×

For Fed stress tests, US banks form a study group

Executives at the biggest U.S. banks are sharing notes with each other before their next round of tests with federal regulators.

Banks are struggling to figure out what exactly the U.S. Federal Reserve is looking for when it conducts its annual "stress tests," which measure how banks will hold up during times of economic turmoil, bank executives, former Fed officials and consultants involved in the process told Reuters.

In gatherings organized by industry groups as well as more informal forums, executives say they have swapped tips about everything from how to best communicate their data regulators evidently appreciate robust summaries -- to how to project legal losses in a hypothetical downturn.

Read MoreOil's market impact: Barrel half empty, half full

The Federal Reserve deliberately keeps quiet about how it measures lenders' performance during downturns, to prevent banks from finding loopholes in the process that would allow them to take more risk, senior regulators have said publicly. It has given banks a little more information recently, but many executives still gripe about the tests.

Jason Hawkes | Iconica | Getty Images

"You put something in and one year it's okay and the next year they say 'no,' and you're scratching your head," said one bank executive. The executive, like others that spoke to Reuters, spoke about the stress tests on the condition of anonymity.

A few years ago, banks might have hesitated to share information with rivals about how they measure risk and how they communicate with the Federal Reserve. Their willingness to talk to competitors about these issues underscores just how exasperated they are with the process.

The Fed does not mind the information sharing, because the banks do not share confidential supervisory information and it is not collusion in any legal senseit does not result in price fixing or evidently hurt customers in any other way, the executive said. A Fed spokesman declined to comment.

Read More Rate rises spark volatility warnings

Regulators have multiple tools for keeping banks in check, including global capital rules known as "Basel III," which rule-makers world-wide have been crafting for years. But Basel III is viewed by the Fed as flawed, because it gives so much leeway to banks to measure how risky certain assets are.

The Federal Reserve has more control over the stress test process, which is part of its annual Comprehensive Capital Analysis and Review.

Many analysts believe the Fed will not hesitate to use stress tests to pressure banks to make their balance sheets safer.

"The Fed will likely use the CCAR process as a way to help drive capital requirements even higher for the largest U.S. banks," said Steven Chubak, a bank analyst at Nomura.

As part of the CCAR process, the Fed lists a series of bad scenarios for the financial sectorfor example, U.S. economic activity shrinking by 4.25 percent over nine future quarters and unemployment spiking to 11.25 percent, and banks estimate how much their assets will deteriorate in those scenarios.


They submit reports that run into the thousands of pages to regulators who then make their own estimates for how banks would perform using Federal Reserve models. If the Fed does not like what it sees, the bank cannot increase dividends to shareholders or buy back more stock.

A wide gap

What irks banks is that their stress test results and the Fed's are often far apart, and regulators give little information about how or why they disagree.

For instance, in 2014 Zions Bancorp, a Salt Lake City, Utah-based bank with $55 billion of assets, said its assessment showed it passed the stress tests: in a hypothetical downturn, it would have capital equal to about 5.9 percent of its assets, above the Fed's 5.0 percent threshold. But the Fed ruled Zions's capital ratio would fall to 3.6 percent in its scenario and issued a failing grade.

Executives also complain that strong capital levels are often not enough: well-capitalized banks can still be prohibited from paying out more dividends for "qualitative" reasons, such as a flaw in its capital planning process. Banks includingCitigroup, BB&T, and Ally Financial have failed the qualitative portion of the stress tests in recent years.

To combat the opacity, banks have met in forums hosted by the Clearing House, a trade group for big U.S. commercial and retail banks, the American Bankers' Association, and at other informal venues.

Read MoreWhere to put your cash? A house or a stock

Among the problems they have discussed is how to effectively write their reports, which must contain detailed information on the mathematical models they employ as well as a narrative piece that describes a lot of the methods that banks used to arrive at their conclusions. Executives say it is a challenge to produce a document that clearly and concisely explains everything to Fed modelers and examiners who may be unfamiliar with individual institutions.

One senior executive at a large U.S. bank involved in the stress-test process said one thing the Fed pays close attention to is the changes a bank made since the last submission. Instead of describing those changes in an appendix, the executive has learned the Fed preferred it to be addressed at the beginning of the document.

The executive added that given that the Fed has a limited period of time to review submissions, examiners have told other banks that they prefer to see a more detailed executive summary.

The Fed itself has taken some steps to help banks get a better sense of its expectations on the qualitative portion of the stress tests and how banks should organize their information—its October instructions for the 2015 stress test included a section on their preferential format.

The instructions also contained a "common themes" appendix that explained issues the Fed was looking closely at, such as if assumptions banks used to estimate losses are clearly articulated and how banks account for potential limitations and weaknesses in the models they use.

But some executives say that the more specific information about what it is looking for is not helpful if the Fed does not detail how it reached its conclusions.

"We've had a lot of conversations about methodologies and the approach to transparency in the CCAR process that have been more of a concern," said Chris Halmy, finance chief of Ally Financial, in an October interview. Ally has been seeking clarity "less about instructions and more about the feedback," Halmy said.

Another unresolved issue that banks have been powwowing about is how to estimate legal losses nine quarters into the future. In its common themes appendix, the Fed faulted many banks' projections and said they needed to take into account "possible claims of all types."

Legal losses are notoriously difficult to forecast. Without any more specifics about what the Fed wants banks to incorporate into their forecasts, it is a struggle to produce a result that will satisfy regulators, said one consultant that does stress-testing work for large banks.

"No one really knows what the Fed is looking for" with respect to potential legal losses, the consultant said. "There may be some objections to capital plans next year as a result."