Today the 19 largest banks are getting the results of their stress tests from their regulators.  Should these results be made public?

This is not a simple question.

The big upside is that markets will have more information, and that all market participants will have access to the same information.  This can allow investors, counterparties, and customers to evaluate the health and stability of the banks with which they are doing business.  I have a default presumption that more information more widely disseminated is a good thing.

Downside 1:  Some banks might be just above the bubble but rank low.  They may still be healthy enough to survive and eventually succeed, but if they are disclosed to be among the weakest, that disclosure may cause a run of depositors, counterparties, or investors.  This could push some marginal banks over the edge and cause them to fail.

This is not a trivial concern.  Last September there were reports that investors were “testing” even the clearly healthiest investment banks (JPM Chase, Goldman Sachs) shortly after Lehman’s fall.  Senior policymakers remember that vividly, when panic might have destroyed banks that on paper were solid.  From the perspective of the policymakers who have the information, it is easy to understand why they may be highly risk averse.  From their perspective, not disclosing the information may appear to be the safer course.

There is a response to this downside.  Banks that do well in the stress tests have an incentive to let the world know that.  It may be hopeless for policymakers to think they can protect the weaker banks by not having the government release the information, because the strong banks will do it implicitly by shouting their good news.

Downside 2:  The stress tests might not be well-designed.  If they are poorly designed, overly optimistic, or just misinterpreted by the market, then the government could be injecting bad information into the market.  Government may not be smart enough to design the tests to provide enough useful information to the markets.

More importantly, any stress test is highly imperfect, and there is a risk that the results would create a sense of false certainty.  I read a lot of market commentary while working in the White House, and was amazed at how frequently high-level market commentaries completely misinterpreted or misread data that we released (much less policy statements).

This is a close call, and people whom I respect advise in different directions.  I fall back on my default presumption / instinct, which is to release the information.  There is a small probability of a really bad outcome (a panic/run), and a much larger probability of a good outcome with no run and somewhat better informed markets.  I also think it is easy for policymakers to overestimate the amount of control they have over the situation.

The Wall Street Journal reports the following game plan:

  • Today the regulators are giving results to the banks privately, and asking them not to disclose them.  They are also releasing the test methodology.
  • 10 days from now, the regulators will release results of the stress tests, although “regulators also have not decided how much information will be disclosed May 4.”

On the whole, this game plan makes sense to me.  I am not sure they will be able to hold out for 10 days, and my instinct is that they should release more information rather than less.  I hope market participants will take the time to understand just how imperfect this information will be.

By the way, I presume that the decision to leak that they expect to release the results was made after they knew the results of the tests.  This suggests that the results are generally good.  If the stress tests showed that half of the banks would fail, I presume they would not be signaling a future release.