No Reason to Wait Until 2020 - Get CECL Started Now

In June 2016, the Financial Accounting Standards Board (FASB) issued a new accounting standard to replace the “incurred loss” impairment methodology with the Current Expected Credit Loss (CECL) model, marking a significant shift in the way credit losses on many financial assets—especially loans—are recorded.

The new accounting standard applies to all banks, savings associations, credit unions, and financial institution holding companies—regardless of asset size. It is effective beginning in 2020 for public business entities required to file with the US Securities and Exchange Commission (SEC), and in 2021 for all other public and non-public business organizations. Early application is permitted for all institutions starting January 1, 2019.

What changed?

Instead of looking at what they incurred as a loss, financial institutions will need to predict what they will lose over the contractual life of the financial asset. And while the new standard does not go into effect until 2020 or 2021, the new calculation methodology requires analysis of several years of historical data—which means it’s time to act now.

Critics of the new standard say in the short term CECL could pressure banks’ profitability and increase the volatility of Allowance for Loan and Lease Losses (ALLL). However, proponents believe it will allow financial institutions to accurately project losses from their core businesses.

Transitioning to CECL is a highly complex change management initiative that has far-reaching implications beyond the accounting department. The basis of a successful CECL transition will require much more than coming up with a CECL-compliant estimate. Success will require taking a holistic view of the end-to-end process for setting aside reserves and leveraging the right tools to get the job done efficiently and effectively. Implementation will be particularly challenging for smaller institutions that lack the predictive modelling capabilities that their larger counterparts have.

Viewing losses from a new perspective

Ultimately, CECL implementation requires a philosophical change in mindset: from a backward-looking to a forward-looking approach in setting allowances for credit losses. It is not just a method of increasing provisions against the loan portfolio, it creates an opportunity to gain a better understanding of the loan portfolio.

We realize that CECL is one of the biggest changes to accounting rules in decades. And that means making painful changes within the organization to meet the requirement. But the magnitude of the change is actually another good reason to get the hard work done now and get a head start with calculating provision for loan portfolio.

You may see reasons blocking your organization from moving forward with CECL, but at closer look, they may be unfounded.

Why? Let us explain.

Reason #1: Missing data and deciding which external data to use

Small banks reporting to the SEC, such as community banks, are having a particularly tough time with CECL, since many have grown through acquisition and don’t have clean data sets. Or they don’t have enough data. 

Recognizing this problem, regulators have encouraged these banks to buy data sets from external data providers to supplement their data, but banks are confused about whether they should use state and local data, or national data.

The benefits and risks of external data

Although good quality data is available, the problem with national data is that the loan characteristics of portfolios at community banks may differ widely with averages from across the country.  We recommend companies use both state/local data and national data. This will give them a richer data set to choose from.

Here we would caution that banks should not use external data sets blindly. They must analyze how the external data reflects into their portfolio. If a bank has a portfolio of risky loans, then that won’t be reflected in external data sets. In that case, using the external data would result in misleading CECL predictions.

Financial institutions are also worried about how many years of data they need. In our opinion, five to seven years of data would be ideal. But even if a bank has less than five years of data, it should get started with CECL now. Although reporting is not mandatory until 2020, it’s better to have something in place in advance. By then, the company will have a few more years of data at hand and be ready to go.

Here our advice is simple: Just do it. Document any assumptions you’re making about your data, and note any gaps you have. You can come back later and work through these. “Keep it simple” is a good motto until an organization has gained experience with CECL reporting.

Reason #2: Blind spots, method choices and organizational changes?

At a recent conference on CECL, the audience was asking about moving from data collection to the actual use of a chosen CECL modeling method or methods.

Some financial institutions have settled their questions about internal and external data but seemed stymied by worries that the results of their analyses will be inaccurate.

They were asking themselves, “Did we factor in everything we should? How will we know?”

We would like to calm these worries by saying that projections may indeed be slightly off, particularly at the beginning. 

Getting an accurate reflection of loan losses

No matter how well financial institutions arrange their data, or whether they have a year of history or three years of history, companies will only be able to see the quality of the results after they have done multiple runs in their model. Only then will they be able to see if the model calculates the expected credit losses accurately.

In our opinion, it will take at least six months of parallel running of the ALLL system and CECL to see how the data compares from the two approaches and understand any anomalies.

That brings us to another point: selecting the modeling methodology.

Selecting the modeling methodology

Compared to the other two choices – a Probability of Default method and a Discounted Cash Flow method - the Historical Loss Rate method requires far less computational power.

The Historical Loss Rate method works by applying the data from historical losses directly to the future and then make gradual adjustments to reflect your own loan performance.

It is up to each financial institution to decide which method to apply and whether it wants to mix and match the methods for different loan portfolios. As long as the details are documented, and the results are supported and reasonable, the regulators have stated they will accept it.

Training and governance

Another point is training and working across organizational divides. It is a complex task to project loan volumes and yields on a quarterly basis and to regularly incorporate new data from new loans. It cannot be done in isolation within the organization.

From a governance perspective, it’s critical that the entire organization is involved, including loan officers, finance, IT, the risk and credit departments, accounting and the CFO.

Banks will also need training, process automation and awareness of what CECL entails.

For us, this is another strong argument for starting now: Often the most difficult changes are the ones that involve asking people to work differently.

Reason #3: Questions of segmentation? 

When it comes to segmentation, we feel your pain. This is a step in the process that looks easy, but it’s not.

In the past, segmentation of assets was often based on the business units. In other words, assets were classified as real estate if they were held in the real estate division.  Now, classification will be by loan characteristics – e.g. the behavior of a loan. This includes collateral type, geographic location, indirect vs. direct loans, size of the loan, and measures of consumer credit risk.

For CECL, we recommend organizations do not classify their loan assets too granularly. That only adds complexity and it’s not necessary in the beginning.

We believe lenders will have to come back to the subject of asset segmentation. Segmentations are never set in stone. But now is the time to move forward, not to debate the intricacies of segmentation.

For us, CECL implementation is not just an exercise to meet regulatory requirements. We strongly believe that early adopters will reap additional benefits, such as smarter decisions on capital allocation.

And, over time, banks will be able to re-align their loan portfolios and optimize their performance.