The issue is not that SVB has loaned too much to companies that can't pay it back.

The issue is that SVB took in a lot of deposits in 2021-2022, and locked them into 10 year bonds at 2.5% interest. Now in 2023, a lot of their customers are burning money (withdrawing funds) and aren't raising money (depositing funds), and so they need to start liquidating the 10 year bonds.

The issue is that the 2021 bonds are discounted because interest rates have grown since, and they're out some money unless they can hold the bonds to maturity. So they're raising money to bridge the deficit to allow themselves to hold more of the bonds.

Right, I get the loans don't feature right now, but the reason why depositor behavior is so concentrated is because they're mostly in the same industry affected by the same market forces. To have large numbers of loans out to that same industry at the same time is additional risk, whether or not that particular risk caused problems this time.

EDIT: Elsewhere in the comment thread people have clarified that the bonds are the big problem on the assets side. They had a bunch of low interest rate bonds that now have to be sold at a loss because interest rates are much higher.

Why would they buy 10-year bonds when their depositors are startups, many of whom won't exist in 10 years? Seems like shorter-term T-bills would have been more appropriate, but I know next to nothing about finance.

Because the yield curve is usually upward sloping.

Banks lend for long periods on the expectations that deposits that are withdrawn by some customers are balanced by new deposits from others.

But of course in the current environment there aren't a ton of startups depositing recently-raised funds.

They made a gamble that if they needed to redeem them early that they would be able to do so without much penalty and that they wouldn't need to redeem too much if any (new deposits would cover old assets).

They lost on multiple fronts, interest rates rose so much that it became a huge loss to sell these bonds, the tech market slowed down meaning that fewer companies were getting infusions of cash and more of them were burning through their cash piles and the final death knell was that word of all this got out and it started a bank run.

> They lost on multiple fronts, interest rates rose... the tech market slowed... word of all this got out

The first two are the same thing and the latter is not really an "event" but rather the legally-mandatory observation of the first two? (Unless they want to commit fraud, of course.)

ISTM the issue was that their core product was as shallow as their customers'.

Shorter term bonds typically yield significantly less. Except the past year or so we’ve had “yield inversion” where the shorter term bonds are actually paying out more than long term bonds. This is an indicator of incoming economic downturn as it indicates institutions need cash now and are willing to pay for it.

Right, Talking to some people in the banking industry and the fear is no one know which other banks are in similar situation.

hint: every bank

Most banks dont have the same cash flow in/out profile that SVB are facing thus wont have the same issues.

It was a major red flag! You're right about their liabilities, but their assets are mostly longish duration bonds. When interest rates go up their main customer base (startups) will have a harder time raising money (bad for them) and the value of their bonds will go down (because new bonds offer more interest). And now they're forced to sell these bonds at a loss to cover the money outflow (customers leaving). They're deep in the hole with no way out.

Basically SVB failed basic risk management, and they deserve to go down. You can't promise your customers they can take their money out at any time and then turn around and invest that money in longer duration assets you can't liquidate. Just irresponsible.

Execs selling millions worth of shares in the past weeks is also classy.

So a slow version of Terra Luna.

Resilient system design doesn't just apply to engineers.

it's a totally different situation to Terra/Luna

The Terra/Luna system would work very well when Luna's cap >> Terra.

The SVB system would work very well if the interest rates stay low, when they could liquidate their 10 year bonds for same value (and their start up would be more likely to turn profits as well).

These are all systems designed with assumption that turned out to be false.

SVB has an extremely low loan to deposit ratio (~.45). Most of their assets are in the form of treasuries, bonds, etc.

Depositor concentration is definitely high but that’s not particularly uncommon in banks (though for one of this size it is). But that concentration is what fueled their growth as well.

> Most of their assets are in the form of treasuries, bonds, etc.

I believe you know what you're talking about, but if that's the case why are they struggling to raise capital to pay depositors?

(I'm asking as an ignorant outsider to banking in general and SVB in particular.)

What doesn't seem to be so explicitly stated is that due to rates going up, the bonds are worth less than they were when rates were high. So they wouldn't have enough money if they turned the bonds into money.

Guy lends you 100 at 3%, you buy 100 of bonds than pay you 5%. Guy asks for his money back, your bonds are worth only 80, big problem.

One of the biggest banks in the world failed to account for a capital management risk scenario where interest rates go up? Insane

Practically every single bond - which banks effectively have to buy and hold for regulatory reasons - has reduced in mark-to-market value every single day the last 2 years.

Personally I think they are nuts for having (reportedly) such a long average duration of bonds in their portfolio, but they were also paying a pretty decent interest rate on savings to their customers - that money has to come from somewhere.

Silicon Valley Bank was nowhere near one of the biggest banks in the world.

https://en.wikipedia.org/wiki/List_of_largest_banks

Not even amongst the largest banks in the US:

https://en.wikipedia.org/wiki/List_of_largest_banks_in_the_U...

