> the bank purchased a large amount (over $80 billion) in mortgage-backed securities (MBS) with these deposits for their hold-to-maturity (HTM) portfolio. Almost 97% of these MBS were 10+ years in duration, with a weighted average yield of 1.56%.
> with the rise in Fed rates, the value of SVB’s MBS plummeted. This is because investors can now purchase long-duration "risk-free" bonds from the Fed at a 2.5x higher yield. Precisely, with the rising US Fed interest rates, the value of existing bonds with lower payouts fell in value.
The second paragraph explains what the original article summarized as "The trouble is that when rates started to go up, mortgage assets got hit hard".
1. Bank buys lots of mortgages (well, assets backed by mortgages) at 0 interest rates. Let's say those mortgages are $100 each.
2. Fed raises interest rates. Now new mortgages are shiny (they return more money) and those old mortgages are shitty, so now they are worth $75 ea. Normally not a problem as they can wait for those mortgages to mature and get $105.
3. Tech sector experiences a crunch. All the sudden more and more companies are taking money out of their bank account.
4. In order to process withdrawals, SVB is now forced to sell the mortgages it bought for $100 for $75. If this continues SVB won't be able to find enough money to pay depositors.
> 2. Fed raises interest rates. Now new mortgages are shiny (they return more money) and those old mortgages are shitty, so now they are worth $75 ea. Normally not a problem as they can wait for those mortgages to mature and get $105.
It's not that new mortgages are shiny and old mortgages aren't; it's that $100 in 5 years (or whatever) is now worth less. It's easier with zero coupon bonds, because there's only the maturity, so a $100 zero coupon bond that matures on date X is worth $Y, regardless of when it was issued; and $100 mortgage that completes on date X needs to have each payment adjusted for current value. Clearly, a smaller series of payments is worth less than a larger series of payments, even though the principal amount is the same.
Yeah really its based on long-term t-bills and the risk premium.
Because safe interest rates went up, people can get a better deal on long term debt unless they get a discount on the bonds from SVB so that in the end the SVB bonds work out to a the current interest rates, if held to maturity.
This is how bond prices move opposite to the external interest rate environment.
There is also a risk premium since t-bills are considered perfectly safe, while these mortgage backed securities are likely now carrying a risk premium so buyers want an even higher effective interest rate out of them.
It isn't really new-vs-old, it is just the interest rate environment. If there was zero new issuance of debt people would still be trading the old debt at the same interest rates and the value of the bonds held by SVB still would have been underwater. Of course if there wasn't enough debt to satisfy the appetite for buying debt then the prices would rise and get bid up, but we'd see this as a lowering of interest rates in these markets (which isn't happening).
> SVB is now forced to sell the mortgages it bought for $100 for $75.
Not quite. The article addresses this in detail. They bought mostly hold-to-maturity (HTM) bonds which could not be sold.
(Technically such bonds could be sold, but selling would trigger a portfolio revaluation of all their HTM bonds, which would have made them immediately insolvent.)
Here is a quote I found in the article: "The decision dictates whether the securities are designated as “held-to-maturity” (HTM) assets " - HTM was an internal designation - an accounting convention. It does not have anything to do with the bond terms.
I'm not an expert. But it appears to be an accounting designation that shows up in filings, so it's not purely internal. You are correct that they are not part of the bond terms, which I was unclear on before.
"HTM assets are not marked to market... they remain glued to balance sheets at amortised cost regardless. By contrast, AFS assets are marked-to-market... Sell even a single bond out of an HTM portfolio, however, and the entire portfolio would need to be re-marked accordingly."
Essentially, when interest rates go up, current bonds temporarily fall in value as people can buy higher yielding ones. For a variety of reasons, SVB needed to sell the bonds at a loss now to cover deposit withdrawals, etc.
- FDIC has implied that insured funds (up to $250K per account) will be released within 7 days.
- It's entirely unclear how long it will take to recoup uninsured funds (above $250K per account).
- It's also unclear how much of uninsured funds will be recouped (although most people believe the figure will be above 80%)
If given the above fact pattern, your company can't make payroll, then it will have to raise emergency cash to do so. Whether or not that's feasible isn't an answer anyone here can provide.
this is incorrect: "FDIC has implied that insured funds (up to $250K per account) will be released within 7 days." Specifically, the FDIC has explicitly stated (rather than implied) that insured funds will be released on Monday (rather than within 7 days). https://www.fdic.gov/news/press-releases/2023/pr23016.html
The same statement also says "The FDIC will pay uninsured depositors an advance dividend within the next week." which I assume means some percentage of the uninsured funds based on a worst case scenario of the recovery.
