THE FED GETS ITS OWN STRESS TEST / THE WALL STREET JOURNAL EDITORIAL

The Fed Gets Its Own Stress Test

Under a lawsuit threat, the bank regulator agrees to change its bank capital tests.

By The Editorial Board

The Federal Reserve building in Washington. Photo: joshua roberts/Reuters

Regulators rarely concede mistakes, but that’s essentially what the Federal Reserve did Monday by announcing it will open up bank stress tests to public comment. 

Consider this another victory for the conservative judicial project.

“In view of the evolving legal landscape, the Federal Reserve Board will soon seek public comment on significant changes to improve the transparency of its bank stress tests and to reduce the volatility of resulting capital buffer requirements,” said a Fed press release, adding that “the framework of administrative law has changed significantly in recent years.” 

It sure has, and the bank lawsuit is a welcome example.

Thanks to the Supreme Court’s West Virginia v. EPA(2022) and Loper Bright(2024) precedents, courts no longer must defer to regulators’ interpretations of supposedly vague laws, especially those involving major questions. 

Judges are slapping down agencies when they make law by regulatory fiat.

That’s what the Fed has been doing with its capital stress tests, as banking groups argue in a lawsuit filed Tuesday. 

Banks say the Fed is violating the Administrative Procedure Act by dodging notice and comment on its models. 

Our sources say the Fed sought to head off the lawsuit by agreeing to change its testing process.

The banking groups welcomed the Fed announcement, but their lawsuit notes that “the deadline for a court challenge to some of the government actions undergirding the current stress test process is February 2025.” 

And they can’t be certain the “recently-announced reforms will provide a timely remedy to the harms arising under the current system.”

Fed stress tests arose from the 2010 Dodd-Frank Act, which requires regulators to evaluate balance sheets of big banks every year to ensure they can “absorb losses as a result of adverse economic conditions.” 

In practice the tests arbitrarily punish banks based on flawed Fed models and impose costs across the economy.

Behind closed doors, regulators design models of hypothetical economic shocks, including their effect on GDP, unemployment, inflation and asset values, among other things. 

Regulators then run bank data—e.g., lending portfolio, capital, reserves—through their models to project how banks would withstand the hypothetical shock.

Bank performance on these tests is used to set a “stress capital buffer,” which is a de facto penalty. 

Banks must hold more capital if they score lower on the test. 

The penalty varies from year to year depending on the Fed’s model design and is layered on top of myriad other bank capital requirements. 

UBS’s required stress capital buffer doubled between 2022 and 2024.

The tests do a lousy job of measuring whether banks can withstand actual economic risks. 

Consider the Fed’s 2022 test, which projected short-term interest rates would remain near zero and the 10-year Treasury would drop to 0.75% in the first quarter of 2022 from roughly 1.5% and remain unchanged during the next two quarters.

Yes, the Fed stress-tested banks for lower interest rates even as it prepared to increase rates. 

Many regional banks scored well, though they faced large unrealized losses when the Fed raised rates. 

The panic in spring 2023 among mid-size banks was one result.

Early in the first Trump Administration, the Fed proposed opening up the design of its test models to public comment, which would make them more useful in gauging the strength of bank balance sheets. 

Because banks can get dinged under Fed models for making certain business loans, it’s only fair to let affected business weigh in.

Yet the Fed in 2019 walked back its proposal for more openness owing to supposed worries that this would cause “model monoculture” and banks would build their balance sheets to game the tests. 

The banks dismantled this concern in public comments to the Fed proposal and again in this week’s lawsuit.

Banks can’t improve their performance without making long-term changes in their portfolio, which is impractical since the Fed’s models change every year. 

What’s more, “if models accurately identify the relevant risks, conformance with the model (and thus better performance on the stress tests) should in fact reflect a lower risk profile,” the lawsuit notes.

If the Fed is truly worried about herd behavior, it could revisit its risk-weighted capital rules. 

The stress tests are redundant given the Fed’s other capital requirements, and they may encourage moral hazard. 

But if Congress won’t eliminate them, the least the Fed can do is make them more open and useful.