You’ve Been WARMED: Why SEC Filers Should be Leery of the WARM Method for CECL

Earlier this month, FASB published a Q&A related to the applicability of the Weighted Average Remaining Maturity (WARM) Method for CECL compliance.  This method is widely considered to be the simplest approach to CECL since it most resembles what many banks use to calculate their ALLL today.  FASB’s Q&A confirms that the WARM method can be utilized for CECL, but the usual host of caveats leave such a wide gap for interpretation you can drive a truck through them.

The vast majority of the “SEC filer” class of community banks with less than $10 billion in assets appear to be homing in on WARM as their go-to method for CECL.  This is after spending ridiculous amounts of money on a software-as-a-service “solution” (SaaS) that they thought was going to be their CECL panacea, only to discover that their SaaS was nothing more than a glorified calculator, and that the tough work was just beginning:  figuring out how to gather their data, pool their loans, select what methods to use for those pools, and then what assumptions to feed into those methods.  Yikes!  By the way, I don’t blame the community banks– the software vendors too often oversold their solution.

Facing this quandary, some of these community banks have since spent even more money with consultants, including the Big 4 accounting firms, to help guide them.  These same consultants are now advising the banks to gravitate toward the WARM method because of its simplicity and what seems to be its mounting momentum toward preference, highlighted by FASB’s Q&A piece.

The WARM method certainly makes sense for the $200 million community bank that is privately held and does not have a concentration in CRE or Ag lending.  After all, the stakes are low from an investor/audit or regulatory perspective.  As a result, the cost of deploying better methods exceeds the benefits provided by those methods.  However, if you are a community bank that is an SEC-filer with greater than $1 billion in assets and a CRE concentration, I would be very careful about putting all your eggs in the WARM basket for the following reasons:

