Capital Ideas Blog

Bank stress tests: How much should be revealed?

By Francine McKenna
March 20, 2014

From: Blog


The US Federal Reserve has just announced results of the first two banking stress tests, which are supposed to tell regulators if large bank holding companies have enough capital to survive another economic downturn. The results suggest that, in the aggregate, the 30 bank holding companies would suffer big losses.  


The 2010 Dodd-Frank Act mandated bank stress tests, but didn’t specify how detailed the results should be (or if they should be bank-specific). In a forthcoming research paper, Haresh Sapra, professor of accounting at Chicago Booth, and his colleague Itay Goldstein of the University of Pennsylvania discuss the pros and cons of full disclosure. Market discipline should improve if stress-test results provide insight about a bank’s ability to withstand market shocks. 

“Public disclosure of a bank’s financial condition enables market participants to make informed decisions about the bank and such informed decisions, in turn, discipline the bank’s actions,” Sapra and Goldstein note. However, bank-specific disclosures may encourage bankers to game the process by dressing up their balance sheets. That might get them a passing grade, but reduce long-term shareholder value.

Stress tests provide more information to regulators, allowing them to better monitor banks and intervene earlier to recapitalize weak or insolvent banks. Regulators intervened late during the last crisis and panic ensued when it was difficult to distinguish a solvent bank from an insolvent one. That damaged regulators’ credibility. It might be improved by disclosing the stress test methodology as well as its results.

Sapra and Goldstein worry, though, about the unintended consequences of full disclosure. If tests are not properly designed, markets could panic. Stress tests have two implied goals that may be incompatible. The microprudential goal is met if regulators can be sure an individual bank has sufficient capital to absorb potential losses and remain solvent. The macroprudential goal is met when we’re reassured the banking sector can survive a systemic crisis, even if the survival of any one institution is not necessarily guaranteed.

The stress test disclosure debate has been dominated by the goal of improved market discipline without considering potential downsides, Sapra and Goldstein argue. To address the issue, they make several policy recommendations. First, test models should not be disclosed so banks can’t “study to the test.” Secondly, test disclosures should include detailed information about the risk exposures of each bank by asset class, country, and maturity, so the market can evaluate whether the bank engaged in sub-optimal behavior to pass the test.

Disclosure of aggregate results instead of bank-specific results promotes risk-sharing activities in the interbank market, reduces bankers’ incentives to take undesirable short-term actions, minimizes the possibility of the “self-fulfilling prophecy effect” of a warning, and maintains regulators’ ability to glean the information they need.

But if the goal is to protect investors and other stakeholders and promote the stability of individual banks, individual results have to be disclosed, even if regulators try to limit their unintended consequences.

The Fed returns on March 26 with results from part two, an exercise it says should “ensure that institutions have robust, forward looking capital planning processes that account for their unique risks and sufficient capital to continue operations throughout times of economic and financial stress.”

We can only hope.

Archive