FDIC's Hoenig Proposes 'Full Scope' Big Bank Exams
Two experts, Tom Hoenig and John Dugan, have very different takes on how policymakers should end too big to fail. But figuring out which one is right is daunting.January 30
FDIC board member Thomas Hoenig proposes corresponding a bank's required capital minimum to its Camels rating and subjecting big banks to more detailed exams.November 30
Tom Hoenig, a FDIC board member, reiterated his doubts that too big to fail has been eliminated, and offered his prescriptions for strengthening the system.November 15
WASHINGTON — Thomas Hoenig, already well known for demanding that Wall Street banks should get back to basics, is proposing the same mantra for their examiners.
Just as Hoenig advocates for more tangible measures of banks' capital and liquidity, the Federal Deposit Insurance Corp.'s vice chairman has stepped up calls for examiners to do a better job verifying a big bank's safety and soundness through a more "systematic review" of its nuts and bolts.
The scope of big-bank exams is inevitably restricted by the breadth of what goes on inside a giant company as compared to the regulators' limited exam resources. But Hoenig says relying on a bank's internal findings, using data modeling and stress tests and reviewing targeted risk areas — now key components of large-bank exams — are no replacement for a thorough look at its operations.
"If you're going to know the condition of an institution, I think you're going to have to examine it," Hoenig said in a recent interview with American Banker. "Otherwise, you are kind of working on the honor system. I'm OK with that to a point. But, 'trust but verify' is the first rule. We need to do more verifying as supervisory authorities."
Hoenig says targeted reviews, which allow examiners flexibility to focus on certain risk areas they view as most relevant for in-depth monitoring, and other elements of today's big-bank exams, should be retained. But he says they should be de-emphasized in favor of a meticulous look at a bank's operations based on a statistically valid sample. While it would not be possible to cover as much material relative to the size of the bank as examiners see at smaller institutions, Hoenig says, the idea is to move large-bank reviews "in the direction of a community bank exam."
"You can't necessarily anticipate where the risks are emerging and therefore it becomes more difficult for a targeted exam process to really work, because then it almost becomes random," he said. "What I'm advocating is you go and reassert what I call a 'systematic review' and that will give you a fundamental understanding of the risk."
In addition to Hoenig's well-established critiques of the impact big banks have on the safety net and of the multijurisdiction process of strengthening bank capital requirements, his recent speeches have also included support for wide-scope exams.
Targeted exams "can be useful, but they are limited in scope and have been adopted because the largest firms are judged simply too large and complex for full-scope examinations," he said in public remarks on Nov. 30. "However, full exams are doable. Statisticians, for example, have long been designing sampling methodologies for auditing and examining large bank asset portfolios and other operations, providing reliable estimates of their condition, and at an affordable cost."
In the interview, Hoenig said his plan would call for the creation of a roving team of examiners to go from bank to bank doing systematic reviews — in addition to the embedded teams of resident examiners. The team would be able to help identify what the embedded examiners should analyze more closely in their targeted reviews.
"Start with a systematic review — that is, you do a general review to a certain degree — and from that you have a better opportunity to see where risk might be larger than what you anticipated," he said. "Then you would do a deeper, targeted review."
Essentially, examiners would spend more time studying individual files to verify the quality of a bank's internal reports about its risk management capability.
"If the reports show daily limits, you would test to see that those daily limits weren't exceeded," Hoenig said. "If they say, 'Here are our underwriting standards,' you could go in and find that … actually, no, they're not ensuring that the collateral is of a certain type. You want to think about this in a systematic fashion."
Hoenig's idea about the scope of exams is just beginning to spark chatter among D.C. policy watchers, with some supportive and others skeptical.
"It's the old-school bank exam. Just like you saw in 'It's a Wonderful Life,' the examiner puts his eyeshades on, sits down at a desk and starts counting everything," said Karen Shaw Petrou, managing partner of Federal Financial Analytics.
But Petrou questioned whether more comprehensive exams are the answer to keeping big banks in check. She said some of the supervisory failings leading up to the crisis — such as chief regulators not acting on problems flagged by examiners and leadership of firms themselves not having a good handle on risks — may not necessarily have been stopped with more comprehensive exams.
