Emergency bridge loan for SVB customers
brex.com
brex.com
For brex this has to be the best marketing ever. Sign up a bunch of customers with real businesses at someone else's expense.
The amount of risk here is significant. I think this is a desperate play by a company in a struggling industry.
losing 20% of your cash-on-hand, where that's above $250k. For most startups, that's going to be a haircut their investors take, where the founders can say "yeah, literally none of this was our fault".
For mature, profitable businesses with recurring income, this is going to bite, but they haven't lost 20% of their customers or 20% of the amount of money they expect to get paid next month.
For startups, where the money was investment, unlikely their investors are going to blame them for this loss. It might shorten runway by up to 20% for some pre-revenue startups, by less for startups with actual revenue.
Not actually clear who is taking the downside
> This credit line is funded by 3rd-party capital (and not Brex directly)
> This credit line is funded by 3rd-party capital (and not Brex directly), who are working with Brex to minimize the impact of this event to the startup ecosystem.
It sounds like a bunch of VCs are gambling on the return, while also knowing it helps stabilize the system that built (and likely holds, in some way or another) their wealth.
In a regime of rising interest rates this gives them pricing power on the loans where they were just passive depositors previously. Since SVB loans were essentially only to companies who were already customers of these VCs, you could view this move fairly cynically.
If they bought them near the peak value what they can actually sell the income streams for is going to be a lot less than what they paid.
In the meantime, impacted account holders can get their IOUs and borrow money to make payroll.
1) When yields go up, treasury prices fall. 2) When yields go down, treasury prices rise. 3) The only way you get the basis cost for a treasury back is if you hold to maturity.
When a bank run happens, you (if you are bank) need cash. Lots of it. If you own assets, you have to sell those assets to come up with that cash. SVB had to sell ALL of their notes for less than they bought them for. See #3.
The losses by SVB were realized well before today. This has been going on for a while now. Outside of $250,000 per account type per person, the only additional money folks will get back is whatever a bank sale comes up with, which won't be as much as most folks think, since the core assets were already sold off. If I had to make an educated guess, most startups/investors will lose more than 50% of what they started with. Remember, high interest rates. SVB will NOT be sold at a good price. This is not a market for sellers, it is a market for buyers. Again, sky high interest rates and many of the very investors that could make that sale possible had money at this bank. That means a lower sale price.
If I lend you 25% of your SVB deposit secured against your SVB deposit, I should still be fine (as long as my claim comes early enough in any ensuing bankruptcy).
Long term treasuries have declined 40% in value since they peaked in 2020.
https://www.google.com/finance/quote/TLT:NASDAQ
Banks hold a lot of treasuries as part of their capital requirements. So long as they intend to hold them to maturity, they don't have to mark them to market.
Banks collectively hold about $620 billion in loses in held-to-maturity securities right now since they have declined rapidly in value during the Fed's interest rate hikes.
https://www.aol.com/finance/why-silicon-valley-banks-crisis-...
If you can hold them to maturity, you're fine. If you need to sell them to raise cash (to cover other losses or during a bank run) those losses become real.
And if your startup had $10m in cash in SVB, but now you find out you're only getting back, say, $7m, that's very bad news for your business.
> Long term treasuries have declined 40% in value since they peaked in 2020.
This is inaccurate. What you linked to (TLT) aren't bonds, these are bond funds. The way a bond ETF works is that they have a stack of bonds that track the benchmark interest rate. They periodically sell off their old bonds and buy new ones, they don't just sit on them and wait for them to mature. That means TLT's NAV goes down when interest rates go up because they're selling lower interest rate bonds, and buying higher interest rate bonds.
Treasuries cannot lose value just like cash. They are as good as cash in almost every context. You can always sit on them until they mature and you'll get the full amount plus interest. You cannot lose money this way.
Where you can lose money though is if you have to liquidate them sooner. Why would someone buy a 3% 30y treasury for $X from you when they can get a 4% 30y treasury from the source for the same price? You have to sweeten the deal by paying out the difference in rates. This is where you can lose money.
SVBs issue was a mismatch in durations. They had too many demands for money out now, and too little available now. They have plenty of money coming in the future, but that's too late.
Reply to two basically correct statements, make a claim that's completely false, and then — and here's the genius — say a bunch of true statements that come around to support the view you're disagreeing with!
Be careful with absolute statements like that. While it hasn't happened so far with USA treasuries, its equivalents in many countries have lost value in the past.
