WASHINGTON — The Federal Reserve's annual stress testing regime violates federal administrative law, a coalition of banking groups said in a white paper released Thursday, likely previewing a court battle that could have far-reaching consequences for the program.

The Committee on Capital Markets Regulation, which includes big banks and several industry trade groups as its members, argues that because the results of the Fed's Comprehensive Capital Analysis and Review stress tests can form the basis of an order preventing banks from issuing dividends, it acts "as a de facto binding capital constraint on banks." Any such constraint would be subject to notice-and-comment rules under the Administrative Procedure Act, the paper says.

"The committee … recommends that the Fed follow the APA's procedural requirements in developing key aspects of its stress tests because doing so would result in better public policy outcomes and reduce the threat of legal challenge to the Fed's actions," the paper said.

The group's members include representatives of JPMorgan Chase, State Street, Wells Fargo, Citigroup and Deutsche Bank as well as groups like the Clearing House Association, the Securities Industry and Financial Markets Association and the Independent Community Bankers of America.

The Fed's stress tests hinge on the annual issuance of hypothetical dire economic scenarios through which each systemically important bank must hold minimum levels of capital for nine consecutive quarters. Those stress scenarios are presented by the Fed roughly six months before the stress tests are conducted, and outline baseline, moderate and severe scenarios. The Fed then runs not only its own stress test of the banks, but analyzes the banks' own internal stress tests for quality and accuracy.

The paper argues that the assumptions that underline the stress scenarios are published without any public input, and have "vastly differed from reality." The paper also argued that the scenarios "fail to consider the most salient economic and financial market risks" that institutions face.

What is more, the models upon which the Fed's internal stress tests are based have never been disclosed, and so despite "limited" efforts to disclose certain aspects of the models, the institutions subject to the tests lack a detailed understanding of what exactly they are being subjected to.

"In our view, both the assumptions and the models should be categorized as rules," the paper said. "There is no apparent reason why public participation in the development of the assumptions would be impractical, unnecessary or contrary to the public interest. With regards to the models, we believe that the exceptions for guidance and interpretative rules would not apply, for the same reasons that they would not apply to the assumptions; the models are treated as binding and impose substantive obligations to hold capital based on the design of the models themselves."

Whether the paper will culminate in a lawsuit remains unclear. The paper clearly offers the Fed an opportunity to change its stress testing regime in order to "reduce the threat of legal challenge."

But while banks have chafed at the annual stress testing ritual in the past, they have also held up their successful completion of the tests as evidence that their institutions are safe and secure.

Jamie Dimon, JPMorgan Chase's president, chairman and CEO, said Monday that he is "very much in favor of stress tests" and praised their severity. Dimon said his bank now holds enough capital to pass the severe stress scenarios for all 31 systemically important financial institutions that undergo the tests.

"They make it severe — not just severe in terms of [major losses] but assuming we make a lot of mistakes in the process," Dimon said. "JPMorgan has $500 billion in capital today. That's enough to bear the losses of all 31 SIFI banks."

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.