  1. WARM is highly likely to get a COLD response from regulators. If you are a bank that meets the criteria above, just think about how much scrutiny you have received over the years in terms of the breadth of your risk management and stress testing.  You are being held to a much higher standard and are expected to have strong analytical command over your loan portfolio.  Stress testing has been a focus; banks are expected to stress their CRE loans in a forward-looking manner by using loan-level data as opposed to yesterday’s charge-offs.  Do you think your regulator will be comfortable if you have robust stress testing, which is not even required for purposes of financial reporting, coupled with the WARM method, which is ultra-light on quantitative analysis and is predominantly backward-looking?  Net net – the level of sophistication regulators will expect for CECL starts at the standard you are being held to for stress testing and other forms of risk management.
  1. WARM is way too dependent on qualitative factors. The dependency on Q-factors is already a huge problem.  In fact, the auditor’s auditor – the PCAOB – is already starting to crack down on auditors to limit the subjectivity caused by reliance on Q-factors within the loan loss reserve. The WARM method might increase the loan loss reserve under CECL because it takes an annual loss rate and multiplies it by a longer period than one year, but purely increasing the loan loss reserve was not the purpose of CECL.  The spirit of CECL was to make the loan loss reserve a forward-looking mechanism – just read Page One of the standard.  CECL was introduced in response to the 2008 Financial Crisis to force banks to increase reserves for the next recession.  As a result, the most important component is the forecast adjustment, because it’s here where the forward-looking component is captured in CECL.  The WARM method applies the forecast adjustment purely in a qualitative manner.  Don’t forget that FASB was initially reluctant to make the WARM method available under CECL for this very reason.  Reading between the lines, FASB ultimately relinquished this apprehension, primarily to appease the $200 million, non-PBE bank with a low risk profile, not the $1 billion, SEC filer bank with a CRE concentration.   Is that explicitly stated by FASB?  No.  Would you bet against this though?  I wouldn’t.
  1. WARM will cause many banks to become over-reserved under CECL. WARM is simple, but  simplicity comes at a cost.  The first two issues with WARM described above are focused on it being over-valued as a tool to achieve the first objective of CECL, which is compliance.  However, even if compliance is achieved, WARM’s analytical flaws create a situation in which the bank has a weak hand to support and defend its loan loss reserve.  These analytical flaws ignore loan-level information, which contain the true risk characteristics, while using historical loss information as the centerpiece, not the starting point.  That weak hand will cause banks to err on the conservative side to satisfy auditors and regulators, padding the reserve with excessive qualitative factors.  And this is in an environment where regulators, auditors, and bank investors are becoming increasingly concerned with the state of the future economy, so look for pressure on reserves to increase.  Achieving compliance at the cost of having to over-reserve is a disservice to shareholders.  CFOS, be warned.
  2. WARM breaks down in a recession. WARM is easier to implement today, when loans are comparable to those that generated low annual loss rates over the last five years.  However, if a recession becomes imminent, then those deploying the WARM method will have to choose between two very bad options:  (1) figure out how to increase the forecast adjustment, which has initially been estimated by throwing a dart at a board (see point number 2 above), or (2) utilize annual loss rates from the last recession, despite the fact that it was the most severe recession in over 60 years and today’s loans have different underwriting standards, were originated under different economic conditions, and are of a different mix than recession-era loans.  The net result will either be an uncomfortable group of auditors and regulators, an unnecessarily high reserve, or both.  WARM’s fatal flaw – the lack of utilization of loan-level data, will rise to the surface.  Nobody has a crystal ball, but to properly estimate how your loans will perform in the next recession from a directional and relative (versus other banks) perspective, you must perform forward-looking analysis on those loans.  The only way to even have a chance of doing this with any degree of accuracy is to utilize loan-level data.  This is also another reason why it is critical to sync stress testing with CECL to be able to tell your story to auditors, regulators, and investors.
  3. WARM offers ZERO strategic value. Banks that focus on WARM are squandering a massive opportunity to use CECL as a foundation to develop their data and analytics capabilities for the future.  The 2020 decade will be all about winning with data and analytics using techniques such as machine learning and artificial intelligence.  This is not just where the banking industry is headed, but where the world is headed.  Banks need to think about how to integrate CECL with stress testing, asset/liability modeling, loan pricing, strategic planning, and even M&A.  All these processes are driven by the same information – the individual loans and deposits (i.e. the zeros and ones) that construct the balance sheet of a community bank.   Failure to do so will mean the beginning of the end for many of these banks because they will be at a competitive disadvantage in the new world.  Because WARM fails to utilize loan-level information in a meaningful way, this will become impossible to do.  Yet the tragedy is that with the right analytics, banks can essentially create ‘alpha’ relative to their peers, maximizing risk-adjusted returns because they can deconstruct credit risk.  This allows them to price loans appropriately, avoid certain loans altogether, or pursue loans in bad times when the rest of the market has curtailed their lending.  I often hear that banks are planning on starting with the WARM method, but then gravitating to a better method over time.  While this makes sense, my concern is that the compliance mindset will infect the process, and banks will only utilize a better method because their auditor or regulator forces them to do so (see #1 above).  The result will simply be a bank going through the same motions except using a different method, as opposed to taking a strategic approach and demanding strategic value and insights that are utilized across the bank from their CECL systems.

I understand why banks are gravitating toward the WARM method.  It’s the simplest method, most like the existing method, and is not data intensive.  However, SEC filers that have gone down this path should treat WARM as a temporary solution that buys them time to develop better methods.  How much time remains to be seen, but I wouldn’t assume more than one to two years.  You don’t want to start from scratch when that time arrives, which is why banks should be planning and developing better methods to use over the longer term.

Adam Mustafa is CEO of Invictus Group.  Invictus consults with banks to develop turn-key and complete CECL solutions, as well as to assist banks who have purchased third-party software by providing them with external loan-level data contained within its proprietary BankGenome™ system, a methodology for reducing reliance on qualitative factors, and expertise on preferred methods such as the PD/LGD methodology.