No free lunch. Treasuries yield more than bills because they’re less liquid.
T-Notes - 2y-10y
T-Bonds - 20y-30y
Only if your definition of risk is limited to default risk.
It looks like a lot of their holdings were in MBS and a lot of the treasuries were recently sold at losses (which necessitated raising capital, which appears to have set this whole thing off), but there is no reason this couldn't have happened if they only held treasuries anyway.
Then the government starts selling those bonds a lot cheaper. To buy the same bonds you have today would only cost $900k.
Even though the 2030 value of those bonds is the same, the 2023 value just plummeted. (And they will gain more per day to eventually make up the difference.)
When your customers demand their money, you have to give them 2023 dollars.
A bond that you can redeem early has the safety of cash here. A bond that you can't redeem early does not.
Back of the napkin, take the 10yr and 30 yr spot prices today for issues from 1-2 years ago, and that's your max haircut. I believe some are trading at 70 cents , so we are talking about 30%. And that's worst case (not all assets would have sold at that price, but better).
Remember this is highly liquid assets. Not some exotic stuff. Its a big loss, but not 50%
Brex's offer is collaterized up to 25 cents per dollar. SO, unless SVB lost huge money elsewhere, there's no way SVB lost > 75% mostly off their MBS portfolio (or other assets, for that matter).
Do you know this? That's a number I (and some other posters) pulled out of a hat as a reasonable thing to do, but I don't think any of us had any sources for it?
Depends upon how much was liquidated and already went to paying out fleeing customers at 100% of deposits. The average haircut could be 30%, but late to move depositors could have it worse (this is why you participate in a run on the bank).
Until the FDIC pores over their books, nobody can say anything sensible about what they own, what they sold, what they sent out and what is secured as collateral.
> As of December 31, 2022, Silicon Valley Bank had approximately $209.0 billion in total assets and about $175.4 billion in total deposits.
And I'm not trying to claim that the balance sheet necessarily entirely reflects reality
December 31st, or even March 1st, are virtually irrelevant to a bank collapse today. Assets were sold at a deep discount, presumably some were pledged, and deposits fled.
SVB's carcass still has significant assets, so Brex can set that % number based on what they know about SVB's state, to make the loans almost risk-free.
I'm not a banker and have no idea if this is actually what Brex is offering. That public page doesn't discuss terms at all, which makes sense given the chaos. But there's certainly ways to structure these loans without much risk.
Balance to take 2-3 months. Will require an actual application and paperwork.
Depends on the price they got for those assets.
California has declared it insolvent. It literally by definition is now.
from the article op linked, in "Findings of Fact", by the Commissioner of Financial Protection and Innovation:
> the bank is now insolvent
This is nonsense and contradictory
The bank was insolvent. FDIC took over (due to insolvency) and has no obligation to make the bank solvent again (other than the 250k insurance limit)
If SVB literally had a way to hack the time-space continuum and wait out for asset prices they own , to stabilize ("maturity") , or to pay them back in full ( a loan)...SVB would STILL likely lack enough funds to pay back their deposits.
The issue wasn’t the selling of assets, it was the panic that their actions took to prop up a balance sheet hole. It’s not like Enron or Lehman.
I contend there is no private entity credit that would go near SVB because precisely they realized more losses than they had more depositors, and such credit would have never been secured at the top of the pile with the FDIC lurking nearby.
No credit facility would save a doomed bank after the realized losses. It was a matter of time.
The panic was not the cause. The panic was always going to happen. A public company would have never been able to do a firesale/equity raise without inducing a panic in the first place.
At market price, not face value.
If you need $80B before your 10Y bonds mature and borrow it today, you're paying minimum 4% interest. By the time your $80B in bonds mature you owe $118B
If you're saying it doesn't matter to FDIC because they can get "free" credit, that's equivalent to a bail out
> This credit line is funded by 3rd-party capital (and not Brex directly), who are working with Brex to minimize the impact of this event to the startup ecosystem.
Those deposits are backed by the assets on the balance sheet at SVB. They went under because they ran out of liquidity, not because they're massively upside down or their assets are crap. Those assets are largely medium-duration treasuries (10Y IIRC). They will pay out face value as they mature (they are the definition of risk free), and the losses on their balance sheet being reported are marked to market. Assuming they have to sell them immediately - which they did have to in order to meet withdrawals.
