In this guest post Ricardo J. Caballero, the Ford International Professor of Economics at MIT, offers an alternative scenario for banks that fail Treasury’s stress tests.
A central aspect of the Obama administration’s financial stabilization plan is to stress test the major banks and to recapitalize the ones that do not fare well. Analysts’ responses to this plan have been lukewarm at best. Some claim that the aggregate scenarios used for the tests are too rosy. Other analysts think this to be a creeping nationalization policy. Either way, the news for the stock market is awful and, as a result, we have observed massive wealth destruction since the plan’s announcement. Rather than help to contain fear, uncertainty has now risen to a new level, further complicating the financial health of the very institutions that were supposed to be stabilized under the plan.
I argue here that these perception problems can be solved by more directly addressing the core fear issue. This requires a clear policy principle: Whatever we do, financial institutions should not be left holding the downward aggregate risk.
Under this basic principle, the stress test policy would be slightly modified as follows: Each bank would be subject to stress tests as planned (I would advise giving some weight to even more extreme aggregate scenarios than those that have been proposed). In this case, however, the result of the stress test should not be used to determine how much more capital the bank may need in an extreme scenario. Instead, it would be used to determine how much insurance the bank needs to buy from the government to protect itself against these scenarios.
The immediate impact of such an apparently small modification would be to narrow the risks perceived by banks and their stakeholders. This would encourage banks to lend because the insurance policy would replace the need for massive self-insurance induced hoarding. It would also encourage private recapitalization. By insulating the financial system as much as possible from the downward risk of the current macroeconomic environment, the policy would break this deadly downward spiral where a worsening macroeconomic environment weakens financial institutions, which in turn further weaken the macroeconomic environment, and so on. With this principle firmly in place, survival risk would be extensively reduced allowing investors to make their decisions based on the long-term prospects of the particular institutions. Without this, investors and banks will remain mired in uncertainty and fear, relegating the natural forces of recovery (and there are many of them) to the sidelines.
A key dimension of the policy principle is to think not only about the particular institution in trouble but about the systemic implications of the policy. For example, some could argue that one way of breaking the perverse feedback between the health of a financial institution and the macroeconomic environment is to nationalize the bank (which means almost by definition, that the full insurance for the losses rests upon the taxpayer). However, while this would indeed work for an individual bank, it would exacerbate, rather than reduce, the aggregate risk faced by non-nationalized institutions because now, a negative macroeconomic shock would make it more likely that they too would be nationalized. It is important not to fall into a fallacy composition where what is good for one institution in isolation may be just the opposite for the system as a whole. In other words, the principle goal should be the removal of the aggregate downward risk not only from individual financial institutions but from the leveraged institutions as a whole.
Who should pay for this insurance? Those most directly protected by the policy. Certainly the shareholders are first in line, but there is no reason to stop here. All preferred shares and debt that are being discounted by uncertainty should also contribute, perhaps by retaining some of the interest payments to service the insurance policy. The government should charge fairly for this insurance, using the same probabilities it assigns to the different scenarios used in the stress tests, and it should mandate that the banks acquire this insurance once they have converged on the particulars of the stress test.