Another issue is that this isn't just about the company not having funds, but also funds that may have been transiting through SVB today. At least one payroll provider was affected.
The FDIC's statement shared that: As of December 31, 2022, Silicon Valley Bank had approximately $209.0 billion in total assets and about $175.4 billion in total deposits
The question is what haircut the FDIC will take on the $209b when they forcefully liquidate all securities. If the above data is still accurate, the FDIC can make depositors whole as long as they don't take more than a 16% haircut when liquidating.
SVB recently liquidated their AFS bond portfolio at ~90c on the dollar. And, this traunch of securities is very sensitive to interest rate changes.
Their HTM bond portfolio consists mostly of securities that could be sold today for ~80c on the dollar.
Given the above, it seems likely depositors will be made close to whole. Then again, who knows what's hidden from public eye.
FDIC only covers $250,000 per depositor per bank per account type, so it depends on how many employees are making how much money. $250k is enough to pay about 20 employees who make ~$150k a year for another month, to give you a rough idea. So it depends on how many employees your employer has and how much they get paid on average. Also it depends on how much incoming revenue your employer has.
What isn't being discussed is that SVB was probably seeing an initial drawdown in deposits because VC funding has locked up and startups are just burning through cash and can't raise money. SVB was likely a bet across the entire startup economy that funding wouldn't dry up.
This probably foreshadows waves of startups failing as they run out of runway.
2.a They sold some of the assets bought at step #1 at a loss (dont know why).
2.b Flubbed the PR game while trying to raise to cover the loss
2.c Bank run!
If 20 people give you $1,000 each totaling $20,000 and you just know you can make $50,000 by holding it for X amount of years thanks to the economic gods, what happens if your $20,000 suddenly is only worth $5,000 thanks to you having to sell said asset much sooner than you would like?
This is bad. This is real bad. ...and I don't think we're going to know the depth of the badness for a while. Think the credit crunch that precipitated the 2008 crash. Though this will be localized to tech startups and it's only the kinds of people who read Hacker News who are going to be affected. A massive wrench has been tossed into the economic machinery of Silicon Valley and other tech hubs.
A bunch of companies now have no or little operating cash and all they're going to have on Monday is $250k each.
That’s not true. The FDIC will be disbursing money above and beyond the $250k to account holders. It won’t be the balance of their account but they’ll also get a piece of paper laying out how much more they’re potentially owed. As the FDIC winds the bank down, it will use that money to make account-holders whole. The big losers are SVB shareholders and bond holders, as they’re going to get the biggest haircut.
ADP isn't going to take that paper to cut paychecks. AWS isn't going to want that paper either. The best you can do is take that paper and get a loan with a stupid high interest rate and that's not going to happen quickly.
The Fed is having an emergency meeting Monday. My guess is they loan the bank-formerly-known-as-SVB as much as necessary to cover depositors needs for the next year, and get paid back as the banks assets are sold off.
> they’ll also get a piece of paper laying out how much more they’re potentially owed. As the FDIC winds the bank down
That will take months at least. In the meantime they'll have this non-liquid piece of paper. I suppose they could try to pay employees with IOUs, but I don't think that many employees would be able to afford to stick around for very long.
> It won’t be the balance of their account but they’ll also get a piece of paper laying out how much more they’re potentially owed.
Yes, they’ll get an “advance dividend” based on whatever surplus cash is available beyond what is necessary to cover the insured balances, plus a “receivership certificate” accounting for the rest, and, to quote the FDIC press release [0], “As the FDIC sells the assets of Silicon Valley Bank, future dividend payments may be made to uninsured depositors.” [emphasis added].
The FDIC will pay an advanced dividend next week. We have no idea how much that will be but it is probably conservative. Some companies may not be able to make payroll today or early next week. Some may not be able to pay their vendors for the things that allow their businesses to run.
It may only be a couple day disruption but what if your employees quit? What if your vendors cut your access and your customers leave?
What if you don't get made whole but get back 80% of your cash? Now your runway is 20% shorter in a climate where investors aren't around.
> It may only be a couple day disruption but what if your employees quit?
Even if they don’t, in California there is a statutory penalty payable to the employee of $100 to each affected employee for a first, non-willful late payment of wages (second or subsequent violations, or any willful or intentional violations, have a much greater penalty – $200 plus 25% of the wages not paid); that’s not a lot (if you didn’t happen to have a prior payroll hiccup bumping this into the “second or subsequent” category), but its still a hit.
