2) Smaller banks have simpler balance sheets. That makes them easier to audit for regulators and for investors. It makes it harder for them to hide outsized risks in “off-balance-sheet instruments”.

3) Most importantly, as Rockfish points out below, is that TBTF institutions have artificially low borrowing costs. Creditors and counterparties know that they will always be made whole, so they have no reason to care if the firm is taking outsized risks. This is a distortion of the market that increases systemic risk. If we made sure that no particular firm was TBTF, lenders and counterparties would be far more likely to impose the discipline that they should have been imposing all along.

Perhaps what you meant was 10 medium size banks instead, so let’s go with that. The idea is that if JPGS were, J, P, G & S instead, they wouldn’t all be exposed to the derivative in the same way, so any one of them (or maybe 2) could go under without jeopardizing the entire economy, and even if the government needed to step in, it wouldn’t need to pledge nearly as much money.

That being said, I think that if you take an existing JPGS and split it up, the interconnectedness between them would considerably higher (necessarily) than the currently criticized web of connections between banks, so their default correlation would be very high, defeating the whole purpose of the split.

While the idea is noble, I think it would be very hard to split it up and get significant benefit from a systematic risk perspective. As is, it’s not like the government only had to support 1 or 2 large banks throughout the current crisis. Between the various programs, help was doled out to a fairly large number of institutions.

I think the TBTF moniker is totally sidestepping this issue. If you break up banks and no one bank is TBTF, when you take correlation into account, you will still find small sets of banks which, considered together, are TBTF.

There are certainly plenty of other issues, such as the implicit government guarantee at no cost to the institutions involved, etc., but I think TBTF is a bit of a red herring.

At the end of the day, I think I agree with you.

Case A: A huge bank (call it J.P. Goldman-Stearns-Bank) underestimates the risk in a new popular type of derivative security and loses billions. Since their collapse would pose a systemic risk to the financial system, the government has to step in, at huge cost to taxpayers.

Case B: A thousand small local banks underestimate the risk in a new popular type of derivative security and lose billions. Since their collapse would pose a systemic risk to the financial system, the government has to step in, at huge cost to taxpayers.

Why, exactly, do we prefer Case B to Case A?
Why, exactly, do we think Case B is less likely than Case A?

Al Franken’s Washington DC phone (Ask for Franken)
202-224-5641