Hacker News Reader

Comment by mkmk | Hacker News Reader

Oldie but goodie on what actually happens when a bank is taken over. https://www.npr.org/2009/03/26/102384657/anatomy-of-a-bank-t...

View on HN

This case is unique because of the sheer volume of non-FDIC insured deposits. Substantial risk of depositors not being made whole for a while, they’ll probably get all their money but it will still be bad

It's a little unique because it's a top 20 bank failing. As far as I can tell, it's not super unique in terms of % of fdic insured deposits.

As a whole it's like <30% FDIC insured in the whole US banking industry.

> they’ll probably get all their money but it will still be bad

They may not. Even for banks where most assets are FDIC insured, you'll see that not 100% of deposits are returned. Selecting a random recent one:

https://closedbanks.fdic.gov/dividends/bankfind/Dividendinde...

~95%.

Looks like 2.7% of their deposits were > 250k. So yes, it seems relatively unique (there are only 5 banks on this list with <15%).

https://twitter.com/GRDecter/status/1634208652595699713?s=20

You've got that backwards. Reread the tweet you linked to:

"Only 2.7% of SVB deposits are less than $250,000. Meaning, 97.3% aren't FDIC insured."

Not a good situation at all.

Aren't fully FDIC insured, yes?

Yes - if you have 5M$ in the bank, you'll have 250k$ of that insured. Not that comforting.

But if you put in $250k because you're responsibly splitting things across banks, and the interest you've made on it raised you to $252k, you're counted amongst the 97.3% and it's not a big deal.

I'm sure some people are out meaningful amounts of money, but I'm not sure what a typical account looks like.

It's 97.3% of deposits ($), not _depositors_

Hm, you may be right, though "deposits are less than" is a really weird way to phrase it in that case.

> Substantial risk of depositors not being made whole for a while

Emphasis on for a while. Stock and bond holders (man, did the latter miss the ball) will absorb losses.

As of Dec 31 the total liabilities were $195bn.

Given that they had only $5bn of long-term debt - and $173bn of deposits - it seems that bond holders cannot absorb much (even though it's not very likely that they receive anything).

That's $23 billion of non-deposit liabilities that will be junior to the deposits.

Possibly forever, if the assets aren't there. A senior claim on $0 is still worth $0. There may be sufficient assets to pay back depositors, but we don't really know at this point.

What stock holders? Market cap was 2.5 billion pre-market, a small fraction of the 175b of deposits (as of Dec 31, 2022).

> market cap was 2.5 billion pre-market, a small fraction of the 175b of deposits

Market cap doesn’t matter. Balance sheet equity.

All their money? I doubt it.

SVB owned a bunch of mortgage backed securities and treasuries - both of which are down 30-50% from their peaks, which coincides with our last venture capital boom. SVB was buying at the top because they had so much capital to deploy.

Maybe instead of cashing it out, depositors could be given a treasury worth 30% less than SVB paid for it, that pays out 1% a year in interest, redeemable at face value in just 20 short years.

They weren’t insolvent yesterday, they just had a rush of withdrawals. It will probably be upwards of 90 cents on the dollar, but that amount of money locked up in a possible credit crunch coupled with a loss of confidence in banks and the current startup environment is very bad

They weren’t insolvent because banks are allowed to pretend treasuries haven’t lost value so long as they intend to hold them to maturity.

Across all banks, there are over $600b in such losses. If they can hold those securities until they mature, all is good. If they need to sell them earlier, the losses become real.

https://twitter.com/grdecter/status/1634091448743219201?s=46...

In SVB's case I believe they are mostly mortgage backed securities (ironic, eh) rather than treasuries but yes. I wonder if this will spur any reflection on the accounting rules for HTM securities.

I bet they bought mortgage-backed securities because returns on everything safer were essentially zero. Crazy-low interest rates for too long created this problem.

> If they can hold those securities until they mature, all is good.

That's an odd bit of logic. If interest rates stay the same and nothing of interest happens for the next 20-30 years, then those banks will lose the spread between the interest on those instruments and the interest they're paying (SVB was paying 4.5% on savings!) for 20-30 years. That money needs to come for somewhere. The net present value of that loss is a very similar number to the unrealized mark-to-market loss of value of those instruments.

It's almost tautological that being able to hold those instruments to maturity requires that they remain solvent for the term of those instruments, which means that the money to cover their losses must come from somewhere.

Now everyone involved can gamble that interest rates will go back down in a few years and those short interest rate positions will recover, but that's a gamble, not a certainty. Of course, the Fed does have a bit of an interest in pushing rates down if needed to avoid bank failures...

> They weren’t insolvent yesterday

Didn't management/CEO/someone high up in the company say something like "We're safe unless everyone pulls out their money" the other day, indicating that they were de facto insolvent?

That just indicates they didn't have liquidity to cover every deposit, which is true for basically every bank that doesn't charge you to store your money in a vault.