> It may only be a couple day disruption but what if your employees quit?
I guess they don't have to worry so much about that what with all the tech layoffs over the last 6 months and with a lot of other statups in the SVB boat, who's going to be hiring?
Is there even time for some lending entity to do enough due diligence to figure out which of the startups are even worth gambling on? That could take weeks or months. In the meantime payrolls have to be met, leases and light bills paid.
You aren't gambling on the startup if the collateral is the receivership certificate / advance dividend. Those should be quick to verify. Any loans up to $250k would be even easier.
Some firms will make extremely low buyout offers knowing that at least a few businesses will have no choice but to take a massive unknown haircut or go bankrupt.
It's the indirect effects that'll get you though. Consider my company has nothing at SVB, but 50% of their customers do. Still gonna be pretty bad when half your clients can't pay you.
Yeah, but you're pulling that 50% number from thin air. How many companies do you think actually have 50% of their customers banked by SVB?
SVB had ~$200b in total assets, less than 1% of the total assets in banks across the U.S. alone.
There are definitely going to be indirect effects, but HN and other voices in early-stage tech are predisposed to overestimate the impact on the broader economy.
The article gave two numbers I didn't know about, and both came as a surprise:
> As at the end of 2022, it had 37,466 deposit customers, each holding in excess of $250,000 per account -- and -- The bank does have another 106,420 customers whose accounts are fully insured but they only control $4.8 billion of deposits
So SVB had only about ~150k banking customers. And of those, less than 40k are actually affected by this debacle. But those were concentrated enough to make SVB the 20th largest bank in the US.
Yeah. We're not an SVB client nor are we in Silicon Valley, but we're already auditing our accounts payable and client lists. It's likely we aren't screwed, but our contract accountant who told us to start doing this kind of bookkeeping said some of his other clients have massive counterparty exposure to this bank failure.
the FDIC does more than hand out 250k checks. They will find a buyer for this bank who can successfully manage its assets. and most likely will payout uninsured account holders at least 90 cents on the dollar. Probably even all 100 cents. Remember the bailouts the goverments made to banks on 08, we didnt lose a cent on those. the govt made money. We are gonna be ok
If the Twitter saga has shown anything, it’s that companies can afford not to pay on their contracts, at least in the short term. Amazon will give them a grace period because a customer paying a few weeks late is better than the customer going belly-up.
Can you expand on that? The FDIC per account insurance limit is definitely $250k. Are you saying that the FDIC has other protections in place, or something else?
Apparently lots of VCs and startups use SVB as their bank. Could someone explain why would startups not choose relatively safer big banks like BoA or Chase? What are the advantages of using SVB as the primary financial institution?
Venture-backed startups are weird in terms of cashflow compared to traditional businesses.
They appear out of thin air with no history and millions of dollars. They then constantly burn money in the tens to hundreds of thousands of dollars a month until they disappear or get injected with millions more dollars. Often, traditional banks don't know how to deal with these entities and it sets off tons of alarms in their fraud and risk management departments.
SVB and related departments know the business and financial models of startups a lot better and are more willing to help.
i really wonder if that was a problem. My last startup banked with Wells Fargo for a while all the way from pre-seed to A, and it seemed to be fine. Anyway, my number 1 ask from any bank is not to fail, before anything else, and that didn't really work out great here. If SVB was a top-5 bank it probably would have been bailed out by the govt. Not great, but probably safer for the customers.
This makes complete sense, however, why would big banks not want to compete against SVB by establishing a smaller division that handles startups? Obviously, the market is small compared to their usual markets, but there is still a lot of good money to be made.
> why would big banks not want to compete against SVB by establishing a smaller division that handles startups?
Because such startups tend to, when they can’t raise funds easily (=nowadays), withdraw all their cash deposits at the same time (=what’s happening at SVB), which puts it in bankrupt. So, small market, but so much variance in deposits that it’s risky.
Bank of America couldn't care less about our XX million/year revenues. Literally wouldn't take even basic requests for services. We moved to Keybank who has a much better rep with small businesses, and service (and pricing) has been orders of magnitude better.
I'm assuming a similar thing was going on with SVB vs. other large banks.
Big banks do not understand the cashflow situation of startups. They flag standard fundraising as fraud, drug money, or money laundering (or simply won't open an account in the first place).
They understand it perfectly well, it really isn’t that complex. You are repeating SVB marketing. Bank of America works with thousands of start ups and I’m sure all the other large banks do as well.
