Silicon Valley Bank collapsed today in the second-biggest bank failure in US history. Why? Because it put a ton of money into long-dated government and mortgage bonds. This would normally be pretty safe:
When the Federal Reserve began raising rates last year, however, those holdings became less attractive because newer government bonds paid more in interest. That might not have mattered so long as the bank’s clients didn’t ask for their money back.
But at the same time as interest rates were rising, the environment for start-up funding dried up, putting pressure on the bank’s clients — who then began to withdraw their money. To pay those redemption requests, Silicon Valley Bank had to sell off some of its investments at exactly the wrong time. In its surprise disclosure on Wednesday, the bank admitted that it had lost nearly $2 billion when it was all but forced sell some of its holdings.
It's not clear yet if SVB really did anything wrong. It invested in government bonds and then had to sell them at a loss when the startup market stalled and clients wanted to withdraw money. Clients were spooked by the reported loss and began a full-scale run on the bank.
Normally US government bonds are considered pretty safe, and SVB reported a Tier 1 capital ratio of 15% last quarter. This is well above the 6% Basel III requirement. Its Tier 1 leverage ratio was 8% compared to the Basel III requirement of 3%. Even after the $2 billion loss those ratios would have been 13% and 7% in the absence of any other losses. Overall, SVB looked pretty strong.
If anything, this makes the whole thing a little scarier. It's one thing for a weak, feckless bank to fail, especially in the face of a huge economic bust like the one in 2008. It's quite another for a strong bank to fail during normal economic times. The obvious conclusion is either (a) this was just a weird one-off failure, or (b) things are worse than we thought.
I don't have the chops to know which it's more likely to be. I imagine it will partly depend on whether depositors get most of their money back. If they do, it's a blip. If they don't, there might be follow-on problems. I'm sure we'll learn much more over the next few days.
Wasn't Silicon Valley Bank FDIC insured?
Yes, so deposits up to $250,000 are insured. Very few consumers have deposits greater than that, but businesses routinely hold more than that in the bank, and they are the ones who could be in trouble.
Deposits less than $250K are 2.7% of SVB deposits.
https://twitter.com/GRDecter/status/1634208652595699713?s=20
That's a lot of non-insured deposits. Without a buyer to step in, that means depositors will need to wait to recover what they can. FDIC says those depositors will get a receivership certificate and recovery of funds will be contingent on sale of SVB assets. Good luck.
We were WaMu customers when it failed. The hand-off to Chase happened (at least for us) without a ripple.
I have just received my 3rd payment order and $30,000 that I have built up on my laptop in a month through an online agent…!u91) This job is good and his regular salary is much better than my normal job. Work now and start making money online yourself........ Just Copy And Open My Profile Name Link,
Google paid 99 dollars an hour on the internet. Everything I did was basic Οnline w0rk from comfort at hΟme for 5-7 hours per day that I g0t from this office I f0und over the web and they paid me 100 dollars each hour. For more details
visit this article... https://createmaxwealth.blogspot.com
Yikes, I have been paid with paychecks drrawn on SVB. I've been to one of their very few tellers to cash a check drawn on them, to make sure the money was there. The company I was working for was going into bankruptcy.
"It's not clear yet if SVB really did anything wrong. It invested in government bonds and then had to sell them at a loss when the startup market stalled and clients wanted to withdraw money."
Even if this is all that happened, it seems like a pretty big failure of enterprise risk management. Rising rates leading to the fall in market value of long-dated loans/bonds is exactly the kind of scenario a bank needs to be prepared for. Rising rates would also impact their client base disproportionately. If they didn't prepare for this by managing their interest rate risk appropriately, I'm not really sure what their risk management function was doing.
But what would they do?
I think it's more an example how regional and specialized banks have specific risk.
They should have hedged their bets with securities that would do well if interest rates went up. Instead they stacked everything in securities that would lose money if rates went up. And they did this while cultivating a client base that was dependent on cheap capital and would be in trouble if rates went up. And they filled up on business accounts over the FDIC insurance cap, which made them vulnerable to a bank run if their balance sheet faltered.
They clearly never considered what would happen to them if interest rates went up.
This is what I wanted to say. They were not being willfully reckless, but they were incompetent. Changes in interest rates are a very basic thing to consider, and they clearly didn't.
