Indeed, the forecasting component is a major component of what makes CECL different from the incurred-loss method.

However, utilizing or applying a reasonable or supportable forecast is different from getting the forecast right, Moore said. A lot of financial institutions get tripped up on the idea that they will need to substantiate that they are correct in their forecasts, he said. “The reality here is that all we have to do is apply a forecast,” he said. “I personally cannot think of a client that has been meaningfully criticized [for their forecast] to the point of having to do something different on their forecast inputs.”

External sources for forecasts are OK

The important aspect is to follow a process in forecasting and in applying the forecast to the estimate, Moore said. “That process could be going to public sources like the Fed. It could be a partnership with a private provider, and we have integrations with some of those.” The National Credit Union Administration also provides economic data related to unemployment and housing price growth at the state and city levels.

Unemployment rates, for example, could come from:

  • the Federal Open Market Committee’s forecasts
  • a state economic agency
  • a local university that can capture the nuances of your market

The forecast is bound to change, particularly given the current economic environment, Moore said. But the application of the forecast should remain consistent through various environments.

Related Subhead

Backtesting enables CECL flexibility

The varying nature of forecasts makes preparation for CECL even more valuable, especially if the work allows the credit union to backtest its assumptions before CECL is effective.

Financial institutions that have already adopted CECL have reported that running quarterly CECL calculations in parallel with their existing ALLL methodology helped them make adjustments to their models to fine-tune the accuracy.

The backtesting and monitoring effort will “play really nice with those projection methodologies,” Mills noted. “At a point in time, we did an effort to estimate what future balance positions would look like, what future loss positions look like, and now, we can actually see, how well did we do that?”

Taking steps now to implement CECL affords credit unions the ability to remain flexible throughout the process – not only with data gathering, but also with model adjustments.

Stay up to date on CECL and other portfolio risk topics.

About the Author

Mary Ellen Biery

Senior Strategist & Content Manager
Mary Ellen Biery is Senior Strategist & Content Manager at Abrigo, where she works with advisors and other experts to develop whitepapers, original research, and other resources that help financial institutions drive growth and manage risk. A former equities reporter for Dow Jones Newswires whose work has been published in

Full Bio

About Abrigo

Abrigo enables U.S. financial institutions to support their communities through technology that fights financial crime, grows loans and deposits, and optimizes risk. Abrigo's platform centralizes the institution's data, creates a digital user experience, ensures compliance, and delivers efficiency for scale and profitable growth.

Make Big Things Happen.