Accounting pop quiz: What's worse — "significant" credit problems or "not insignificant" credit problems? And which would require higher loss reserves?

It sounds like a fine distinction, and a nonaccountant might have to squint hard to see why it matters. But it has become a key point in an esoteric — though potentially important — dispute over how to reserve for potential loan losses, and some think it could significantly raise bank M&A costs if it is not resolved.

The Financial Accounting Standards Board, a nonprofit that writes accounting guidelines, has proposed new rules for the treatment of loan losses that some say would unfairly require banks to hold excess reserves on loans and securities they acquire.

These critics say the rule — Proposed Accounting Standards Update Subtopic 825-15, for those keeping track — would in effect force banks to account for expected losses in acquired loans twice: once when moving the loan to their balance sheet, and a second time when they reserve against it.

Critics call this "double counting" of loan losses, arguing that the expected loss is already baked into the fair value of the loan when it goes on banks' books.

Mike Gullette, vice president of accounting and financial management for the American Bankers Association, called the proposed rule a "tax on mergers" that would require the buyer to hold as much as 5% in excess capital.

Others say that these objections are, in short, completely bogus, and that "double counting" is not real. The critics misunderstand how the rule will work, they say.

FASB itself is divided. One board member has threatened to vote against the entire new loss model if double counting is not fixed, while others deny there is such a thing as double counting at all.

This disagreement bubbled over at a meeting last month between FASB's seven-member board and its Investment Advisory Committee, a nine-member advisory panel that has raised concerns about double counting in the past.

A debate about modifying the rule turned into one about whether double counting was real, then into a discussion of hypothetical examples — interrupted by frequent cross-talk — about how the proposal would play out.

"We've had this debate probably 25 times, and I still get confused," one board member said at the meeting, according to a recording. "Without seeing [examples] in front of us, we're just going to continue to talk over one another."

FASB is now preparing these real-world examples for its investment committee, in the hopes of clearing up the fog around the issue. The final draft of the rule is expected at the end of the year, and the deadline for implementation would likely be a few years later.

"The Board believes that its recent tentative decisions — and the drafting process of the final standard — will address" the perceived problem with double counting, FASB wrote in an email to American Banker.

The dispute highlights the complexity — and, for nonaccountants, the opacity — of the debate about accounting for bad loans. The banking industry and FASB have been skirmishing over this issue since at least 2012, when FASB unveiled a draft of a new loan-loss model to replace the current accounting treatment.

Most agree that the existing rules should change. The current incurred-loss model, under which banks increase reserves only when losses begin, leads them to set aside too little in good times, then massively ramp up reserves in a downturn. This prevents banks from adding to their reserves until too late, and causes peaks and valleys in earnings and liquidity.

But not everybody agrees that FASB's proposed alternative would be an improvement. The so-called Current Expected Credit Losses, or CECL ("Cecil"), model would require banks to reserve against estimated losses that have not yet occurred.

It's likely that this change will result in higher reserves, but estimates of how much higher vary. Comptroller of the Currency Thomas Curry said in 2013 that reserves could rise 30% to 50% on average.

Other estimates range from around 10% on the low end to double current reserve levels.

That prospect has, predictably, alarmed the banking industry. The first draft of the rule got 355 comment letters, many of which said the loss model would require excessive reserves and that the method of forecasting losses is too complex.

FASB has made many changes to the proposed rule in response to these comments. But on the issue of acquired loans and the alleged double counting, some in the industry feel FASB has dragged its feet.

Gullette, for instance, said the board has "not been very straightforward" about fixing the problem.

Others besides banking lobbyists have raised concerns about double counting; groups including the Securities Industry and Financial Markets Association (known as SIFMA), did so in their comment letters.

Even the American Institute of Certified Public Accountants, a group that can hardly be said to misunderstand accounting, wrote in 2013 that the FASB's proposal "includes a fundamental flaw because it double counts expected losses."

FASB Investment Advisory Committee member Joe Stieven, of the investment firm Stieven Capital Advisors, has been the most outspoken critic of the board's treatment of purchased loans, and has grown frustrated that the board has not resolved the committee's concerns.

"We raised this as an important issue in June of 2013, and we raised it again at the end of last year, and we raised it again at the beginning of this year, and we felt like we really didn't get much response," he said at the meeting on May 11. "I think there is some response now, but I'm still not clear on whether that's good enough."

Stieven and the ABA both want FASB to treat all purchased loans, impaired and unimpaired, the same way, which they say would eliminate the double-counting problem.

In April, FASB's board members voted five to two against doing this. Instead, the board sought to resolve the problem through a subtle, and highly accountant-like, alteration: it changed the threshold for acquired loans that banks will not have to book reserves on from those with "significant" deterioration to those with "not insignificant" deterioration.

The change would exempt slightly more loans from the alleged double counting. For Stieven, though, it's not enough. He called it a "compromise" that failed to eliminate the double-counting "loophole."

"It's better than it was, but I just don't understand why we have to have a loophole there at all," he said at the May 11 meeting.

Even for observers without a dog in the double-counting fight, the controversy adds to the frustration over FASB's rulemaking process. While FASB has responded to many of the banking industry's concerns, its protracted deliberations have left banks in a three-year limbo on very important issues, said Tim McPeak, senior risk management consultant for Sageworks, which advises banks on accounting issues.

"We're three years down the road and they're not giving us a ton of insight," he said. "We're all on the edge of our seat waiting to see what they're going to tell us."

Of course, writing a massive and complex set of rules while trying to address the concerns of hundreds of constituents should not be quick and easy. Stieven says while he disagrees with FASB on double counting, he has "a high degree of respect for the complexity of the issues they have to deal with."

Banks can do little but wait and see how — or if — the so-called double counting issue gets resolved in the final rule. There is not much they can do to prepare, beyond making sure they will have the right information at hand to comply with the final rule, said McPeak.

"The common question banks ask is, 'What can we do today?'" he said. "You should just try to capture as much data about your loan portfolio as you can and try to store it as well as possible."

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.