Investing in long-term government bonds? That is the most recklessly dangerous balance sheet a bank could have. How could there not be a simple financial risk metric that would highlight this instead of all the other banking tools that merely disguised it? Duration gap? Income gap?
"Investing in long-term government bonds? That is the most recklessly dangerous balance sheet a bank could have."
I'll have to see your work on that one, everything I know and found online says otherwise. I mean, it's not like these bonds are crypto or NFTs, -those- would have been the most "recklessly dangerous."
The thing about bonds is that they are frequently traded early for less than face value. When interest rates rise sharply (like in the last 2 years), existing bond prices fall sharply. I mean, who wants to own a 1.5% interest, 5-year bond when a simple savings account is paying 4.5% interest? SVB made a huge bond bet back in 2020 that interest rates would remain low. They were wrong. Now, a regular person could just hold off and collect the bond face value in 5 years, but SVB needed money NOW and so had to sell off their bonds for cheap.
Bonds in themselves are safe and predictable. As in many things, the risk is generated in the secondary market.
I am just so surprised by what has happened in the banking system. I had no idea that banks could have such massive unhedged interest rate risks. My basic conception was that all banks would have 24/7 trading desks continuously monitoring treasury markets which would compete against each other to manage risk. SVB completely ignored the treasury market even when they took a 100 B naked long position in long term treasuries? It would never have occurred to me that they might just buy 100 B in long treasuries and just dispassionately watch as they lost 20 B on a marked to market basis. If banks are that uninterested in the value of their assets perhaps they should contract out the management of these assets to the web to a globally competitive virtual trading platform.
What surprised me even more was that this was not even just SVB. SVB flushed 20 B on their unhedged Treasury position, while the entire tab for the banking system was over 600 B! I would have never predicted that. Even more surprising was that during 2022 central banks made it explicitly clear that they were going to raise interest rates no matter what: They fully committed themselves to controlling inflation. How do you realistically call a bluff when you are playing monetary policy poker with central bankers?
In terms of how the interest rate risk could have been hedged, I am not completely sure of the exact mechanics, though futures options, etc. would seem to be a possibility. Also, I wonder whether it would have been sensible at some point to sell 2 year forward futures contracts on say 10B in 10 year Treasuries etc..
As a former bank examiner, and current CFO of a bank, I guarantee that the bank had a lot of loans that were very illiquid and probably worth far less than the face value of the loans.
They should have, and I am sure they did, run an interest rate risk forecast showing the expected changes in income and capital if interest rates went up 4%. The regulators would have beat them up if it showed a disaster. It probably wasn't pretty, but it wasn't a disaster.
SVB's particular problem was similar to Continental Illinois back in the early 80's; it didn't have many core deposits. It was funded with deposits that were very interest sensitive. So their cost of funds went up significantly as rates rose. Most banks, almost all banks, make more money as rates rise because the rates on loans go up faster than the rates on deposits.
So, SVB didn't have anything to fall back on as depositors left the bank. It needed to sell assets because it couldn't use their loans as collateral because people didn't trust the value.
I hope the FDIC screws the uninsured depositors. Far too many times the regulators pay everyone like nothing ever happened.
Melrose Credit Union, the largest lender of taxi medallion loans, failed and the foolish regulators paid all of the uninsured depositors. The problem there was the regulators allowed a very high concentration of taxi loans that ended up being terrible when the value of a medallion dropped from $1.2 million to about $0.1 million.
SVB had too high a concentration in loans that didn't hold their value.
I could go on, but I doubt anyone is still reading.
Still reading, and would love to hear more.
I second to continue. The Secret Sauce ™ is the complexity, and stories of one of the cooks are very interesting.
Please continue. I may know a whole lot more about banking than the typical person on the street, but basically, I don't know that much.
So you're saying this is a sui generis (because of the nature of its deposits from start-ups and the large rise in the Fed rate) and not a black swan signaling a previously unknown systemic risk?
No, he is saying that the bank did not have a broad deposit base (i.e. savings deposited with them, whatever the specific form), what is called core deposits, typically retail / small depositors who rarely move and typically have low rotation; instead they were reliant on market funding and jumbo deposits from corporates that quickly move their deposits (and are large values moving) to higher rates.
"Deposits from start-ups" is a misunderstanding. Corporate deposits.
Their funding model was risk exposed while also lending concentrated in risk.
It's a classic industrial specialist bank issue. Absolutely nothing novel.
This puts alot of leverage assuming their loans lost value...
