I’m not super well informed about this space, but my understanding was that SVB’s issue wasn’t that its depositors were from the tech sector, but rather that it had put a ton of capital into investments that are significantly less valuable (on the open market today) now than they were a year ago due to increasing interest rates.
I’m curious how the tech sector matters here specifically?
The tech sector has been withdrawing money in aggregate (because they continue paying payroll and rent but haven't been raising as much money recently) which increases liquidity pressure, and furthermore tech depositors are more skittish due to experience with FTX and thus prone to panic bank runs. They also have larger average account sizes, way beyond FDIC limits, which makes them more likely to need to withdraw money to protect their cash.
SVB did have some issues with losses but they likely were still solvent; the bigger issue was just a lack of liquidity and a sudden bank run - 45 billion (out of ~175 billion in deposits) was withdrawn in a single day before they ran out of liquidity.
It’s possible to void past transactions, but usually due to fraud, favorable preference, or insiders. I think the preference rule only applies if it is the borrower (the bank) who decides who is getting paid. Making good on transaction obligations in the order they occur doesn’t seem like a decision of the bank.
Having a 90 day window before insolvency to void transactions would probably smooth out the insensitive to make a run on the bank and give the bank time to find funding. More people would keep money in on the belief they already missed the window, perhaps enough for the bank to to stay solvent. But if that money was used to pay salaries you’re now forced to extract it from people living paycheck to paycheck. Incurring an instant debt for the depositing company would be worse than losing access to cash and even more companies will be sent insolvent.
Instead I think banks should have to get full deposit insurance, if the risk premium is too high then maybe they should consider restructuring to be less risky. Instead they get a free ride by shunting risk to depositors which incentivize the banks to leverage up to the max risk they can get away with.
The bankruptcy code provides guidance for a 90 day look back period from the date of insolvency where transactions can be clawed back or must be repaid. The FDIC will talk to the recipient bank and get a court order to deposit the funds in to a trustee account so that all available assets can be distributed evenly and fairly.
They are very unlikely to do that to normal individual account holders. It would undermine the purpose of providing the deposit insurance in the first place.
I would bet many accounts had balances well over the $250k cap and had larger withdrawals that, if left completed, would reward the first movers in the run on SVB to benefit to a greater extent than account holders that waited.
I agree that people below the threshold won’t see claw backs.
I might be wrong here. I think I'm confusing it with their ability to claw back executive pay. Under bankruptcy the court can absolutely claw back withdrawals, but an FDIC takeover isn't exactly a bankruptcy.
The tech sector isn’t an issue in itself, it’s that (1) all their deposits all came in at once because it’s one sector, so there was a huge demand surge for deposit interest, leading to a supply shortfall of loans they could issue and hence kinda desperately parking the money somewhere, which ended up putting their risk balance off kilter (bonds with interest rate risk); (2) all of their depositors (aka creditors) talk to each other and listen to the same people, so bank runs happen really fast. Compare this to First Republic bank: demand for deposit interest does not surge dramatically because there is finite liquid cash needing to be deposited and so one sector getting a cash infusion comes at the cost of another. It smoothes out. Plus their customers don’t all talk to each other and behave like worst-case bank runners.
All sectors are pretty highly correlated in the cash they have on hand and how they behave with it. It would be equally risky to be a bank that only deals with oil companies. Nevertheless it offers efficiencies for acquiring new customers and new business, so banks do it.
The big concern is the contagion spreading. All banks are vulnerable to failure if deposits are taken out quickly, especially in this market where assets have taken a huge hit over the last 12 months.
FRB wants people to believe the contagion is limited to tech and that they have limited exposure to it, so that folks don’t take deposits out en masse. If people start to think that it’s unstable, then they’ll take money out and it’ll be a death spiral like SVB.
If people don’t get their access to their cash by Monday (or a week later at latest), then any rumor of liquidity issues at any bank will lead to a bank run. Once you know / experience it, bank runs feel like a legit thing to do.