Which I guess makes it about semantics. If everyone takes out there money one at a time, one person each week, over years, they would be solvent. But if everyone did it at the same day, they wouldn't be able to handle it, making them insolvent.

i think by that definition all fractional reserve banks are 'de facto insolvent', which would make this a definition of 'de facto insolvent' that isn't useful for fractional reserve banks, since it doesn't distinguish among classes of fractional reserve banks

and yes, the ceo did say that yesterday, which is likely why this happened

Yes but supposedly if you mark their balance sheet to market, the net is approximately zero, i.e. they have just enough assets to cover liabilities with no reserve. In this case you could say that the reserve requirement is serving its purpose. The stockholders may be wiped out (depending on how much a potential buyer values the goodwill), but the depositors should get all their money back. Worst case I think depositors would take a small haircut.

If they have to liquidate their assets (mbs/treasuries) they’ll probably create at least some slippage, moreso in MBS, that may prevent them from recouping their value at current prices. Alternatively they can try to hold to maturity or spread out their selling but that will make depositors illiquid

I don't think that holding to maturity really solves the fundamental problem though. The reason the MBS have lost value is because they have a low coupon compared to what you can get now. To execute the hold to maturity strategy, they would have to pay the depositors interest with the coupons from the MBS.

Under ZIRP, the depositors weren't getting anything, and so making 1.5% on MBS was fine. In the current interest rate environment, depositors won't be satisfied with a zero yield on their deposit accounts, and the MBS don't pay enough to cover it, so either they lose depositors because they're not paying competitive interest, or they take a loss every day because their investments don't cover the cost of the deposits.

Ultimately keeping the bonds on the books as HTM just spreads out the loss over the lifetime of the bonds rather than recognizing it right when interest rates change, but the result is the same.

Oh fully agreed. They fundamentally lost depositors’ money by making bad investments - holding to maturity is just a way to argue that depositors will be made whole. The depositors may as well just be given the bonds/mbs themselves and allowed to hold to maturity or sell them to meet immediate needs.

If you mark to market hold to maturity assets for any bank there are hundreds of billions of loses, fyi

Yes so the key distinction is this bank had a bank-run against it that forced their hand. Fair to say any number of banks would fail if they faced a bank-run in current environment (depending on proportion of long dated bonds they hold)?

Have there been any cases in modern times, in the last 25 year or so where depositors lost money because they had more than the fdic covered ?

50 cents on the dollar at IndyMac in 2008

https://www.depositaccounts.com/blog/indymac-depositors-are-...

Yes it happens literally every year. It's not usually a huge %, like it might end up being in this case.

SBF was a piker.

I know a couple bank VPs who have been the representative of the take-over bank in absorbing failed banks. They describe it in terms that most of us would recognize as taking over a failed project and re-architecting/refractoring it so that it can succeed (or making the call to discard all work to date and start over with a workable solution).

From the looks of the news stories, it looks like the FDIC isn't as efficient in taking over a bank as it used to be:

https://nypost.com/2023/03/10/nypd-called-to-silicon-valley-...

I wonder if the FDIC is just as efficient but this situation is different.

Possibly a forced rush job after that comment about solvency. I think the FDIC usually prefers to do secret operations under the cover of night to avoid problems. With everyone's eyes on the bank, that wouldn't be an option.

When the FDIC takes over a bank, they usually have some local cops around, to deal with crowd control and calm down upset depositors. See the 60 minutes video. There's a police car parked in front of the bank, but the cops are not doing much.

will companies with money in SVB even be able to make payroll? Seems like this could cause huge issues in the short term and long term

Pipe, Clearco, Founderpath, etc, are going to have a busy month.

Likely many companies won’t be able to do mid-month payroll.

Someone will setup a market for the FDIC warrant on uninsured deposits, but that will take time.

Who do you think backs those Clearco loans?

That's fascinating. I'd love to watch a documentary about that.

60 Minutes followed an FDIC takeover in 2009, including when the agents actually walked into the bank’s headquarters: https://youtu.be/TAE8i40A5uI

I just watched it -- aside from the very interesting mechanics of the FDIC coming in and taking over a bank, I thought Sheila Bair (the then FDIC chair) came off as very knowledgeable, realistic, and critical of the risks posed by big banks. She went on to become a university president after the recession / banking bailout.

That interview with the chairman (chairwoman?) is just great. She straight up says that maybe they shouldn't bail out the big banks and instead apply a similar process to them.

This 60 minutes segment from the Great Recession is the closest thing I know of to that. https://www.youtube.com/watch?v=TAE8i40A5uI

Are clawbacks a thing in this case for the people who did get [some] of their money out in time?

Clawbacks are only poised to happen for big crypto failures (FTX) because the people who held money in those institutions aren't technically "depositors," and are considered a more generic type of creditor (which has a bankruptcy clawback).

Bank deposits are special and there will be no clawback. However, depositors who lost money will be first in line at the bankruptcy court.

note there is no bankruptcy court here. SVB will go into fdic receivership and almost certainly be sold off to another bank. Depositors are still first in line to get stuff back, but it's not through bankruptcy court. There is no court involved.

Reply on news.ycombinator.com