The Federal Reserve’s stress testing of big banks may be becoming too predictable to accurately catch a crisis that could severely disrupt the financial industry.
So says a working paper by Paul Glasserman and Gowtham Tangirala, two Columbia University researchers, written when Glasserman was a consultant to the Office of Financial Research, a new arm of the Treasury Department.
The authors analyzed the results so far of the Federal Reserve’s Comprehensive Capital Analysis and Review (CCAR) and the Dodd-Frank Act Stress Testing (DFAST) program, and released the report ahead of the Fed’s first round of 2015 testing. (Initial results are to be posted on Thursday.) They concluded that projected losses by bank and loan category within the stress test results that have been made public “are fairly predictable and are becoming increasingly so.”
“But whereas the results of stress tests may be predictable, the results of actual shocks to the financial system are not, and herein lies the concern,” the authors wrote. “The process of maturation that makes stress test results more predictable may also make the stress tests less effective.”
The two outline several options to mitigate this, including just accepting greater predictability, because if bank portfolios change slowly, then their capital levels would likely change slowly as well.
“And a predictable process still has value: the stress tests require banks to invest in resources for thorough risk assessment with overall benefits for financial stability. The CCAR process includes much more than stress testing, and the other dimensions of the CCAR review may take on greater relative importance than the stress test over time.”
Still, the authors remain concerned with “a routinized stress test” because it could likely lead banks to “optimize their choices for a particular supervisory hurdle and implicitly create new, harder to detect risks in doing so.”
To reduce predictability, the stress tests could require banks to qualitatively describe how they would handle adverse and severely adverse scenarios, and significantly expand the overall number of scenarios.
“Third and most ambitious, the stress testing process could be expanded … to include knock-on and feedback effects between institutions, and interactions between solvency and liquidity, leading to a richer set of outcomes than can be achieved through a fixed set of stress scenarios applied separately to each bank,” the authors wrote.
“Such a process, though difficult to implement, would respond to changes in the financial and economic environment and would be less likely to get stuck in a predictable outcome.”
A Wall Street Journal story said the Fed currently tests both quantitative and qualitative responses to certain scenarios, but, “in recent years, the qualitative side of the Fed’s annual tests have been what has gotten some banks in trouble.”
Indeed, Citi failed its Fed stress test last year, as the agency rejected its request to return capital to shareholders over qualitative shortcomings.
Featured image: Thinkstock