Timothy Zimmerman is no longer fearful of a new accounting approach to loan losses.
Zimmerman, the CEO of the $488 million-asset Standard Financial in Monroeville, Pa., and chairman-elect of the Independent Community Bankers of America, was among the bankers who lobbied the Financial Accounting Standards Board hard to make changes to the proposed Current Expected Credit Loss accounting standard.
The bankers worried that CECL, which will replace a system that relies on historical data with a forward-looking one that will require lenders to forecast credit losses when a loan is booked, would lead to a spike in loan losses and pressure to buy costly modeling software.
That was 18 months ago.
Last spring, however, FASB agreed to a number of revisions that put most bankers’ fears to rest. Most significantly, the amended version made it clear that financial institutions would be given flexibility to choose whatever tools they felt they needed to determine their allowances. In other words, no costly modeling mandate.
Zimmerman, who once labeled CECL a “dangerous proposal,” now believes that banks can benefit from the new standard. In fact, he is pressing bankers to push forward with efforts to comply with CECL in the wake of a study by Jack Henry & Associates that found that nearly two-thirds of banks and credit unions haven’t started the process.
“You have to keep it in perspective,” Zimmerman said in a recent interview. “Everybody’s doing a loan-loss calculation today. It isn’t like this is a new thing that now we’re going to do something no one has ever done before.”
Here is a transcript of the interview, edited for length and clarity.
How do you think banks should approach CECL implementation?
For community banks, my sense is that the implementation process is getting started now. The two key groups are the accountants and the data-processing providers. Banks know they’re going to need some vintage-year information for disclosures in the footnotes, and to some degree to help calculate what their current expected credit losses will be. As you might imagine, with 6,000 community banks and a lot of independent accounting firms, we’re starting to get different answers about how it might all work.
I think banks got the message that they shouldn’t start with buying a model. … I think they should determine what they think makes sense for their bank, coordinate that with their accountants particularly, because they’re the ones that are going to sign off on this eventually. Then, if you both decide that your unique situation is so complex that you need a model, you can start talking to the companies that provide models and things like that.
When do you expect the pace of implementation to pick up?
It will pick up as we get closer to 2020, and it’s going to get a lot more serious. You’re going to see banks do some sample disclosures and I think you’ll see some cases where companies will elect early adoption — only after their accountants sign off. To me the key at this point is making sure each bank is in alignment with their accountants on how to calculate it and how to disclose it.
As long as regulators are seeing well-developed calculations and the accountants are signing off, I don’t think you’re going to get a lot of pushback.
Do you get the sense that data requirements are being perceived as more onerous or complex than anticipated?
The answer to that goes down to the individual bank level. My bank uses a third-party core processor. There’s a real strong report writer that is included and we have people who develop customized reports all the time. When we were working on CECL last year, I went to our financial staff and said, “I want you to see if you can do this. Here’s what the disclosure looks like. See if you can do it.”
They went into the system and basically came up with what our vintage-year disclosure would look like. They said they can do it and that it’s not going to be too bad. It seems like if you have a system that has a good report writer and your staff knows how to use it, the data is sitting in your system, you can pull it out. I think in cases where banks are on different systems and maybe they haven’t done a report like that, it might be a little bit more difficult.
The level of complexity obviously changes based on the bank’s business plan and scale. Clearly there are cases where larger, more sophisticated banks with broader-based portfolios may need a model or maybe they have a model already and will have to modify the model they have.
Are you saying CECL may not be as bad as some had feared?
Everybody is doing a loan-loss calculation today. It isn’t like this is a new thing that no one has ever done before. You have to keep it in perspective.
I think the big concerns about reserves being 50% higher have been mostly allayed. Everybody breathed a big sigh of relief once FASB made some of those changes to the standard before they published it last June.
Would it surprise you if any small banks were early adopters?
It would surprise me probably if a lot of them did, but it wouldn’t surprise me if some of them did.
One thing that got lost in the whole discussion is that a lot of community banks have wanted to look forward. If I’m an agriculture lender in the Midwest, I know what’s out there [and] what the trends are for commodity prices and things like that. If I’m concerned about what might happen in the next couple of years, I wasn’t supposed to consider that under the old rules. It was a historical-based, incurred loss model. Under this new standard, you’re allowed to provide for that.
If you have granular information in your market about what’s going to affect your bank, you can prepare for that. I think in some cases, people are anxious to go ahead and do that.
There has been very little discussion about the benefits that might flow from CECL.
Right. All the discussion has gone in the other direction. There is high interest and there is still concern about it. But if you do it correctly, you’re going to provide forward-looking calculations that you weren’t able to before. There’s a plus here.
Is Standard planning to report under CECL in 2019?
No, we start in 2020.
That’s still several years away.
It is, but there’s a danger in that. When I’m going around the country representing ICBA, the message I’m delivering is don’t wait. You should be having these conversations now. It’s probably not a good thing if you’re completely relaxed and you figure it’s out over the horizon.
If you haven’t had a conversation with your accountants and your data processing company, and potentially with your regulator, you should. You should take the initiative and get that conversation going to make sure everybody is on the same page. You don’t want to get a big surprise.
Have the accountants been sharing this message with clients?
I get the sense, especially with the regional firms, that they have groups that are actively working on this. They’re getting inquiries from their clients. They’re making contacts, too. I think the better ones are because, as we know, time flies. It’s going to be 2020 before we turn around. I think there’s a dialogue, but it’s happening quietly.
In the beginning, based on the rhetoric, people who sell models figured this was going to be very lucrative. Now it’s crystal clear that you don’t have to do that in most cases. So if you were ramping up to sell models, you’re going to be disappointed. There will be models bought and run and they’ll do a nice job [but] for the vast majority of community banks, I don’t think that’s going to be the answer.
Is CECL one of the major challenges facing banks over the few years?
I don’t think it’s going to be the sea change we heard about originally. It’s not that you can snap your fingers and just do it, either. You have to think through what you want to do, starting with what’s required. Then you have to figure out how your business plan stacks up in level of complexity. How many different types of loans am I making? Do I have homogeneous groups?
The other question is how far out in my market can I realistically look and have a good idea of what’s going to happen. Remember, the standard says you only go as far forward out as you can reasonably predict, then you revert back to the historical record.