Now that the bank is in the FDICs hands, those medium-term treasuries need not be sold, and other liquidity options exist to make un-insured depositors mostly or entirely whole. One such option is just selling all accounts to a big guy like JPM, the way WaMu was handled in 2008. Or the FDIC can swap the treasuries for cash and since they have no time pressure, just wait until they mature.
That's just a math problem. If you assume all deposits are at least 250K, you can get a floor for that number. If 97% of deposits exceeded threshold, then the average deposit is at least $8.3 million.
85% of accounts weren't FDIC insured
https://time.com/6262009/silicon-valley-bank-deposit-insuran...
I'm sorry for the people in that position but it's a distinction with a difference.
Yes some startups will go bankrupt and most of the remaining ones won't be profitable any time soon.
Surely Brex isn't counting on the profit from operations of the borrower to repay the loan, but rather that company's deposits at SVB being released at some time in the future.
Not sure how this works out in practice, IANAL, etc.
It was the same situation with LTCM, with their counterparties (large banks) rushing to liquidate before LTCM became insolvent, which the Fed feared would send shockwaves through the financial system. The banks quickly, collectively agreed to do the right, responsible thing and stopped the fire selling, and injected some emergency capital into LTCM, and then gradually unwound LTCM's postions (and made money in the process).
Seems like you haven't needed to think about systems design as a VC! "Yeah go to the same place as everyone else I invest in, what could possibly go wrong?!"
Do mid-level software engineers typically diversify across every single point of failure? Architectures? Programming languages? Compilers? Too much diversification itself can be a point of failure.
As long as you have a sufficiently robust insurance backup (the FDIC) choosing a robust single point of failure seems like an okay idea for business operations (it's not like it's literal life support). The people at the FDIC aren't machines; they can do what's necessary to fulfill their function in a timely manner.
I think you're unfamiliar with the Long-Term Capital Crisis, and the speed with which it was resolved (two days: Sep 22-23). In fact, it's partly because the NY Fed and big banks acted so swiftly and responsibly, that most people have never heard of this and don't realize how close we came to a financial crisis:
>The Fed came to be concerned that if LTCM’s extensive list of counterparties tried to exit their positions at the same time, it would create a rapid and widespread sale of assets, a fire sale, which could potentially impair the economy.
>On September 22, the New York Fed invited a core group of three firms to a meeting to discuss the LTCM situation. The core group, later expanded to a fourth firm, formed three working groups to consider possible solutions, one of which came up with the idea of a consortium approach. A broader group of thirteen firms was invited to the New York Fed that evening to discuss the approach. The firms disagreed over how much each firm should contribute to a rescue package and could not commit to such an effort on such short notice (Siconolfi 1998).
>The talks on a combined rescue reconvened on the morning of September 23, but were soon halted by news that an investor group led by Warren Buffet had made an independent offer to buy out the firm's partners for $250 million and subsequently inject $3.75 billion capital into the fund (Loomis 1998). This appeared a clean solution to both the creditors and the Fed, and McDonough advised Meriwether that it was likely his best bet (Schlesinger and Schroeder 1998). By the 12:30 p.m. deadline, however, the offer was not accepted due to reported legal issues.
>With no other solution in sight, the talks resumed with more haste inside the New York Fed. The consortium ultimately came to an agreement at about 6:00 p.m. on September 23. Together, fourteen firms put up $3.625 billion in capital in exchange for 90 percent of the fund’s ownership (two firms included in the talks declined to participate).
https://www.federalreservehistory.org/essays/ltcm-near-failu...
Coincidentally, this took place ten years before, and nearly to the day (Sep 15th), of the collapse of Lehman Brothers in 2008, which had also been especially exposed to LTCM in 1998.
Does svb have a similar client list?
No, the consortium members actually made money by slowly unwinding LTCM's trade book. It's not clear that a fire sale of LTCM's collateral by a rogue counterparty would have done the same, and at best might have just minimized their losses had LTCM completely collapsed.
I was on the phone with portfolio companies, friends and clients yesterday ensuring they pulled their funds to a back-up bank account. SVB was a national bank. It's not a charity, or even a local bank serving a niche community. It was a big bank, and a badly-run one at that. Nothing they do isn't done by others. I believe in my friends and their missions more than anything SVB was up to. (I'm not a VC.)
