The day before the FRB released its stress-test edicts, the FT suggested that the tests are little more than “public theater.” If there is a show on, though, it’s one where the first act was written by George Bernard Shaw and the second was written by the South Park guys — each fine on its own, but incoherent and contradictory when taken together. Like this theatrical spectacle, the stress test has two parts – quantitative and qualitative. Fine on their own, but together a mish-mosh that creates conflicting expectations and costly disappointments.
Here’s why: the stress test started as the first macroprudential standard. In 2009, it was a horizontal, quantitative review that pitted the very largest banks against scenarios the Fed thought could befall them. Judged against each other instead of by each bank on its own, the goal was first to see if the biggest banks could withstand acute stress and, if they could, to reassure markets on a question then all-too-critical to a financial system poised over the brink of another systemic crisis. Banks were of course forced to raise capital if they fell short of the 2009 standard, but each was judged by the FRB’s horizontal pushes and pulls, not by additional factors unique to the individual institution.
Although regulators hope to find a way to make macroprudential regulation counter-cyclical, it has so far been deployed largely through these stress tests. With each bank compared on common criteria to others, outliers are identified and forces that could pull the system apart are, it is hoped, also seen enough in advance to blunt their force. For these purposes, it’s fine to wash out differences that are immaterial to macroprudential decisions and to normalize certain assumptions that, while perhaps significant to one or another bank, do not loom large in the stress scenarios.
But, when macro standards go micro, these bank-specific factors loom large. The stress tests moved from macro to micro when the Fed decided not only to judge banks against each other based on quantitative standards, but also to add the qualitative element embodied in the Board’s capital-planning rules. These are micro standards because each bank cannot distribute shareholder capital unless or until it satisfies supervisors on very specific points such as the extent to which astute, independent corporate governance reigned over the distribution decision.
We might still have a coherent drama with a two-act stress test if the quantitative criteria linked well to the qualitative ones. However, we actually have a play in which the actors who are auditioning to be dancers get rejected because they can’t sing the solo. Factors that make a horizontal test work for horizontal, macro purposes are at fundamental odds with those needed to judge each bank on its own. Does it make sense to demand that banks go to the trouble, not to mention the huge expense of buying a ticket to a two-act play if the real drama comes only at the end of the second act? Yes and no.
Yes, because horizontal stress tests are vital ways to see how systemic banks withstand stress and which banks show themselves as good or bad against a norm that, for all the complexity of the stress-test criteria, is still a crude simulacrum of actual systemic stress. No, because counting on these Fed-imposed, limited, artificial templates to tell us much about each bank is like asking George Bernard Shaw to save Kenny.