MST 2018 Annual Lender Survey Tracks Progress as Financial Institutions Prepare for CECL Accounting Standard

Again this year MST used its annual survey to focus on the new ASU 2016-13 accounting standard, CECL (current expected credit losses), set for implementation in 2020 and 2021, which will dramatically change how banks and credit unions determine their reserves. 

According to the survey, almost all institutions have started preparing for CECL, though none has finished. Few, especially smaller institutions, are as far along as they should be, note CECL experts. 

“That’s not far enough along for many institutions,” MST Senior Advisor Paula King said. “If you are an SEC filer you will be estimating under CECL on March 31, 2020. You’ll want to start testing your models by the first quarter of 2019, which gives you less than a year to secure external help if you need it, budget for your transition and draft your roadmap, which will include gathering and assessing your data and re-segmenting your pools.” 

Again this year, the MST annual lender survey tallied responses from a wide range of institutions. Participating lenders have locations, or at least are doing business, in all 50 states. Half the respondents are in institutions of under $1 billion in assets and another 31 percent in institutions of between $1 and $5 billion in assets, in contrast to the firm’s 2017 survey, where the majority of respondents indicated assets in the $1 to $5 billion range. Slightly more than 14 percent this year fall into the $5 to $20 billion range, leaving just a little under 5 percent reporting assets of more than $20 billion. 

Getting started on CECL

When asked to “describe your progress in preparation for and transition to CECL,” more than half the respondents, 55 percent, indicated they were “having internal discussion/meetings.” They are assembling CECL transition teams. The teams almost universally include members from Accounting and Credit. An additional 58 percent include an associate from Finance, and almost half have a member from Lending.  

“We’ve included at least one person from every department … that will be affected by CECL,” one respondent wrote. “This gives us different perspectives and allows us to see from all angles.” 

Still, fewer than 40 percent of the respondents said they had a chief officer on their CECL team, and only 20 percent included a member of the board of directors. 

The 2018 survey does indicate some progress. Twenty-eight percent of all respondents were evaluating prospective third-party partners to help with their transition, and 31 percent had already engaged third-party support. Twenty-seven percent had reached the point of testing prospective CECL methodologies. 

Download the full survey results whitepaper.

Data for CECL

Much of the consternation about complying with CECL has revolved around data. Do institutions have enough data? Do they have clean data? Do they have the right kind of the right amount of clean data? In 2017, only 20 percent of MST survey respondents reported believing they had sufficient data to estimate under CECL. Thirty-two percent of this year’s responding lenders report being confident they have sufficient data. And while more than 70 percent are still compiling data, 15 percent have already determined they will need more data than they can assemble internally. 

In answering questions about how far back institutions will look to gather data, the largest percentage, 38.2, reported they will use a five- to seven-year look-back period. 

That might be insufficient, MST’s King warned. 

“Ideally, we’d like to see enough data to represent a full economic cycle,” the period between economic peaks. The current cycle, according to the National Bureau of Economic Research, began with the fourth quarter of 2007. Only 38 percent of respondents plan to use a look-back period of eight years or longer. 

Pooling

Almost all reporting institutions, 85 percent, do at least some pooling by product type. Fifty-seven percent segment loans by risk ratings, and another 52 percent by collateral type. CECL will require more granular segmentation, and almost two-thirds of survey respondents said CECL will cause them to restructure their pools. 

“Institutions should consider segmenting their portfolios as much as possible while maintaining statistical significance,” noted MST Senior Advisor Chuck Nwokocha, defining “statistical significance” as a combination of size in terms of the concentration of the portfolio as a percentage and number of loans. 

“This is a chance to capitalize on the CECL implementation to structure new pools that appropriately segregate the risks,” pointed out MST Senior Advisor Shane Williams. He added that CECL presents “an opportunity to not only meet current needs but prepare for future growth and product changes.” 

Methodologies being considered for CECL

The end game of CECL preparation plan is to identify a methodology or methodologies that are appropriate for the lender and its loan portfolio. Most survey respondents, 81 percent, have used a historical loss approach to estimate at least portions of their loans under the incurred loss accounting standard, and interestingly, more than 58 percent say they are considering historical loss as a methodology under CECL. That option received the most votes with vintage coming in second with 47 percent considering its use. 

“You can start with historical data,” King pointed out, “but you have to project data over the life of the loan, so there’s a lot more involved than historical performance, including applying qualitative factors and forecasting adjustments.” 

“Choice is often driven by a data gap analysis,” Nwokocha explained. “The analysis allows us to identify currently available data that should be maintained and consider whether any additional data has to be collected or maintained to implement CECL. 

“Another driver is the appetite of the institution,” he said, “whether they want to do only what is minimally necessary to identify a working methodology or are looking to fully delve into what would be best for their institution. Appetite is influenced by capacity and resources.” 

Impact on Allowance

Only 45 percent of survey respondents said they think CECL will increase their reserves. That’s an increasingly optimistic view. In 2016, 93 percent of respondents said they thought CECL would result in higher reserves; in 2017, 66 percent said their reserves will increase under CECL. 

Williams questions the optimism, pointing out that at least two factors should add to an institution’s reserves. 

“You are going to add investments to the reserve, which you did not include before,” he said. “You are estimating loss over the life of the loan versus a set emergence period; by definition that new total has to be at least as much as before and should be a higher loss estimate.” 

Thwarting progress

What’s giving lenders pause in their transition? For some, the job is beyond the skills or time they have to dedicate to it. 

“We suggest … an outside vendor,” one lender wrote. “… it could get complicated doing the math in Excel and (Excel) leaves room for user errors.” 

Others would like more prescriptive direction from the FASB or their regulators. 

“(We need) … more scenario based interaction tools. How to do the methodologies, compare and contrast, how to explain what we’ve learned, to be confident to discuss with examiners. More examples to help develop our strategies and implementation bank-wide.”