I have to say, Bank of America was a great bank for my startup. We wouldn't have gotten a bridge loan with them or anything like that but for operating cash management there was never an issue.
> Big banks do not understand the cashflow situation of startups. They flag standard fundraising as fraud, drug money, or money laundering (or simply won't open an account in the first place).
It's not like they blacklist you the moment your attempt to open an account raises an eyebrow. Show up to BofA or Chase with a copy of your contract with your VC, in addition to their check, and it would be odd for them to refuse you an account.
Uhhh yes. Most commercially viable businesses at scale rely in credit. If you're making serious money (call it $10M+ in revenue), you very likely have some credit facility with your bank. Whether you use that credit or not, is another thing.
I'd expect that in many cases they are forced to as part of the venture funding offered by SVB. F.e. Secure a $10M venture note (in connection with Series A) but use SVB exclusively for banking.
Weird or interesting how SVB survived 2000-2002 tech crash , and also 2007-2009, but 2022 was too much. Every crisis or cycle is different. Unique challenges, threats, etc. This is why it's so hard to predict these sort of things. It's impossible to anticipate all the ways things can go wrong.
The author seems to allude to a difference between then and now is that it wasn't quite as easy to withdraw money back then, usually you'd have to at least make a phone call, or head over to your physical branch. Now you could just open the website/app and do it there, immediately.
Another point seems to have been that Silicon Valley bank have a customer base which is tightly knitted, so if any bad news gets shared in the circle, it gets around quickly. Maybe they weren't as popular the last crashes, or as impacted as other sectors, so it would be easier to avoid trouble. But specifically this time, the trouble was related to them, so they got hit the hardest.
Maybe... but SVB's immediate issue is that they lent money when rates were zero and those loans decreased in value when interest rates went up. This type of issue can only be caused by a decade-long zero-interest environment where people stop believing rates will ever go up again. It's interesting how every financial crisis has its roots in the specific economic conditions of the time.
Even so, no matter what cause, if people are unable to withdraw money + won't hear any rumors about a bank run, would SVB be in the scenario they are experiencing today? As it seems like the bank run is what did them in, not whatever magic they did with the deposited money.
Yes. Their issues started when they needed to sell their assets available for sale. This happened because interest rates made it harder to raise money, which caused their clients to draw down their accounts faster than they anticipated. The bank run at the end is more a symptom of SVB's issues rather than a cause. Everything about the bank's problems can be traced back to believing interest rates would be 0 forever.
They were insolvent before the bank run, that's why they needed the firesale and emergency equity raise to stay afloat. If they were able (allowed) to hide that fact and get the money quietly elsewhere, we may never have known. Or perhaps they would have continued to lose money and fail later, with depositors taking a larger haircut.
> quite as easy to withdraw money back then, usually you'd have to at least make a phone call, or head over to your physical branch. Now you could just open the website/app and do it there, immediately.
The main reason some of the other startup banks such as Ramp or BREX exist is because SVB was kind of a dinosaur when it came to cash management. Their online cash management apps were very basic and rudimentary. And they were a lot slower than these new online banks at responding to customer requests.
So now you're saying that bank (a very conservative investment bank from what I've been told) was moving too fast?
That’s because in those situation, SVB didn’t put $91b of its deposits into mortgage bonds right before interests rates soared. Those bonds are now worth a lot less (and will be sold for even less still to other banks that want to pick them up on the cheap).
Every crisis is different, but hedging so much money in one basket seems to be the real undoing here. I’d posit that the bank, which was much smaller in 2000 and 2007, didn’t make those sorts of moves then.
It might be impossible to anticipate which way exactly things will go wrong, but I’d say it’s pretty easy to anticipate the ways things can go wrong. I see this as at least ignorance, probably negligence, and at most, fraud, on the part of a lot of folks involved.
This has been the most widely predicted recession in history as far as I am aware. The Fed has openly and deliberately said they will raise interest rates until something breaks. Investing in assets that are negatively correlated with interest rates under these conditions is a terrible idea. Professional bankers should know this.
Startups are far less negligent, but still, could diversify their deposits with multiple banks (or use “real” banks) to mitigate their risk. Startups are risky enough; having sound money is important and I wouldn’t be risking it with a regional bank.
The Fed is negligent for treating financial markets like penny stocks. They had nothing else in their tool chest to fight inflation, though, so I can’t entirely blame them.
Politicians and the Fed are negligent for creating so much liquidity that it spurred such immense inflation. Anyone with a macro-101 level of economics could have foreseen this.