...and nearly none that it's not just the value of those loans, but that others will take them as collateral.
Melrose Credit Union, the largest lender of taxi medallion loans, failed and the foolish regulators paid all of the uninsured depositors
Why was it foolish? It's not as if the bank's owners were being being made whole. People kvetch about "moral hazard" but that cow has long since left the barn: we shouldn't provide any deposit insurance at all if moral hazard is a sufficiently serious concern.
The assertion "we shouldn't provide any deposit insurance at all if moral hazard is a sufficiently serious concern" is simply a fallacious position, the excluded middle. There is a trade off on moral hazard and costs and it is not a black-white, on-off trade-off - rather there is a trade-off between escalating social costs for insurance against loss as that reduces sophisticated actor incentive to monitor their own risk inducing loss of rich actors to be "socialised"
It is simplistic borderline innumerate reaction.
The poster I agree with as there the jumbo-depositors - that is over the insurance deposit limit are generally corporate or high-net-worth and the social and economic cost of unlimited coverage to them when there is not a legal obligation, and further becomes incentive to risk taking with gain privately captured while loss is paid by the public- the very thing the Left cries about normally in a baking crisis.
is simply a fallacious position
No. It's simply one you happen to disagree with.
"It's not clear yet if SVB really did anything wrong. It invested in government bonds and then had to sell them at a loss when the startup market stalled and clients wanted to withdraw money. Clients were spooked by the reported loss and began a full-scale run on the bank."
Well, sure. But if it's just a run on the bank--just a liquidity problem--then you'd reasonably expect that some larger institution would snap up SVB at a discount.
The fact that they couldn't find a buyer, even at fire sale prices, seems to point pretty strongly toward they're insolvent, doesn't it?
Banks normally have all the liquidity anyone would ever need.
They can borrow against their first mortgages at FHLB. They can borrow against many types of investments from the Federal Reserve. They have lines of credit secured by basically all of their assets.
Sometimes the system has a system wide failure like the Great Recession. But that can't be the problem today.
I assume SVB was not a Systemically Important Financial Institution which would have brought a far higher level of scrutiny but I am totally not qualified to talk about them.
https://en.wikipedia.org/wiki/List_of_systemically_important_banks
I guarantee they continually performed simulations about interest rate risk. ALCO is a critical item to regulators. (I am glad it is. It wasn't back in the early 80's and that was a major cause of the S&L crisis.)
So why did the bank run out of liquidity?
Normally, you raise deposit rates and loan rates when you have a liquidity issue. You get more deposits and you make fewer loans. Why didn't it work here?
Probably it did work but depositors will get scared when you raise your rates too high or too fast.
Will you put $248k in my bank if I pay you 10% for 6 months? Yes. Will you put $24.8 million in my bank at 10% when only $250k is federally insured? I hope not. You would know I was desperate or crazy. Either one is sufficient to know that you should avoid me.
When you need money tomorrow, or even next week, you are going to get terrible prices for your loans. I lent you $1,000,000 on your $3 million house for 30 years at 7.1% and you have an 820 Credit rating and make $750k a year. Normally I could sell the loan but it would take some time to sell it at par. Why am I selling it now and demanding cash within 2 days? Is there something wrong with the loan? Or can you negotiate a better price because I am desperate? Probably one or the other.
SVB had few mortgages and had a bunch of fairly complicated loans to non-public companies. Those loans are about 1,000 times harder to sell than a generic mortgage.
So, SVB couldn't sell many more securities, couldn't sell many loans, and needed cash NOW to pay the depositors who are leaving.
In some ways, it could be a simple run on the bank.
My bank has a ton of core deposits so we are not at risk of losing too many deposits at one time. I really am not familiar with how the bank would work without significant core deposits. It is harder for them than for me. That would magnify the problems mentioned above.
Even if the most complicated loans are solvent, that difficulty in selling or using them as collateral means they're no bulwark against needing liquidity.
Unless they are pledged as collateral in advance then they are illiquid. Even so, getting more than 50 cents on the dollar is almost impossible unless it is a plain vanilla first mortgage.
Yes, Financial Times highlighted they dodged systematically important - that charming regulatory arbitrage you have in the USA. So escaped Basel III full application.
No bank could survive a bank run like that.
We could survive a 40% loss of deposits but that is because we do have some plain vanilla mortgages and we could raise insured deposits by paying high rates.