I know people who had their checking and savings at Washington Mutual when it failed in 2008.
The FDIC arranged for Chase to take it over. Nobody lost any money, and nobody I know had any problems with checks clearing or withdrawing cash at ATMs or paying their rent or mortgages.
I think the demise of Washington Mutual is still considered the biggest bank failure in US history, it happened while quite a lot of other bad things were happening, yet depositors ended up being totally fine. The systems in place for this are actually really good, and the FDIC is fast and competent.
The FDIC almost always pays insured depositors within a few business days of a closing, usually the next business day. Payment is made either by providing each depositor a new account at another insured institution or by issuing a check to each depositor.
FDIC says SVB customers will have access to insured funds on Monday, so seems like the answer to your question of how long it takes is “next business day”.
No rational reason, however during bank runs rationality tends to go out the window and even banks that are perfectly safe on paper can find themselves in trouble.
If First Republic could say they could meet all depositor demands, they would have. But instead this is the most they can say to stave off a run.
And to their credit, the fact that they're not tech heavy means the tech firms pulling from SVB aren't likely as scarred and pulling from First Republic. But First Republic has a ton of HNWIs too who are not insured.
This isn't Bank of America serving mom and pops, it's rich people who will pull and are uninsured by FDIC.
Matt Levine reports that the reason they had to buy those low-yield investments that plummeted is because it’s tech-sector customers had too much money in the boom times. Too much deposits means they need to buy a lot of something, and in the boom times that was low-yield stuff.
Sort of the other way to handle it would be to say “we can’t find risk-free yield for the volume of cash we just had deposited, so deposits now get 0.8% instead of 1.0%”
Which is kinda fine? Means you might lose some business as others chase yield. But I feel like most startups don’t actually have that much cash in the bank so they shouldn’t really be chasing yield anyway.
It didn’t take a genius to predict interest rates were going to rise. Locking cash away for 10yrs in very low % return vehicles seems stupid?
> It didn’t take a genius to predict interest rates were going to rise. Locking cash away for 10yrs in very low % return vehicles seems stupid?
Everyone is a genius in hindsight. You could have made millions out of a few thousands if you were able to predict an interest rate regime change. But where are your millions?
I locked in a 30yr fixed mortgage at 2.49% right before the spike because it seemed obvious? Not sure if that counts, but that should be enough cred that I made a large bet on my prediction when it mattered.
Yeah plenty of folks I know felt similar, myself included. I bought a home during the pandemic. Prices had increased a fair amount in my area which worried me, but the ability to lock in such a low rate for 30 years more than offset that concern.
you don't need a to be a genius to have known that the 2.4% interest rate won't last for ever. Everyone in america with a mortgage knew that and rushed to re-finance. So why did bank managers not know. i think its more like, they had no other good options to park their cash. So they took the best of a bunch of bad options
They would've had to go to 0%. As cormacrelf mentions, short-term treasuries in 2021 were yielding < 0.1%. That's a pretty hard sell when inflation is running 8-9%.
Some banks did exactly that. Even now, when SVB is advertising 4.5% rates on business checking [1], First Republic Bank is offering 0.01% on Business Interest Checking [2]. But note that SVB's failure impacts all sorts of household names like Roku, Roblox, Coinbase, Stripe, while FRB reports that tech is only 4% of their business. Companies that don't offer yield lose out to companies that do in the yield-chasing competitive marketplace.
We're observing some form of anti-survivorship bias, where risky behavior was incentivized by the market, so market participants had to engage or get pushed out of the market, and so now we hear about the risky behavior because that's what failed. We're not talking about banks like FRB or Wells Fargo that offer 0% on their bank accounts.
That’s a sign of mismanagement if true. Commercial/retail banks (as opposed to investment banks) usually make basically all their money on lending. Deposits (as liabilities) need to be so low risk that they basically don’t do much more than break even.
