Given these implications, the policy choices are easy. Why allow these long term consequences in the name od stiffing depositors?

I believe the mechanism he's talking about is that large, uninsured depositors will wire funds out of regional banks that are "small enough to fail" to more diversified national banks that are designated "systemically important banks", which means they're too big to fail.

The MBS will still be worth their face value at maturity.

They didn't buy MBS, but they have very similar assets/liabilities. They have $17B more "assets" than liabilities, but $166B are in loans and only $4B in cash, meanwhile they have $176B in deposits that are about to be withdrawn heavily. They are going to need to liquidate those loans and you want to guess how much of a % haircut they're going to take on a liquidation of that scale, for similar assets to the MBS SVB purchased last year?

This is a very different phenomenon than what happened in 2008. In 2008, banks owned a different kind of MBS that was poorly-underwritten, poorly documented, and truanched in a way that made their value extremely sensitive to various model assumptions. This made them extremely illiquid, meaning that if you tried to sell them in volume you would have to sell at a large discount relative to the value of the expected discounted cashflow of the security. (This is not true of 2023 MBS. These MBS are a totally different species. In 2023 rates rose, the value decreased, but we can be extremely certain of the value and they are extremely easy to sell at little discount to this value).

Contagion happened in 2008 because when there was a run on bank A, bank A had to sell its illiquid MBS at a large discount. This reduced/made uncertain the value of bank B's similar MBS, which triggers a run at bank B. In that sense, the bank A run causes the bank B run. There's no spillover mechanism in this 2023 scenario: SVB's selling its treasury or MBS portfolio doesn't meaningfully impair some unrelated bank's assets. To the extent that some unrelated bank is in trouble, it's because they face correlated macro shocks, not because there's a causal spillover.

> Consumer deposits have an average account size of less than $200,000 and business deposits have an average account size of less than $500,000

… and far less of their assets are likely to share SVB's duration exposure:

> The investment portfolio is less than 15% of total bank assets.

So instead of being able to hold to maturity, SVB had to sell assets they bought for $100 for $80. Perhaps they could have hedged against interest rates rising better than they did or had a more short term securities or capped account balances. But the fed could have been faster to act in raising rates to allow for more gradual in raising rates as that would have been less disruptive.

Now the Fed should set up a program to take these long term assets off of member banks balance sheets. Do so at a discount, make the members pay for their bad investments but prevent failures.