SVB couldn't.
Certainly but one can question if a run would have started had the point of departure been different, SVB would not have had the funding profile it had if it were under Basel III.
On other hand its specific market and stresses ... perhaps regardless would have been stuck.
How could the 16th largest bank in America with over $200 billion in assets have been seemingly so illiquid? Is there no banking requirement for a bank to have substantial near cash highly liquid (e.g., T-Bills) in secondary reserves? It is such a shallow illusion (only to be believed by people with no understanding of finance) to think that long term government debt is somehow a safe investment. In fact, it is one of the most leveraged, high risk investment available.
See NeilWilson, there are regulation but their funding model got in trouble - and this bank is a regional one under lighter regulation than the names you likely know.
It's not clear yet if SVB really did anything wrong.
Peraps. Adam Tooze suggests regulators were, in fact, asleep at the wheel:
https://twitter.com/adam_tooze/status/1634549484544749571
Regulators can not prevent all single institutional failure (or rather it is inefficient use of resources), they can prevent structural market risk which is the appropriate focus.
SVB was in the process of a capital raise that had the bank run not accelerated, might have paid off and completely avoided. As Financial Times reporting rather clarifies: https://www.ft.com/content/6943e05b-6b0d-4f67-9a35-9664fb456504 - had the run not accelerated or had they started their placement effort maybe a week earlier perhaps the run would have been avoided as they were close before both run and stock price collapse accelerated.
Fundamentally as NeilWilson has explained, their extreme reliance on jumbo deposits above insurance limite (to quote comparative from the FT)
"https://www.ft.com/content/6943e05b-6b0d-4f67-9a35-9664fb456504
At the end of 2022, SVB estimated that almost 96 per cent of its $173.1bn in deposits exceeded or were not covered by FDIC insurance. By comparison, Bank of America has estimated that around 38 per cent of its $1.9tn in deposits were not covered by FDIC insurance."
Useful comment in FT on the manner in which the situation developed with SVB as a collateral damage of market pecularities from the Covid period
Opinion Banks SVB’s collapse is not a harbinger of another 2008
ROBERT ARMSTRONG
"https://www.ft.com/content/0e1c671f-7998-4a55-a44e-edf07193616d
SVB’s problems began with the investment boom that followed the start of the coronavirus pandemic. As the go-to bank for California venture capitalists and start-ups, it was flooded with billions of deposits from young companies flush with investors’ cash. There was so much money — almost $130bn in new deposits in 2020 and 2021 — that SVB could not lend it all out. Instead, they invested much of the money in long-term US government-backed bonds. The bonds have no credit risk, and because SVB’s deposits cost almost nothing, they were profitable, too, despite paying only a few percentage points of interest.
But this balance sheet structure could only work while rates remained low. As the Federal Reserve battled inflation and rates rose, deposits became more expensive. In just the past year, SVB’s deposit costs rose from 0.14 per cent to 2.33 per cent. Meanwhile, the yields on its long government bonds didn’t budge. A profit squeeze was looming."
The placement decision relative to an inrush of liquidity is more understandable when put in a specific timing context and a specific market context of an entity itself highly exposed to a specific sectoral niche and with a very unusual specific funding as well as lending structure.
SVB's failure is a blip. Like LTCM's failure was.
It would seem to me that the 2020-2021 run up in deposits (primarily from tech start ups parking their VC funding) would be a giant red flag for a financial institution, and that investing the proceeds in long term T bonds to be highly risky given the likely instability of the level of deposits.
Complex loans to mostly a limited industry (Silicone Valley) also seems to be a highly risky and mostly illiquid, especially if they are not in a form and structure that can be readily syndicated. Banks earn a much higher ROE by syndicating loans they originate and skimming fees on sell down. I never experienced large banks willing to retain a large hold position long term, and usually needed assurance from their syndication team that the desired hold position can be realized.
Perhaps I am wrong, but I never heard of large investment grade banks funding themselves with loans directly secured by assets, but are funded by unsecured borrowings supported by covenants and a negative pledge.
I do feel sorry for those entities who may have been contractually required to leave large balances at SVB as part of their business or borrowing arrangements with the bank.
I saw an interesting little side note yesterday - in addition to some startups here in WA (readily predicable), some of the wineries may be in trouble. Wineries getting VC money?!? Yow!
Pingback: Was SVB management incompetent or just unlucky? – Kevin Drum