Transitioning Your Incurred Loss Methodology to CECL
Regulators have urged institutions to leverage their current methodology for the allowance in the transition to CECL but make changes to it, primarily with respect to life-of-loan and forecasting requirements. The FASB expects the transition to be scalable, but that “the inputs to the allowance estimation methods will need to change to properly implement CECL.”
Compared to current GAAP, which requires you to reasonably estimate the amount of loss that has been incurred (upon meeting the probable requirement), CECL is more about what cash flows you expect not to collect at origination. And while current GAAP sees all loans as good until proven otherwise, under CECL, all loans are originated with some measurable risk of default.
How to Pool Loans
CECL requires loans to be pooled or segmented according to shared risk characteristics for measurement. Start that process by looking at how you are analyzing your risk segments now and how they will line up for CECL. Most institutions are using a call report structure on which to base their pooling. CECL will require a more granular approach than that, but you can start there as call codes contain much useful information. Your initial pooling decisions are likely to change as you move through the process, like your methodology choice and Q-factors. You will have to continue to ask yourself if your pooling structure remains appropriate given your risk profiling.
Things that may bring about a change in pooling structure are new products, changes in underwriting and standards, and loan acquisitions. The way you look at risk under CECL is going to change. CECL will force you to peel the onion back to understand why a loan went bad and that will give you a clearer picture of your risk categories. Most banks haven’t been accumulating the data they’ll need for CECL, so by year four or five you will have that data and have honed your model to truly reflect the risk in your portfolio.
The Q-factors you apply now to your allowance will likely be those you’ll use for CECL. Additional consideration will be needed to address projections under the reasonable and supportable forecasts requirements. There is a good chance that the current regulator recommended minimum Q-factors will grow in number. The key is being able to defend what you use. As you look at your loss history, identify the factors in your market that caused those particular losses and that should give you a road map to identify a new or expanded list of Q-factors that address your portfolio.
Q-factors will be applied to pools, which reasserts how CECL is more about pools than looking at your whole portfolio in aggregate. Q-factors will be vital for smaller institutions, and it will be important overtime to ensure the Q-factors you use reflect your existing portfolio.
In deciding on a methodology, start with identifying what you can pull out of your incurred loss model. Spend more time deciding what fits you rather than which methodologies are most popular. Ask yourself what you want to accomplish as an institution and which model gets you there, for example, do we want to get a lot of business intelligence out of this?
Weighted average remaining life is gaining steam as a methodology. At first, it was thought it wouldn’t be complex enough, but it now appears regulators are more open to it, more willing to approve it for smaller institutions with portfolios that aren’t very complex.
New computational CECL issues include:
- Contractual term adjusted for prepayments
- Reasonable and supportable forecasts
- Limited usage of “no allowance”
- Purchase credit impaired (PCI) loans become purchased credit deteriorate (PCD) loans
- Debt securities accounting
- Elimination of directional consistency and
Visit Abrigo’s portfolio risk resources for more information on the incurred loss model and prepare for the transition to CECL.