Wealth externalities and limited liability in banking

Arnold Kling writes:

Suppose that you sell your shares in Shakee Bank to me today, and tomorrow Shakee has to be taken over by the FDIC. Am I liable for losses, which probably were caused by decisions made before I bought my shares? Suppose that Shakee has accumulated $8 billion in losses, and all its shareholders of record obtained their shares for a collective $0.10. What happens then?

I don’t envision the FDIC being eliminated, but say the government is in a position to be picking up some potential bondholder losses.  Under one version of the reform, bank shareholders already have posted extra collateral, by requirement of the law.  That is somewhat like higher bank capital requirements, with the twist that there is now a new legal class of bank capital.

That is an improvement over the status quo, but it’s not the most innovative form of the proposal.  One alternative version is for the government to outsource the enforcement to the bank itself.  For instance the regulator can say: “as insolvency approaches, the bank is liable for 1.5 to 1, it can come up with the money any way it wants. If it can’t come up with the money, we will take the major shareholders of record, say a year before the event (or consider a more complicated weighted average of this variable) and send them an income tax assessment for 2-1.”

The bank might preemptively organize like a partnership, or it might apply its own collateral and capital requirements to the shareholders, or it might find some other way of meeting the obligation.  Banks would compete to find the better solutions.

In response, many people fear banks trying to set up with only hobo shareholders.  That would avoid the 2-1 or 1.5 to 1 or whatever, because hobos don’t have extra assets to attach.  I just don’t think those banks will become major money center institutions because the quality of shareholders really does matter at some level.  For instance such banks could not have wealthy, highly motivated, equity-holding CEOs.  Most likely hobo banks would stay small and thus skirt the too big to fail problem or maybe they would not exist in the first place.

One problem with traditional capital requirements is that the government ends up making comovement-inducing ex ante decisions about which assets count toward satisfying the capital requirement.  Remember AAA CDOs in America and AAA government securities in Europe?  Under non-limited liability, only cash is accepted but it only has to be delivered ex post in the case of failure.  The regulations themselves need not create the same kinds of uniformity, misjudgments, and excess systemic risks up front.

One tricky question is how to apply non-limited liability to foreign banks operating in the United States.  This is a problem with all regulatory schemes based on less than perfect international coordination.  The first cut approach is to insist on non-limited liability for U.S. operations, though of course evasion and reclassification of operations may occur.

Mark Thoma adds lengthy comments.  Here is a very relevant paper by Claire Hill and Richard Painter, and a blog post by them.  Here is Suzanne McGee.  Here are some debates on non-limited liability in economic history, including work by Lawrence H. White.