Do I need to start implementing CECL now?

New Call-to-action

cecl.pngIf you fail to plan, you plan to fail. I heard this phrase so many times during my childhood (from a wonderful, loving, and … organized father), that I’d made a goal to never repeat it in my lifetime. Well, here I am, repeating it.


Anybody who has had CECL on the mind, knows that this phrase aptly describes the attitude lenders should have in preparing for CECL adoption in 2020. A successful CECL implementation will require early planning on a number of levels: methodology, potential changes in reserve amounts, and data storage, to name a few. I’d like to focus on data storage, as that is a topic of question and concern for many institutions and is really at the heart of whether you need to start implementing CECL now. So when is it necessary to begin storing data? And what data should be stored? The answer is a resounding and very unsatisfying “it depends.”


It depends largely on the methodology you choose to adopt, and thanks to FASB’s pursuit of flexibility in CECL regulations, there are more possible methods than there are Land Before Time movies (to see just how many that is, click here). If you’re using a PD model built by a third party, such as Visible Equity’s model, you may not need any historical data. In fact, it is probably more important in that scenario to have comprehensive data on your active portfolio as of the report date (FICO, LTV, etc.). On the other hand, if you plan to use your own data to build a PD model, you will need a lengthy and complete history of loan-level data (for all loans -- not just those that charge off). The story is similar for loss rate methods. Because loss rates are typically calculated on your own historical data, life will be much easier come 2020 if you start deliberately storing data now. Obviously, you will need to store charge-off data, but storing prepaid and matured loans are also important in calculating an accurate loss rate. Though there is still some flexibility in methodology within the loss rate realm, you will likely need time-specific values of the following for all loans:


  1. Loan type
  2. Origination date
  3. Credit quality indicator (i.e. grade)
  4. Unpaid balance
  5. Charge offs
  6. Recoveries


It is important that you not delete this data after a loan closes. The longer your archive of historical data, the more accurately you will be able to make predictions about lifetime loss.


Long story short, start now! Data storage is cheap enough that the benefits of building up your data far outweigh the cost. As the world is becoming more and more data-driven, good data will benefit you in CECL and beyond.

Rachel Messick

Product Manager/Data Scientist at Visible Equity