[jjeaff]

No, I understand the liquidity risk involved in having too much tied up in long term treasuries. But I am yet to see evidence that any bank could have withstood a run of that magnitude. Nor have I seen much evidence that most other banks have significantly less liquidity risk than svb did.

[tripletao]

It's solvency risk, not liquidity risk. When interest rates are 4%, having $1000 ten years from now means I have 1000/1.04^10 = $676 now. If my current liabilities exceed that, I'm insolvent, not illiquid.

Liquidity is about bid/ask spreads, disorderly markets in which the price becomes temporarily irrational. The SVB's problem was simply the time value of money, that the liquid and economically rational price of their long-term bond-like assets is lower than they wished it would be.

This paper estimates that 190/4800 ~ 4% of banks would have deposits at risk if half of uninsured deposits were withdrawn. That means 96% of banks wouldn't. The SVB's situation wasn't completely unique, but it's far from the norm.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4387676

[LapsangGuzzler]

> Nor have I seen much evidence that most other banks have significantly less liquidity risk than SVB did.

In traditional banking, rising interest rates are a good thing because it means that banks in turn get to underwrite loans at higher interest rates, which positively affects their bottom line. SVB's problems were twofold: A) they had a one-dimensional investment strategy that was adversely affected by rising rates, and B) outstanding loans made up a very small portion of their business relative to their size, which made it so that they weren't able to capture meaningful value from rising interest rates. The latter is actually pretty rare for a bank, which shows how uninterested they were in actually functioning like one.

[prottog]

> But I am yet to see evidence that any bank could have withstood a run of that magnitude.

I think you're right that no bank can withstand a run of that magnitude, and it has been pointed out elsewhere in the thread that entities that can do so aren't really a bank anymore. However, the bank run only occurred because it was public knowledge that SVB had terrible duration risk, so it's somewhat of a chicken-or-egg problem.

All the findings that came out since then points to minimal duration hedging on SVB's part, so all in all it sounds like a bank that was poorly managed, perhaps adapted too well to a ZIRP world, and was ripe for a run to happen.

[tripletao]

> I think you're right that no bank can withstand a run of that magnitude,

They're not right. No bank could sell assets quickly enough to withstand such a run, but that's why the Fed serves as lender of last resort. Even under previous policy, the Fed would lend against the mark-to-market value of the collateral. So in theory any MTM-solvent bank could survive any run, so there was no incentive to start the run in the first place.

The SVB was MTM insolvent due to that excessive duration risk, so it couldn't do that. It's not the only MTM-insolvent bank (see the link in my other comment), but that in combination with its unusually high fraction of uninsured and thus flighty deposits was apparently enough to start the run.

[jjeaff]

Thanks for the posted paper. But if SVB was able to borrow against the mark to market value of their collateral, I don't understand why they couldn't remain solvent. Because they had plenty of assets, even at mtm value to cover a 50% run. Even all the way to nearly a 90% run, it seems.