Richard Fisher and Harvey Rosenblum, president and research director respectively of the Dallas Fed, offer their prescription for mitigating the risk of banks that are “too big to fail”:
First, we would roll back the federal safety net—deposit insurance and the Federal Reserve’s discount window—to apply only to traditional commercial banks, and not to the nonbank affiliates of bank holding companies or the parent companies themselves, where the safety net was never intended to be.
Second, customers, creditors and counterparties of all nonbank affiliates and the parent holding companies would sign a simple, legally binding, unambiguous disclosure acknowledging and accepting that there is no government guarantee—ever—backstopping their investment. A similar disclaimer would apply to bank deposits outside the FDIC insurance limit and other unsecured debts.
Third, we recommend that the largest financial holding companies be restructured so that every one of their corporate entities is subject to a speedy bankruptcy process, and in the case of banking entities themselves, that they be of a size that is “too small to save.” Addressing institutional size is vital to maintaining a credible threat of failure, thereby providing a convincing case that policy has truly changed. This step gets both bank incentives and structure right, neither of which is accomplished by Dodd–Frank.
I think the first two prescriptions are boilerplate. The root of the problem is that whatever the laws the institutions are seen as too big to allow to fail. “Don’t do that” isn’t enough of a prescription.
The third, unfortunately, is just wishful thinking. It can be stated more succinctly as “Just do it!”. It presupposes that breaking up the elephantine big banks into chewable pieces is a shared goal. Quite to the contrary, if the big banks are broken up and rendered small enough to be allowed to fail where will politicians and apparatchiks go to rake in the big bucks after (or between) their period of what we laughingly refer to as “public service”?