CECL – Applying qualitative overlays that are transparent, robust and repeatable

CECL – Applying qualitative overlays that are transparent, robust and repeatable

By Stevan Maglic, SVP, Head of Quantitative Analytics, Regions Bank

CECL 2019 is taking place in New York City on 27-28 March, 2019 – find out more here www.cefpro.com/cecl

Can you please tell the Risk Insights readers a little bit about yourself, your experiences and what your current professional focus is?

I’ve got over 20 years of experience as a practitioner in risk management and line-of-business functions at Merrill Lynch, BMO, ABN AMRO, and most recently at Regions Bank.  I have worked extensively developing and implementing economic capital and other credit-related models.  I have also worked in the credit portfolio management space, developing and implementing credit trading and risk mitigation strategies.  Since the financial crisis, I have focused on developing champion/challenger stress testing methodologies, but more recently my focus has shifted to CECL and the integration with stress testing, economic capital, risk rating, valuation and other methodologies across the firm into a comprehensive framework.  Over the course of my career, I have spent considerable amount of time in front of regulators, model validation, audit, and internal stakeholders, defending methodologies and methodological choices.

What, for you, are the benefits of attending a conference like CECL and can you provide an overview of what you hope to discuss at the event and why the topic is so important?

Since most firms are in a mad dash to prepare for going live with CECL in 2020, now is a good opportunity to hear what others are doing and compare notes before everything gets locked down.  Although many firms have settled on methodological choices and made design decisions on the production environment, firms are only now starting to focus on qualitative adjustments that might be needed specifically for CECL in the setting reserves on an ongoing basis.  During my talk, I will describe what unique qualitative adjustments are relevant to CECL, above and beyond adjustments that are used in the incurred loss approach.

What are some of the challenges associated with developing models with the capability to project more accurately than CCAR models?

For the most part, CCAR models are accurate and do perform reasonably well.  However, the challenge is that CCAR models have been built with a specific design criteria and specific variables (e.g., unemployment rate) set by the Federal Reserve Bank.  For this reason, more can be done to produce a more forward-looking loss estimate.  For instance, number of hours worked decreases before the unemployment rate is impacted, and as a result can contribute to a more forward-looking model.  Since accurate forecasts are more relevant to CECL — rather than performance under stressed macroeconomic scenarios — making use of more anticipatory macroeconomic variables is more desirable.

Why is it important to develop models that are more anticipatory than your CCAR models?

During periods of economic stability, having an accurate forward view of the economy is not so critical.  However, as the current bull market continues into record territory, anticipating the next recession has become an active discussion at most banks.  In this case, how do we ensure that we are able to adequately anticipate the next recession and have adequately increased reserves to the appropriate levels? If we can’t do this, rather than having reserve levels that are “too little, too late” — as the current ACL process is often characterized — reserve levels could be instead “way too late” under CECL.

In terms of qualitative overlays, what non-linearity and multiple scenario considerations need to be made by financial institutions?

To some extent, this one comes down to semantics.  While most firms intend to make use of multiple scenarios, it is not clear if the effect of additional scenarios will be represented in the overlay or as part of a blended base forecast that includes probability-weighting of scenarios.  Either way, unanticipated changes in the economy can be represented as an added stress in the base scenario variables or — if multiple scenarios are used — in the scenario likelihoods.  To the extent that this process cannot adequately capture inflection points in the economy, then additional adjustments are needed in the form of an overlay.  What I am say here is that all of this needs to work together consistently.  In my talk, I will go over some of the alternatives in getting all this to hang together.

How can financial institutions best leverage their early warning frameworks to make decisions on qualitative overlays?

At Regions, we have existing CCAR models as well as a forward-looking EWI framework that we have designed to be more anticipatory than our CCAR models.  The process that we are currently developing will take in all this and additional relevant information to inform the overlay process.  Since judgement still needs to play a critical role in setting of reserves, we imagine a scorecard framework to support a robust, transparent, and repeatable overlay process.  At the conference, we will be going into more detail on how we are thinking about this, new methodologies, and how a framework like this could work.

Finally, in your opinion, how do you foresee CECL impacting the industry once all institutions have moved to implementation?

Current ACL methodologies vary considerably between institutions and CECL should bring about somelevel of standardization.  However, the FASB standard is not prescriptive and methodologies under CECL will still vary, but perhaps not quite as much.  In terms of impact on business activity, even if issues of capital relief are addressed, products like the 30 mortgage and other CECL “hogs” will likely be impacted in some way.

You may also be interested in…

Sign up for your free Risk Insights account: