In the wake of the financial crisis, many economists are trying to come up with creative new ways to deal with systemic risk: the risk of a "wholesale bank failure" and failure of the financial system in general. I just finished reading Judge Posner's recent book, A Failure of Capitalism. In it, Judge Posner makes a convincing case that individual bankers can (and did) make rational decisions that, at least in the aggregate, greatly increase systemic risk. I do not wish to go into the details of that analysis here; I just want to absorb its truth.
When rational actors make decisions that create negative externalities, it often falls upon the government to adjust the incentives to account for those outside costs. In banking, for instance, Citigroup may make certain decisions that increase its risk of bankruptcy to 1%. For a smaller bank, that risk would only be negligibly important: the bank could fail and go into receivership. But for Citigroup, of course, such a failure would have BROADER effects: it would not be able to keep the (many) promises of payment it regularly making to other banks (cascades); it would create a "fire sale" situation wherein bank assets would have to be sold by the FDIC at a sharp discount; and confidence in the economy overall would sharply declare. A systemic risk regulator would intervene to prevent a Citigroup (or one of its similarly-sized cohorts) from taking these individually rational (but systemically risky) actions. Even Tyler Cowen suggests that we might need such a regulator, and it probably needs to be the Federal Reserve. I respectfully disagree.
I think the best way to regulate systemic risk is to use the insurance premiums charged to banks by the FDIC's Deposit Insurance Fund (DIF). In very simple terms, the FDIC charges banks an insurance premium that is used to cover depositor losses when banks fail. Under the current system, under 12 USC 1817, the FDIC charges a "risk-based" premium that is supposed to be based on: (1) the probability that the DIF will incur a loss for that institution (ie, that the institution will fail); (2) the likely size of any such loss; and (3) the revenue needs of the Fund. Trouble is, the premium is only based on the individual size of each bank's risk to the Fund. Therefore, when calculating Citigroup's premium, the FDIC does not include any of the "contagion" effects noted above. The FDIC is not actually charging for the real "reasonably size of any loss" that the bank will suffer from a big, interconnected bank's failure.
I've seen a few different studies outlining how we could actually set the premiums to account for the systemic effects of a bank failure. I'm not going to venture into that. My only point is this: properly scaled, deposit insurance premiums that include systemic risk would obviate the need for any "systemic risk regulator." If banks that create systemic risk faced increased premiums of any significant size, one would expect them to adjust their behavior to reduce the risk. In fact, the best approach may to charge punitively high promotions. One could anticipate that these punitive premiums could quash the moral hazard created by government bailouts; banks would know that they would pay a high price for setting themselves up to be "too big to fail." Best of all, even if a bank was so brazen as to generate systemic risk in the face of high premiums, the money collected from the bank's premiums would be enough to clean up the (system-wide) mess resulting the bank's failure.
Of course, to accurately assess the premiums and let the market work its magic, the FDIC would need access to an enormous amount of information at banks. Not a problem! The FDIC has the right to examine any FDIC-insured institution if the FDIC's board of directors finds the examination is necessary "for insurance purposes." 12 USC 1820 (b) (3). That would simplify the issue of setting up an entirely new systemic risk regulator with the authority to examine the books of market participants.
Still, I recognize there's a big sticking point: any effective premium would probably have to apply to financial institutions that do not even operate with tradititionally FDIC-insured deposits. I also recognize that "excessive insurance premiums may hinder a financial institution's capacity to resolve its bad loan problems and / or reinvest its owned capital." ( Financial Crises in Japan and Latin America , pg. 82.) And, lastly, there's always a chance that FDIC systemic risk premiums would be miscalculated. There is some suggestion, for instance, that the FDIC misjudged the systemic risk posed by the failure of Continental Illinois National Bank in 1984. Even so, I think that's better than just handing the keys over to the Fed, which has enough to worry about (like managing inflation).