Based on a presentation given by Dr. Joseph Breeden, CEO of Prescient Models LLC., at the 3rd Annual Analytics and Financial Innovation Conference (AXFI).
The Financial Accounting Standards Board (FASB) announced the final standard for the Current Expected Credit Loss, or CECL, on June 16, 2016. Thus far, banks and credit unions have only been required to estimate and report potential losses for the next 12 months on loans and leases. This new regulation, however, will require financial institutions to predict and report potential losses for the entire lifetime of a loan.
What it means for your Credit Union or Bank
The most important thing to keep in mind is you will have to report potential losses from the present age of the loan to its furthest non-cancellable end point, or the loan’s “lifetime loss”. FASB’s regulations are “artfully vague”, by making this change sound simple without ever mentioning best practice, validation, or examination. All of which will be of increasing importance as you attempt to explain your method for arriving at the reported loss reserves. In other words, you need to “have a nice story” to explain your reports.
A major challenge you will now have to face is how to calculate these potential losses. Unfortunately, the popular “moving average” method will no longer work. The “vintage model” is the bare minimum example to be CECL compliant. However, reporting with such a minimal method will require you to spend much more time explaining your methods and trying to prove the results are valid.
Most likely, if you are capable of doing so, the best option is to use modeling. This method will be the best to provide accurate results, which is not only beneficial for CECL compliancy, but your own forecasting. The truth is, reporting potential loss over the lifetime of a loan is very “doable”. This will likely require a change in your reporting methods but doing it properly might help your business avoid a lot of cost and effort.
Data is the Key
The CECL regulation has been finalized and will go into effect December 15, 2020. Fortunately, this gives you some time to prepare. However, you should begin preparing for these changes now. Modeling is only part of the solution. A model is like an engine. It can’t run without fuel. Data is the fuel that makes the modeling engine run. Just like actual fuel, quality and volume make a difference. High quality data (fewer errors) and a larger volume of data (more years to better capture trends) result in better models, which means better forecasting, and less extensive explanations necessary in the final report.