It was more of a regional bank, specifically one servicing SV, with at most sporadic branches elsewhere, usually just one in a given state in its largest major city (e.g., one in NY: in NYC; one in D.C.; one in Colorado: in Denver). In many states they have no branches (e.g., Florida): https://www.svb.com/locations
> Nothing they do isn't done by others
Then why did so many startups and VCs bank with them?
> my friends and their missions
I would personally be hesitant to use this terminology for others, unless they're actually doing something truly impactful and impressive, like curing some horrible disease, going to Mars, or at least actually in the military on some mission.
Top 20 by assets and with a branch in New York. That’s a national bank.
> why did so many startups and VCs bank with them?
They were first and did what they did well. That doesn’t make them preciously unique.
> would personally be hesitant to use this terminology for others, unless they're actually doing something truly impactful and impressive, like curing some horrible disease, going to Mars, or at least actually in the military on some mission
Over a bank?
Over a financial contagion that could sap the liquidity and solvency of the startups these VCs have invested in, and possibly cause harm to the wider economy (which I doubt they care very much about)?
The entire VC model is to make investments with 10+ year time-to-profit.
Right now SVB, if fully liquidated, cannot repay all of the deposits.
If I have assets worth $110 today but they are locked up, and I have to pay back what I owe today, then I’m facing a liquidity crisis.
If I have assets worth $90 today, but $110 in a few years, and I have to pay back the money I owe in a few years, then I’m solvent and everything is good.
If I have assets worth $90 today, but $110 in a few years, and I have to pay back what I owe today then I’m insolvent.
Per the FDIC
Still certainly very possible that the analysis is incorrect.
[0] https://dfpi.ca.gov/wp-content/uploads/sites/337/2023/03/DFP...
Same with their mortgage-backed securities.
Ish. On a mark-to-market basis they had insufficient reserves. That's closer to insolvency than illiquidity. The mismanaged duration is closer to illiquidity. But not of the sort a lender of last resort could save them from.
Hold-to-maturity assets are not required to mark to market, for example.
And it's not based on what they paid for it from what I gather, it's based on what the total payout will be when it matures.
Imagine a bank takes demand deposits and pays a variable rate of the base rate minus 1%. The bank then makes fixed rate, 30 year mortgage loans that it intends to service itself at the base rate plus 2% (portfolio loans).
Interest rates then go up.
Bang: that bank just became insolvent. Is this actually the model of the world you want?
I'm not sure what the problem is, this model to me actually seems reasonable. The alternative is to claim that "We are solvent as long as our clients keep deposits in our bank and are happy with an interest rate lower than our competitors for 30 years".
In fact the scenario you're describing is basically what happened to SVB except the interest rates increased more.
I guess the main issue is that generally banks do take a risk by borrowing short and lending long, and hope that things work out in the end.
This is irrelevant in a scenario where every client wants to withdraw their money, because the only way SVB can fulfill that is by selling their HTM assets at the current market price, which will be at a heft discount compared to where they bought them.
When yields rise, prices fall. Remember that.
Now let's say you need to access that emergency fund, and the amount you need is 90% of it. However, what you've found is that treasury prices have fallen so far, your investment is now only worth 50% of what it was before unless you hold it for another 5 years. If you don't sell now, you will be homeless. If you do sell now, you buy a bit of time and can possibly get a loan, and barring that, you will be homeless.
That is what happened to SVB.
If you need a closer date (because you need to fulfill customer cash flow requests), then solvent plus illiquid can become insolvent quickly.
FDIC is VERY efficient. Most probably, the bank will be open on Monday morning, albeit with another owner, and most small depositors and startups wouldn't even notice if it's not for all these 'breaking news'.
The guys that are on the hook here, and their exact job is to manage money, are the VCs. They will most probably loose a relatively small percentage of money (say 10%).
They threw all their Atlas customers into SVB, according to Atlas customers
> Loans are issued by Celtic Bank, a Utah-Chartered Industrial Bank Member FDIC. All loans subject to credit approval.
Stripe itself was, and probably still is, using Wells Fargo for US corporate accounting, per https://qr.ae/pvERsZ
If a victim of SVB tries to use HN as Stripe customer support and asks if they are exposed, will they respond and admit that their business capital funds in Atlas is lost?
> They threw all their Atlas customers into SVB, according to Atlas customers
And removed all mentions about SVB in Atlas. Stripe has been very quiet about their own involvement in SVB for their entire business.