[mathattack]

I agree with everything you say up to “panics do not destroy capital; they merely reveal the extent to which it has been previously destroyed by its betrayal into hopelessly unproductive works”

In todays fractional banking system where banks only hold a fraction of liquid assets to cover their liabilities, a run on a good bank could still put it under. (Lehman, Bear and others were both illiquid and insolvent, but runs can kill good banks too) This is why the FDIC was put up to guarantee commercial banks. Nothing similar existed to protect investment banks.

[nhaehnle]

Not quite. If a bank is solvent but illiquid, they can get a loan from the central bank. This is the central bank's "lender of last resort" function.

So no, a good bank cannot be put under by a bank run.

We will probably never be able to say to which extent the big banks at the time of the financial crisis were still good banks. The problem there was that banks had a massive amount of assets that were indirect (i.e. whose inherent value relied on other assets) and that were structured in such a complicated way that nobody could assess their inherent value.

Before the panic, the inability to measure the inherent value of those assets was ignored because they could be valued according to their market value. With the panic, the market simply stopped doing anything, and there was no market value anymore.

The FDIC is orthogonal - it is an insurance of deposits (up to a limited amount) even at bad banks.

[mathattack]

In theory that's the central bank's job. 2 issues, though...

1 - It can be hard to tell the difference between solvency and liquidity. What's a derivative of an MBS really worth? Or a CDO that's made off of other CDOs that are trading at an undetermined liquidity discount? Or a unique plot of real estate?

2 - The bailout decisions are often political, as well as based on imperfect reads of fundamentals.

Yes, the FDIC provides run protection from both bad banks and good. Protecting bad is the price of protecting the good.

[nhaehnle]

1 & 2 basically elaborates on what I said ;)

And no, FDIC protection intentionally protects deposits at bad banks. This is not an accident. Trying to put the burden of evaluating what a bad bank is onto regular people is not going to end well, so you guarantee deposits at all banks, full stop.

Besides: Deposit insurance for deposits at good banks is pretty pointless, don't you think? It would never be used by definition.

[mathattack]

I think we agree on 99%.

My point is just that a good bank can still have a run in the absence of the FDIC guarantee.

Let's say a bank has $100 million in deposits. They keep $10 million in liquid assets, and lend out $90 million in un-securitized loans to local businesses. There's a false market rumor of something bad happening at the bank, and all of a sudden $20 million in depositors want their money back. The bank isn't able to resell the loans quick enough on the secondary market to make up for the shortfall. This could happen.

The FDIC guarantee protects the bank because there's no longer a need to have the run - everyone will get paid.

[nhaehnle]

I think you're right about our mostly-agreement :)

Just to clarify, I think we have to distinguish between the likeliness of a bank run and the effects of a bank run.

Indeed, the FDIC makes a bank run extremely unlikely. Perhaps this is what you mean by "protecting the bank".

However, even if there were no FDIC, a bank run on a good (solvent) bank would not cause that bank to collapse, due to the central bank's lender of last resort function.

The bank run would "merely" cause a shrinking of the bank's balance sheet, which the bank would have to offset by selling its assets over time.

It is true that a big change in the balance sheet like that could still lead to the eventual death of the bank, e.g. because the bank has high fixed costs (in the form of physical branches, non-fireable employees, and so on) which can no longer be covered by profits from its regular business. However, this eventual death is (a) not certain since the bank has plenty of opportunity to turn things around and (b) a slow death, very much unlike the sudden implosions that people usually think of when they hear "bank run".

[mathattack]

Once a run happens, the central bank has to make a choice on who to lend to based on imperfect information, though. The central bank might say, "I don't know about all these local real estate loans, let's let them fail."

It's the FDIC that removes this possibility.

Thanks for engaging in this conversation!

[frandroid]

Yeah, but unless you have massive disinformation, you can't create a panic on a solvent bank in today's world. The examples you use were ones where capital had been destroyed.

[tspareme]

I guess you have to define what a solvent bank is. I don't think any bank can handle a sustained run given their liability duration mismatch. So I am not sure what do you mean that you can't create a panic. Its always possible to have runs just not very easily.

The financial crisis showed that bank liabilities are really liabilities of the country that the bank incorporates in. So in case of Ireland, Greece, Iceland, it is up to the country to step up and backstop their banks. If a country can not, then you will have panic on a bank. And in order for a country to be able to backstop their banks, the debt of the country must be credible.

[notahacker]

I'd argue a solvent bank is one whose expected return on assets, adjusted for expected losses, is sufficient to ensure it can repay money lent to it at the discount window rate [for all plausible trajectories of the base interest rate over the lifetime of its outstanding assets]

There's no logical reason why a central bank wouldn't lend to such a bank at the discount window even if some irrational hysteria caused its depositors to withdraw en masse, or why another bank not suffering from depositor hysteria wouldn't buy its loan portfolio.

The problem of national governments' economic policy lacking credibility is largely orthogonal[1]; central banks that underwrite private banks print money rather than borrowing it

[1]except to the extent really inept inflation-boosting fiscal policies compel the central bank to make aggressive and unanticipated interest rate rises that drive banks into insolvency.

[tspareme]

The issue is expected loss. No one knew what is the expected loss in a crisis. This is why some of the Lehman debt holders actually made money from the bankruptcy. This is why TARP actually made money. It is because there is a substantial difference in asset price while in crisis and not in crisis. The asset price is what is making the bank solvent.

Of course, the central bank (unless you are locked into a monetary union of course) can lend freely during a crisis. However, it must also be careful as to not trigger inflation or worse yet cause people to lose faith with your currency. There is also moral hazard as well but that's more of a soft issue.

The bigger issue is what happens if your bank liability is many times larger than your countries GDP. This was the case with Iceland or Britain. Then you can't print enough money to make your bank whole.