CECL Q&A – Methodology

The FASB’s Current Expected Credit Loss (CECL) model presents unique challenges for banking professionals. To help institutions prepare, Abrigo launched a CECL webinar series covering data, segmentation, methodology, and forecasting requirements broken down by loan pool-type.

A key component of the series is allowing participants to ask their CECL-related questions and below are several questions with answers relating to different methodologies; specifically, we take a look at common questions for running vintage and migration analyses.

Vintage Analysis:

Do you need to have both the origination date and renewal date? Is it okay to replace the first origination date with the renewal date when it happens?
Yes, you do need both the origination and renewal date. No, it is not acceptable to replace the first origination date with the renewal date when it happens. For example, if we originated a loan in 2012, it had a 3-year maturity and we renewed that credit in 2015. That loan would have a 2012 and 2015 Vintage. Any events that happened to that credit after the renewal should be tied to the 2015 Vintage and subsequently, any behavior before that renewal should be tied to the 2012 Vintage.

Regarding extensions/renewed loans, do you count the extended/renewed life as part of the life of the loan?
In most cases, a renewal would result in a “new loan” for vintage or any life-of-loan analysis. The guidance contained within the new standard (ASU 326/CECL) points to a “more than minor” test contained within ASC 310-20: “This condition would be met if the new loan’s effective yield is at least equal to the effective yield for such loans and modifications of the original debt instrument are more than minor.” In practice, more than minor has been determined to be >10% change in NPV. Also, short-term extensions that facilitate administrative needs are ignored and/or do not meet the “more than minor” test. For example, a 90-day extension would likely not constitute a new loan while a full 3-year renewal would.

Migration Analysis:

What’s optimal for loss rate calculation? 36-months? Longer? Shorter?
You need to run your average life calculations by pool type and understand how your portfolio is performing. There is no ubiquitous rule-of-thumb that would apply to all institutions. Each pool will have a different average life.

When we are talking about migration analysis, are we only concerned with the beginning credit quality indicator and how much of that pool moved to loss OR are we following the pool through all grade changes through the life of the loan?
Loss typically occurs when a credit goes to a specific risk-rating (i.e., Sub-standard); therefore, loss rates in every other risk-rating category except for that one would be zero. So the idea here is, we are freezing a loan with its attribute at that point in time and watching it perform into the future. The future behavior will be pulled backward to that loan’s risk-rating and balance at that point in time.

If you are interested in more answers to CECL, watch the on-demand webinar, CECL Methodology Q&A.