Voters are negligent and ignorant for their unwillingness to elect better politicians and demand change and reform in an area that continually fails them. We are failed time and time again by the banking system because it is fundamentally unable to be guaranteed solvent. We bail banks out or get by a crisis somehow and forget about it until the next time around and then shout “wow, who could have seen this coming?!” as if it is the first time a bank has failed.
What makes this thing even more interesting is that there is viral, wide-spread knowledge about banking - as opposed to 2000 or 2007. A bank run is one tweet away.
A lot of banks have similar asset books and asset values MUST go down as interest rates rise. So, banks are really screwed unless they offer MUCH higher interest rates on deposits ASAP. Because if they don't, most people are very aware of treasurydirect and will simply deposit there.
Do you work for a company that has most of its operating cash in SVB? If so you probably won't see your next paycheck or it will be late. Even if you work for a unicorn there's still a risk because they're not likely to hold more cash than is needed for a few weeks (which is now held up beyond $250k.) The rest of the funds are in bonds or other investments that will take some time to unwind to get cash to meet payroll.
Even if you work for a company entirely banked with SVB (and many companies have a few bank accounts for various reasons) they'll probably be able to make payroll even if they have to take an annoying loan.
Annoyingly, Rippling had a lot of their stuff in SVB so some companies see the debit from their SVB account but no actual corresponding credit into their employees accounts.
If the companies had their account with SVB then rippling is not the issue. Issue arises when their bank is something else but since rippling uses SVB, they see debit in their accounts but SVB doesn't release pay roll.
(Unless rippling forces you for opening SVB account, i don't know, i do not use rippling)
Can someone ELI5 how the increased interest rates hurt the bank's strategy? I've read a few articles now and still can't parse through the language to understand what's happening.
Say you buy a bond for $100 with 10-year maturity. This means the bond will pay $100 in 10 years. What is it worth today?
Let's say you are looking for an investment to pay 1% yield year over year, because that's the going market rate (which highly correlates to interest rates). Then this bond should be worth about $90.53 to you. You buy it at that price.
Now suddenly rates change and everyone is expecting investments to pay 3% year over year. Then an equivalent bond is now worth only $74.41.
That's OK. You don't have to sell your bond. You can hold on until it matures, and just live with the fact that you aren't getting the same yield as others.
But now imagine you need to come up with a lot of money quickly to satisfy some requirement (like customers withdrawing a lot of money). Now you have no choice but to sell your bond. And that means you have to admit you lost the money. And now you may not have enough money to cover your obligations, even though you would have if you could have waited for the bond to mature.
And the reason SVB had more bonds than most is because there was a huge influx of deposits in 2021 from investors going nuts funding startups, but SVB doesn't have many customers that are good candidates for loans. They needed to do something with the money. So they bought bonds instead, because those are "pretty safe".
They were holding some candies that became less valuable because you can buy better candies for the same price in the shop. They needed the money, so they had to sell those candies for cheaper. Some friends of theirs noticed they were selling those candies for cheaper and started getting suspicious and panicky and asking for their toys back.
* Candies - fixed income securities, like mortgage backed-securities and US treasuries
* Candies becoming less valuable - government raising the interest rates, which makes fixed income products, like treasuries and mortgage-backed securities less valuable, because you can get the same securities but with higher returns in the market
* Friends panicking - investors and depositors
* Friends asking for their toys back - withdrawing money from the bank, which starts the bank run
They laid off 30 people in less than 24 hours of this announcement? That sounds like those people were already getting laid off and this was a convenient cover
> 1. Some mighty people/companies will bail out all depositors
Yeah, I'm sure Elon Musk can just write a check to cover this... oh, wait.
Seriously, who do you think would have the capital to swoop in here and cover something north of $170B in shortfall? Especially right now given that money is pretty tight everywhere.
Once its becomes more clear how much SVB assets can be sold for and along what time frame, there will be plenty of firms willing to either: 1) purchase receivership certificates at a discount to expected value 2) issue bridge loans to cover the expected value of receivership certificates.
It's not a bail out in the sense of free money. It's a bail out in the sense of a short term loan, just like after the GFC.
This is the real question. How much are those MBS worth on the open market? Their value decreases every day as the fed keeps raising rates. I don't think anyone can really say what kind of hole SVB might be in right now.
In one of the other threads discussing this SVB was said to have ~$200B in assets, 93% of which were not covered by FDIC. That actually comes out to $186B.
FDIC Takes over Silicon Valley Bank - https://news.ycombinator.com/item?id=35096877 - March 2023 (708 comments)