Small changes to stress testing could have big impact on risk, Tarullo says

WASHINGTON — Former Federal Reserve Gov. Daniel Tarullo broke his silence over several of the central bank’s recent proposals to revamp regulations put in place in the wake of the financial crisis, saying incremental reductions in capital and other changes could reduce the financial system’s resiliency.

Speaking at a conference sponsored by Americans for Financial Reform Tuesday, Tarullo — who is currently serving as a visiting professor at the Harvard Law School — acknowledged that the regulatory apparatus he helped establish after the financial crisis was far from perfect. But the changes that the Fed has proposed since his departure signal a reduction in regulatory constraints without any complementary changes that would ensure the banking system’s resilience in the face of stress, he said.

“I don’t at all think that the state of regulation when I left the Fed ... was the best it can be,” Tarullo said. “But we had made considerable progress enhancing the resiliency of these largest banks, and at least putting a sizeable dent in the Too Big to Fail concern. Unfortunately, I feel that a good bit of that progress could be endangered by a kind of low-intensity deregulation, consisting of an accumulation of non-headline-grabbing changes and an opaque relaxation of supervisory rigor.”

Federal Reserve Gov. Daniel Tarullo
Daniel Tarullo, governor of the U.S. Federal Reserve, speaks during an event at Princeton University in Princeton, New Jersey, U.S., on Tuesday, April 4, 2017. It may be time to phase out one of Wall Street's most feared aspects of the Federal Reserve's stress tests, Tarullo said Tuesday on the eve of his last day on the board. Photographer: Ron Antonelli/Bloomberg
Ron Antonelli/Bloomberg

Tarullo limited his focus to the various changes proposed to the Fed’s stress testing regime, which he said include relatively minor adjustments that could have major ramifications for capital retention by the largest banks if implemented.

One example he cited was a proposal to replace many of the Fed’s post-stress minimum capital benchmarks with a “stress capital buffer” — a concept Tarullo himself suggested in the final months of his tenure at the Fed. The Fed’s proposal differed in at least one important way, he said, and that is in the application of the enhanced supplemental leverage ratio to the post-stress minimum capital position required of the global systemically important banks, or G-SIBs.

“Buried in a footnote in the explanation of the Fed’s proposal was the information that the Fed intended to exclude the enhanced supplemental leverage ratio for what was supposed to be a fully integrated set of capital requirements,” Tarullo said. “This is important, because the binding capital constraint on many G-SIBs has been the post-stress leverage ratio.”

Tarullo said that the stress capital buffer, as he conceived it, would ease some of the balance sheet and distribution assumptions in the Fed’s stress testing models — changes that would likely result in the largest banks having to hold less capital to cover their post-stress minimum leverage requirements. But the application of the G-SIB surcharge to the risk-weighted capital requirements would “more than offset” that capital reduction, he said.

If the Fed’s proposal is finalized without the application of the enhanced supplemental leverage ratio continuing to be factored into G-SIBs’ leverage ratio, he said, the net effect would be to reduce the capital constraints on the biggest banks in a way that would simplify compliance for banks but potentially hurt safety and soundness.

“If you relax the assumptions but you don’t increase the leverage ratio number, that means an effective reduction in the dollars of capital that would need to be held with the same balance sheet by banks that are currently constrained by the leverage ratio,” Tarullo said.

Tarullo was similarly critical of some of the Fed’s proposals to increase transparency in the stress testing regime. In March, the agency released a supplemental document that provides new descriptions of the central bank's stress testing models, information on how hypothetical portfolios might perform when tested through the Fed’s model as well as a range of results from the performance of actual portfolios.

If banks have a clear understanding of how the Fed anticipates losses in a certain asset category under certain levels of stress, they will be able to apply subtle changes to their balance sheets that minimize projected losses without dramatically changing the institutions’ overall holdings, Tarullo said.

“While the code itself has not been released, I suspect that the smart people who work on such things for the big banks now have most of what they need to reverse-engineer the models,” Tarullo said.

Tarullo also said the idea of subjecting the Fed’s stress scenarios to public comment — an idea that has not been proposed by the Fed but that Vice Chairman for Supervision Randal Quarles has expressed interest in exploring — has the potential to undermine the entire stress testing enterprise.

Prior to the financial crisis, he noted, the Office of Federal Housing Enterprise Oversight stress tested Fannie Mae and Freddie Mac, but reused the same scenarios year after year. If the changing and unpredictable dynamic of the Fed’s scenario selection process is difficult to anticipate and frustrating to banks, he said, it is in service of simulating the unpredictability of financial crises.

"Volatility is a feature of a stress test regime — it’s not a bug that needs to be corrected,” Tarullo said. “If the stress test becomes predictable, it ceases to have value.”

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