They're #18 on the US list you linked, I saw other articles saying they were #16 so it probably just depends on the date assets were measured.

The falloff is extremely rapid. They're closer in both an absolute and relative sense to the 40th than even the 10th.

Looks I said high when I meant low. For people who aren't in finance, if rates go up, bond prices go down: an instrument that pays 10% a year on 100 is worth less when there are new instruments paying 15% a year on 100. For the rates to match the original 100 is now lower, because that's the only way you get the same amount of interest per dollar spent buying the bond.

Treasuries and bonds drop in price when interest rates go up.

in 2021 (when interest rates were low) you might pay 100 cents on the dollar for a treasury that payed 2.5% interest. In 2023, with interest rates much higher, you can no longer sell that treasury for 100 cents on the dollar. People can get much better returns by just buying recent treasury. So you need to sell your bonds lower, at say 95 cents on the dollar. That is an immediate 5% loss.

This is known as 'interest risk' and is really the only risk you have when buying safe bonds like treasuries. The risk is completely avoidable if you can hold your bonds to maturity, so it only matters in a liquidity crunch.

It's more like 50 cents on the dollar.

Interest rates went up. That would be fine if they could hold the bonds until maturity, but to honor withdrawals now they are forced to sell at a loss.

Right! Okay, that makes sense. So the failure on the assets end of the balance sheet was relying on bonds that were purchased at times of historically low interest.

Also, in more normal time they could have put more deposits into shorter term debt that would not have had this problem. But rates were so low that would have been very unattractive.

They factually lost 1.8B on 21B worth of investments, this is an 8.5% loss. Unclear why there's conjecture like previous comments saying its a 25% loss.

They did the right thing by selling out now -- likely because they see more rates raising in the future which would decrease it, so they took the liquidity and then raised a bit of money to cover up that loss (raising 2.25B)

No, they did the wrong thing by selling so late.

Agreed -- but are you saying they shouldn't have sold at all? I don't think you can say it was a "wrong decision" at the moment, I agree that it was too late. But you could also go back and say they shouldn't have invested in the long term assets in the first place.

Or you could go back and say they shouldn't have accepted the deposits without having short-term gain solutions in the first place.

I'm saying they shouldn't have bought what they sold in the first place.

The eye-opening number was 70% equity to deposit ratio. They were apparently betting the farm on the "there is no alternative" narrative.

VC's are generally the ones that actively promoted SVB to their investments. There's literally a branch on Sandhill road.

By investors I'm not talking about venture capitalists, I'm talking about the stock market. I'm just surprised that people are just now realizing how bad it is to have the lion's share of Silicon Valley's financial resources centralized in one institution that doesn't have much in the way of other business.

Their success is due to their tight focus though isn’t it? They’ve been around for 40 years and deeply entrenched in the VC ecosystem.

I am very curious whether their poor risk management is new, or if they’ve been playing fast and loose for much longer and only blowing up now.

Since 1980 interest rates have only really ever gone down. Buying long dated bonds at X% and having to sell them a year later at X-1% actually makes you a capital gain.

Old habits die hard.

What could possibly go wrong running a bank by working with the most volatile companies on Earth with failure rates above 90% lol.

They aren't loaning money to the startups (well, mostly). They handled the banking. This current problem has nothing to do with their clientele, it has to do with long term bonds.

They sell banking services to startups... all companies need some way to manage payroll (for example). Regardless of the balance sheets of their clients, they should have been able to manage assets such that they could fulfill day to day banking needs.

It's that failure to manage assets on the bank side that we're talking about. Startups/clients failing should be a normal event at SVB, given their focus.

As far as I understood it, the 10% which paid off made SVB into the type of bank it is today.

They obviously made a bad move in 2020-21 in thinking the public market was going to crash and what would happen to bond yields, but so did lots of people.

If you think SVB is the only bank which is making high risk decisions to get payoffs, I would suggest you level up and look at what ALL these banks are doing. Faith in the system and long-term risk assessment isn't built into the mentality of banks which have so much invested in equity trading vs. boring deposits.

The financial stability of the depositors had ~0 to do with this situation.

This is completely wrong. This was a classic bank run caused by the startups banking with SVB switching from net depositors to net withdrawers as they aren't able to raise in the current environment and obviously none of them actually make money. The withdrawals start, there was a liquidity and then a panic and boom bank dead.

This is true. most startups are trying to bank with the company I work at now and switch from SVB to us, it is a fucking panic and some of these companies are trying to send us 100's of mill cash in hand

> Is this true? And if so, how was this not a major red flag for investors earlier?

The red flag being......?

There is nothing per se that makes this situation unique to startups.

This would happen to any bank that (1) made bad investments (2) caught the public eye.