"Why would" reviewing every loan at Washington Mutual "have been any better than the examiners who should have been able to see the giant pile of turd in the middle of the room?" Petrou said. "You don't have to count each turd to know there's a really big pile."
Others said the problems regulators missed or downplayed leading up to the crisis revealed a flaw in exam methodology.
"It's like taking the computer codes that run the car, but never looking at the tires or the fluid levels or the gaskets to see if they were deteriorating," said Arthur Wilmarth, a banking law professor at George Washington University.
Frank Partnoy, a law and finance professor at the University of San Diego, said more comprehensive exams should at least be tried, and would prove valuable even if they failed. "Why not attempt full-scope exams first, and see if they are viable?" he said. "If they cannot be done, that is an important message for regulators and the public: it suggests that banks are too complex."
But some former regulators say the targeted exam model and examiners' use of data did not come about in a vacuum, and the charge for the agencies now may be to further improve on those elements.
The crisis "may be an indication that the targeting needed to be more precise," said Ralph Sharpe, a partner at Venable and a former director of enforcement at the Office of the Comptroller of the Currency. "It's important that the exam is done in a way that it reveals a bank's underlying problems. If they are revealed, then by all means it should go deeper and broader in scope. … But you can't look at everything every time you go into a huge bank."
Douglas Roeder, who was the OCC's senior deputy comptroller for large-bank supervision and is now managing director at PricewaterhouseCoopers, said the biggest banks are now subject to quarterly updates of supervisory ratings, known as Camels, but also ongoing monitoring from OCC resident teams who design exams based on their knowledge of specific institutions.
He noted that examiners were given more leeway after the thrift crisis of the '80s and '90s "to build their own supervision plan rather than going A to Z every year and trying to figure out how to get that done with limited resources and increasingly larger and larger institutions."
"One thing a full-scope review should not do is overly constrain the exam team from going to look at something at the right point in time it needs to be looked at," Roeder said. "You don't want to strangle the examination process where you've got all this stuff that has to be done and there is an emerging risk in the institution … and you're so constrained resourcewise that you can't stop and go look at that. You've got to have some flexibility."
Yet targeted reviews and a full sweep of an institution may both be attainable, he added. "If full scope requires the examiner to address all the key elements, but then supplement it with the discretionary risk-based process, that might be a way to do it."
Other former senior officials said the agencies have greatly improved their ability since the crisis to use automated data to uncover problems more quickly than would be discovered by reviewing a portfolio manually.
"This is clearly evolving," said Christopher Spoth, who helped oversee the FDIC's supervision program and is now a director at Deloitte. "Looking at the experience of the crisis, such as high losses on residential lending, which was unexpected, supervisory staff at the agencies want to be able to identify that faster while there's time to mitigate an emerging risk."
Hoenig agrees that data modeling can be useful, and he is sensitive to concerns about the resources needed to conduct fuller-scope exams. He said the more abstruse elements of examining big banks, including Federal Reserve Board stress tests, should be included as part of a holistic process.
"I have a lot of respect for the idea that we have these models. I understand models. I think they have value as a check … but models don't acknowledge friction," he said. "Models assume everything works the same way every time, and they don't have errors in judgment. You need to go in and look at the institution, not just the model and the output from the model. While that's important, it's not sufficient.
"I don't think this is revolutionary. I think this is a reasonable step to have a better, systematic understanding of these very complicated and important institutions in terms of their impact on the economic system. I don't think you can model that. For people who think you can, God bless them. But this isn't physics. This is human judgment, and human choices driven by incentives of the safety net."
Hoenig, who says other regulators have been "receptive" to his idea in informal discussions, said concerns about the cost of expanding the scope of exams could be considered in determining the size of the statistical sample. More demand for exam resources theoretically affects bank resources, since the OCC charges banks assessments to fund its supervisory activities. "I see that as a potential issue. But until we do the analysis, I don't know that it's an issue," he said. "We estimate the GDP. … You can get a reasonable estimate of the population by sample. Let's bring in some statisticians to help design the sample size and go from there.
"How much risk we're willing to accept in the systematic review — that is, how big the variance is we're willing to accept — is a decision we have to make given the cost of doing so. But I'm not sure that just having the bank's management present you their deck and their analysis is a better use of resources than what I'm suggesting."