Back in 2021 short term T-bills gave you less than 0.1% yield, hitting an all time low of 0.04% for some timeframes. The bonds they bought gave them 1.5% annual return. The banking regulations didn’t prevent them buying the ones they did. So yeah. That’s why.
I'd rather have a bank do that, than go and invest in subprime packaged mortgage loans. Except SVB got punished for investing in long-term maturation US Treasury
I'd also like to point out that VCs were telling all their portfolio companies "Pack it in and conserve cash as you're not getting another round." That caused a lot of companies that would otherwise have put their cash into an investment vehicle to instead hold it in their accounts.
It will also be interesting to see if someone is going to wind up in jail for initiating the bank run. Someone shared confidential info that started this whole thing.
Was it confidential info? The public earnings call was three days ago, investors didn’t like what they saw and the stock went into free fall, then the run started yesterday.
I think insiders might not even have realized how bad their balance sheet looked. On the earnings call the CEO was talking about how much he likes cycling to de stress, etc
It matters because recent underperformance in the sector led to the initial deposit outflows in the first place. If you’re concentrated heavily in a single sector, you’re likely to see higher correlations in the behavior of your depositors.
Every asset is a risk. Banks should have right to do exactly two things. Keep their customers saving and issue loans. There's plenty of risks even in that activity. Every other thing bank does is just piling up risk to unreasonable levels.
The difference is, who created the risk. When you buy asset from someone else you are buying the risk someone else has created. So you have more risk. Banks are very unique institutions that are allowed to do things no other business is allowed to do. They also should be equally severely restricted so they can't acquire any additional risk they themselves didn't create.
Banks are risk sources in the economy, the risk of "what if borrower doesn't pay in full or at all", the law should do everything to prevent banks from being risk sinks in the economy. Because in current setup banks just buy more an more risk and they eventually collapse one way or the other when risk is en masse converted to cost to the people and the economy.
Issuing a loan is buying an asset. You give the borrower money and in exchange buy their promise to pay it back with interest. And in fact that’s what SVB did, except instead of originating the loans themselves they bought them from the original lender.
To buy good loan you have to check whole supply chain. To issue a good loan you need to check only the borrower and you have every incentive to do so if you can't offload the loan to someone else after it's created.
Exactly. And it's a work that bank has to put in if it can't sell the loan. But if the bank can sell the loan it becomes "check the borrower or bullshit the buyer" and the banks decided latter is less work. And that's why 2008 happened.
> What's really the difference between issuing a loan and buying the same loan issued by someone else?
Risk. When you are issuing a loan you, yourself are in control of the risk you have created. When you buy, you can easily buy massive amounts of risk someone else have created and you can easily be overwhelmed even when you try to do industry standard, responsible, "safe" thing.
I think this was a weird thing where deposit demand was so much greater than loan demand. They had too much money and thought they were doing a safe thing. But failed to consider the time aspect and how rapid interest rate increases nuke them.
Banks do provide a valuable service and having them reject deposits isn’t a solution.
Glass-Steagall only prevented commercial banks from owning non-investment-grade securities - from what I have read, SVB’s portfolio (of Treasuries and MBS) would’ve been entirely consistent with pre-repeal Glass-Steagall.
SVB would have failed even it held only Treasuries on its book. Those are literally used to define the risk-free rate of return. The assets were not risky. It was their duration mismatch that was the problem.
That’s also totally wrong. What do you think the financial difference between originating a loan and buying a bond is? You pay out a principal sum, then (as long as the borrower is creditworthy) you get back the principal and interest.
Unless you only make floating rate loans (which is extremely rare outside of revolving lines like credit cards) then you have exactly the same duration risk problems and you’re actually exposed to much greater credit risk with a loan than you are with a government bond.
To issue a loan you need to individually check the borrower, and that's it. With bonds you put your trust in the whole system that banks apparently (as SBV debacle shows) don't understand (bacause they can't or because they have incentive not to understand).
Also, floating rate loans and mortgages are super common. Maybe not in US.
I’m curious how the tech sector matters here specifically?