Editor’s note: This is an altered version of a post that originally appeared here.

One of the summer highlights for Federal Reserve watchers is the annual ritual in which the Fed’s soothsayers peer into their crystal balls, aka the central bank’s stress tests, to reassure us that the U.S. banking system is robust and getting stronger all the time.

You see, while the future is always uncertain, the results of the stress tests are not. So we can all rest easy … or can we?

This year, the news seemed particularly good. Whereas in previous years there were dunces that failed, the latest stress tests are the first in which all the banks passed. The bank that could be most viewed as a laggard, Capital One, only got a mild slap on the wrist.

Federal Reserve building in Washington, D.C.
There is no sign that the Federal Reserve is inclined to take away the punch bowl any time soon. Instead, the Fed plans to make future stress tests even less demanding. Bloomberg News

The Fed is so pleased with the performance of its examinees that it decided to reward them with a big dividend/buyback party that will allow virtually all net earnings to be distributed. The Fed provides the punch bowl which will be paid for by other bank stakeholders, including taxpayers — yes, the same taxpayers who are still being compelled to subsidize the banks (via “too big to fail”) to take excessive risks and overleverage themselves.

There is no sign that the Fed is inclined to take away the punch bowl anytime soon. Instead, the Fed plans to make future stress tests even less demanding. The days when banks felt challenged by the stress tests are over.

The results for the Fed’s stress tests were published in two stages. The results of the first stage — the Dodd-Frank Act Stress Tests (DFAST) — were published on June 22. These showed how the banks involved would fare under an “adverse scenario" and a “severely adverse scenario” projected over a two-year period ending in the first quarter of 2019. The 34 bank holding companies (BHCs) tested hold more than 75% of the assets of all domestic BHCs.

The “severe adverse scenario” posits a GDP more severe than during the “global financial crisis” (GFC). It also has the unemployment rate almost doubling to 10%, equity prices falling by 50%, equity market volatility surging to 2008 levels, and house prices and commercial real estate prices experiencing large declines. Total projected losses across the 34 banks are $493 billion.

The results for the second stage — the Comprehensive Capital Analysis and Review (CCAR) 2017 — were published on June 28. The CCAR provides a quantitative assessment of a firm’s capital adequacy and planned capital distributions (or “capital plan”), such as any dividend payments and stock buybacks.

“This year's results show that, even during a severe recession, our large banks would remain well-capitalized,” said Federal Reserve Board Gov. Jerome Powell, the official in charge of the stress tests, on June 22.

But I disagree and here’s why.

My doubts start with the observation that the primary purpose of the stress tests program is to reassure the public that the banking system is safe. That is what central banks always say. Yet the stress tests are not some independent assessment of the financial strength of the banking system carried out by experts free to arrive at conclusions that might not suit the Fed; instead, they are part of a publicity campaign by a public agency with its own interests and agenda.

A second problem is that the stress tests place far too much emphasis on a single “severe adverse scenario,” putting all of the assessment’s stock on one discrete hypothetical event, ignoring the breadth of different and distinct potential scenarios. There are multiple scenarios that could have a significant adverse impact on the U.S economy. These include a collapse in the Chinese or eurozone banking systems or a major trade war. Do the Fed’s stress tests provide any assurance that the U.S. banking system could withstand such shocks and remain in good shape? No, because the stress tests did not consider them.

Turning now to the modeling itself, the stress test projections fail a basic reality check: the Fed’s projected losses are too low to be plausible. The projected loss of $493 billion in a GFC-like scenario is barely half the losses accumulated as a result of the actual GFC. The stressed portfolio loan-loss rate of 5.8% is also low for a major stress. Loan-loss rates in crises can be well over double that.

The stress test models in a sense try to achieve the impossible. Frank Partnoy and Jesse Eissinger have made the case that it is no longer possible even for a qualified expert to infer a bank’s true financial state from its audited accounts. If a bank has a large hidden loss that does not show up on its accounts, we cannot expect a stress test to identify it. If the accounting numbers are not right, the stress tests cannot correct for them.

But the Fed’s stress tests have a bigger immediate problem: the standards for passing are too low. This issue is particularly acute for the eight U.S. “global systemically important banks.” The pass standard applied by the Fed to these banks was a minimum enhanced supplementary leverage ratio of 3%. While they were all able to meet that standard, had the Fed applied a 4% minimum the capital plans of only four banks (Bank of America, Bank of New York Mellon, Citigroup and Wells Fargo) would have been approved. At a 5% minimum, only Wells Fargo’s plan would have been approved, and at a 6% minimum none of the banks’ capital plans would have been approved.

The use of a 3% pass standard is all the more curious because the Fed is in the process of imposing a 5% minimum supplementary leverage ratio — unrelated to the stress tests — on the systemically important BHCs and a 6% minimum ratio on their federally insured bank subsidiaries, to become effective on Jan. 1.

Goldman Sachs appeared to exploit the Fed’s low standard to perfection. With a barely passing CCAR supplementary leverage ratio of 3.1%, Goldman managed to persuade the Fed to approve a capital plan that left it with 10 basis points to spare over the Fed’s regulatory minimum. If 3% remains the minimum, there could be a race to the bottom in future CCARs, with other banks converging down to the passing level.

Why is the Fed approving generous distributions when the banks’ leverage ratios are so low? Two words: regulatory capture.

The Fed is so confident that the U.S. banking system is in good shape that on June 25, Chair Janet Yellen said that she did not expect to see another financial crisis in her lifetime. Let’s hope that her prediction does not go the way of the economist Irving Fisher’s famous line, “Stock prices have reached what looks like a permanently high plateau,” 12 days before the October 1929 crash.

The good performance of U.S. banks in the stress tests indicates that banks have mastered the now routine stress test game. This year’s stress tests are great news for bank shareholders, but bad news for everyone else.

Kevin Dowd

Kevin Dowd

Kevin Dowd is a professor of finance and economics at Durham University in the United Kingdom and an adjunct scholar at the Cato Institute.

BankThink submission guidelines

BankThink is American Banker's platform for informed opinion about the ideas, trends and events reshaping financial services. View our detailed submission criteria and instructions.