Fed seeks balance between transparency, toughness in stress tests

NEW YORK — The Federal Reserve Board continues to grapple with balancing the need for transparency in its stress testing exercise with maximizing its overall effect, the central bank's top bank supervisor said this week.

The comments by Fed Vice Chairman for Supervision Randal Quarles came as the Fed has drawn mixed reviews for changes to the stress test regime, including adjustments to the information provided to banks about the stress testing models and the release of a new round of testing scenarios to assess banks' capital strength in a crisis.

Speaking here before a gathering Wednesday night of the Council for Economic Education, Quarles said the changes — which include setting 17 banks on a two-year stress testing cycle — are vital to the evolution of a critical post-crisis tool.

Federal Reserve Vice Chairman for Supervision Randal Quarles
Randal Quarles, vice chairman of supervision at the Federal Reserve, listens during a panel discussion at the National Association of Business Economics' (NABE) Economic Policy Conference in Washington, D.C., U.S. on Monday, Feb. 26, 2018. Quarles offered an optimistic view of the U.S. economy, suggesting it may be on the cusp of a sustained period of faster growth and reaffirming his support for "gradual" interest-rate increases. Photographer: Joshua Roberts/Bloomberg
Joshua Roberts/Bloomberg

“Improvements like these are necessary to ensure our supervisory framework evolves from its post-crisis origins to an effective steady state,” Quarles said. “The question of how best to consolidate the gains from the first 10 years of stress testing deserves the attention and effort of the country's best minds. We should welcome changes and novel ideas, even when they explore stress testing in a new and unfamiliar light.”

Yet the Fed has encountered criticism over both the new scenarios and its efforts to be more transparent. The Fed acknowledged in its statement announcing the 2019 "severely adverse scenario" — a set of conditions in a hypothetical downturn — was more severe than last year’s scenario, which itself was more severe than those of prior years.

In a Feb. 5 blog post, economists at the Bank Policy Institute said their analysis of the new scenario indicated that the hypothetical changes in unemployment rates and GDP were not only more severe than last year’s scenario, but were actually more severe even than the 2007-2008 financial crisis. That potentially goes against the agency’s own guidance that calls for severely adverse stress scenarios to mimic the severity of the financial crisis.

“We have shown that the Federal Reserve’s 2019 scenario is much more severe than the 2007-2009 financial crisis and any post-war recession,” wrote Francisco Covas and Bill Nelson, the head of research and chief economist, respectively, at the Bank Policy Institute. “It is therefore inconsistent with its own guidance for how it constructs the stress tests. And, once again, it was issued without prior notice and the opportunity for public comment, including analyses like this one.”

The other changes finalize a suite of adjustments to what the industry can know ahead of time about the testing process. The Fed will now publish details on a range of post-stress outcomes for actual portfolios and more specific post-stress results for hypothetical portfolios — changes in line with the Fed’s December 2017 proposal.

Dennis Kelleher, president of the Wall Street watchdog Better Markets, said in a statement that the post-crisis stress testing regime made meaningful improvements to the safety and soundness of the financial system. He warned about the unintended consequences of giving banks too much information.

“The Fed’s rigorous and independent stress tests have been one of the most important and successful post-crash tools used by U.S. regulators,” Kelleher said. “While transparency is generally good, it will be incredibly counterproductive if it allows the banks to game the tests. The Fed must be vigilant to ensure that that this critical tool does not become ‘no stress’ stress tests."

Quarles said that the issue of transparency in the Fed’s stress testing program falls into two basic areas: transparency into the Fed’s proprietary models and transparency into how the agency develops the stress scenarios through which banks’ portfolios must be run.

With respect to the models, he said, the U.S. has developed a system whereby the Fed’s own proprietary model is the binding constraint, meaning that there is a risk in over-divulging its secrets. In other jurisdictions, he said, the emphasis is more on monitoring the effectiveness of banks’ proprietary models, which reduces the risk of manipulating the system.

“For me, the concern ... around total transparency of the models is less this issue of gaming — that you’re giving away the answers to the test — and more of a model problem,” Quarles said. “If you have a single model that all portfolios are being judged against, and everyone knows everything about the details of that model, then … whatever mistakes we have made in the construction of our models … those become the fault lines in the entire financial system.”

With respect to the Fed’s stress scenarios, Quarles said the Fed is in a somewhat more difficult bind. The existing practice of developing each year’s severely adverse stress scenario in a vacuum brings with it a risk that banks will tailor their real portfolios to mitigate the losses exhibited in the hypothetical stress scenarios.

Some measure of public input could reduce those artificial changes, Quarles said, but the agency must be careful to make sure the scenarios remain rigorous.

“Frequently, if we are honest about it, there have been unintended consequences with each of the scenarios over each of the last 10 years,” Quarles said. “That could be improved by having more public input into them. The question is, how can you get that public input without giving the questions away?”

Despite their criticism of the new scenarios, the BPI authors nonetheless said the group expects banks subject to the 2019 scenario “to meet the post-stress hurdle rates because of significant increases in capital over the past decade.”

Some analysts said the overall impact of the transparency changes was a net positive for banks, which have urged the Fed to make the models and scenario-formation process more understandable for years.

“Overall I think it's pretty good for the banks,” said Ian Katz, an analyst at Capital Alpha Partners. “The scenarios are tough, but pretty close to what was expected, and the increased transparency is something the banks have been wanting for a long time, and they got it.”

Kevin Fromer, president and CEO of the Financial Services Forum, echoed that sentiment, saying in a statement that the Fed’s recognition that the banking industry is better capitalized than it was at the outset of the financial crisis and commitment to rightsizing the stress testing regime represent positive developments for the industry.

“The Fed has rightly recognized the strength and resilience of large financial institutions,” Fromer said. “The Forum supports federal regulators’ commitment to modernize the current stress test regime and will work to maintain a safe and sound financial system that can be further improved in part through greater transparency.”

But Andrea Usai, associate managing director with Moody’s Investors Service, said in a research note that the result was something of a mixed bag for the industry, though perhaps not in the way that one might presume. Larger banks that are still subject to the 2019 stress testing cycle will be able to prove their resilience once again, she said, while smaller banks that were placed on a biannual cycle may be seen as more of a credit risk.

“The Fed’s stress test announcement is credit positive for the largest U.S. banks because their capital resiliency will be tested against harsher macroeconomic conditions than last year,” Usai said. “But it is credit negative for less complex U.S. banks because those firms will not be subject to a new stress test in 2019, easing regulatory scrutiny and potentially allowing for capital distributions based on last year’s results even for banks whose risk profiles may have changed in the interim.”

For reprint and licensing requests for this article, click here.
Stress tests Minimum capital requirements Regulatory reform Regulatory relief Randal Quarles Federal Reserve
MORE FROM AMERICAN BANKER