Following last December's proposals regarding the residual pieces of securitizations, federal regulators - including the Federal Deposit Insurance Corp. and the Office of the Comptroller of the Currency - are poised to release another, more specific set of guidelines, industry sources say.
The anticipated guidelines - which might come as early as February, a source said - will likely specify the percentage or amount of residuals a bank is allowed to include on its books when calculating assets for core capital purposes.
"There's been a discussion from the regulators lately, particularly coming from the FDIC, about setting an absolute limit, and I think that's what you're going to see next," said an attorney who specializes in regulatory issues concerning securitization. "I don't know exactly where it's going to turn out, but I think there will be, unquestionably, further action soon."
The federal regulators recent concern with residuals stems from the losses associated with recent bank failures, most notably First National Bank of Keystone, which could cost the FDIC as much as $800 million, $300 million of which is related to residuals.
In addition to issues with residuals, officials at Keystone were charged with fraud and illegal disposal of bank documents.
"Keystone had the problem anyway," the lawyer said. "Putting aside shoveling files into a big ditch in the back of somebody's house, you have Goleta National Bank out in California, and others as well. And I've heard the regulators say at various times that they believe there are another 10 to 12 or more institutions out there with the same problems: Where assets are on the books at a certain value which is probably at best ephemeral."
Sources say the problem with securitization residuals is that they are created based on assumptions.
"It's a function of what discount rates you use and other things," the lawyer said. "And if you use improper assumptions, either too conservative, or too liberal, or too aggressive, on your choice of discount rate, etc., your numbers are going to be very different."
The lawyer guesses the next set of guidelines will correspond to legislation concerning similar intangible assets, namely servicing rights.
"When you talk about purchased mortgage servicing rights, and purchased credit card relationships, the regulations have specific limits on how much you can count in your core capital," the lawyer said.
"I don't know if they're thinking about this the same way," the lawyer said. "But it seems to me they have a precedent, where they have built into the regulations a specific limit on this type of asset, so I would be surprised if they don't take on this same type of approach [with residuals].
Problems with the FDIC
According to a consultant in this field, there is currently a great deal of tension between the regulators on the residual issue "largely because the FDIC is just bananas about the losses that they've taken in the failures of the banks in subprime securitization."
"The unfortunate part of the FDIC is that the people that are working on it don't know anything about securitization," the source added.
The source contends that the FDIC is wrongly associating some of its recent losses to residuals.
"This became a really hot topic after the Keystone failure - after they realized how much money they were about to lose - although frankly it's a chicken and egg thing," the source said. "They're going to lose on the residuals there, but I don't know if they're going to lose because they are residuals. I think they're going to lose because of the fraud.
"The FDIC is estimating its losses at about $800 million, which would be the combined value of the missing $500 million and probably the current book value of the residuals," the source said. "But they haven't figured out yet that the residuals loss comes from the fraud, so they're after all residuals. And they see all the residuals as being [interest-only] strips which is wrong, too.
"Once you're into subprime, there's a heavy portion, or a significant portion, that is overcollateralization, which is different than the IO strip," the consultant said. "Again, near as I can tell they haven't distinguished."