Bank Failures: Silicon Valley Bank

Silicon Valley Bank, America's 16th largest bank, is the first of several major banks to fail.

By Robert Thorpe

This is the start of a series on recent bank failures.  I don't know yet how many articles there will be in the series.  It depends on how many more banks fail!  Three fairly large US banks have failed so far, here I'll concentrate on the largest - Silicon Valley Bank.

Bonds

It's first necessary to understand bonds.  This is more difficult that it seems.  If more people in finance understood bonds properly this crisis would not have occurred.  If you already know how bonds work feel free to skip this section.

Bonds are a way that large organizations like governments and big businesses borrow money.  A bond is an agreement to pay a certain amount per year, those payments are called coupons.  A bond usually has a duration - it has an end date.  At that end date the principal part of the debt is repaid.  The buyer of the bond is lending to the issuer of the bond.  For example, suppose that the government sells a bond that pays $2 per year.  That bond has a duration of 10 years and after that time it pays $100 to the holder.  When this  bond is first sold a buyer pays $100 for it.  That means this person will receive $2 per year (therefore 2%) and after 10 years they will get their principal back.

Bonds are frequently issued, usually that is done by auctioning the bonds to buyers.  After that the bonds circulate - people can buy and sell them on a second-hand market.  Anyone who has enough money can buy a bond from this market.  The market in US treasury bonds is one of the largest in the world.  The US treasury issues many different durations of bond.  There are 2 year bonds, 3 year bonds, 5 year bonds, 7 year bonds, 10 year bonds and even 30 year bonds[1].  Since bonds usually pay the same every year this makes things very stable for the government.  Because of this market you can buy a bond of any duration.  You can buy a 8 year bond even though the treasury doesn't sell one because you can buy a 10 year bond that is 2 years old (and therefore has 8 years left to run).

The prices of bonds fluctuate after they are sold. For example, let's suppose that the bond I described above was originally sold for $100. Now, a few years it's now worth $110. That bond still pays $2 per year. But, the interest rate the bond-holder receives is 2/110 = 1.818%. This is the “current yield” or “simple yield”.

Bonds rise in price when interest rates fall, and they fall in price when interest rates rise.  This is because of competition.  Bank accounts are in competition with bonds.  You can obtain interest from both of them.  However, bonds always pay the same coupon.  So, if interest rates on bank accounts rise then bonds become less attractive.  That leads to them becoming worth less.  Similarly, if interest rates on bank accounts fall then bonds become relatively more attractive.  Let's say a bond pays 2% in coupons and the interest rate is 2%. Then the interest rate falls to 1%, but the bond still pays 2%. If you own the bond you're getting an interest rate above the market rate. You can keep the bond for that reason. Or, you can sell the bond to someone else who will likely pay more than the issue price for it because it's paying an above-market interest rate.  For this reason the yield that this second buyer obtains will be less than 2%.

So far I have talked about bonds that pay a fixed coupon.  There are also bonds that pay a coupon that is proportional to the inflation rate.  These are usually called "index-linked bonds" because they are linked to an index of consumer prices like the CPI.  The US treasury calls them "TIPS" which stands for "Treasury Inflation Protected Securities".

Silicon Valley Bank

Silicon Valley Bank (SVB) specialized in banking for high-tech startups.  They also provided banking for the venture capitalists who fund startups.  Startups are risky, much more than half of them fail.  For this reason banks generally don't lend them money.  Instead startups raise money from venture capitalists and share offerings like IPOs.  Most banks operate by lending to the same group of customers who have accounts with them.  SVB could not do that because the startups it worked with were too risky (and they were getting VC backing anyway).  So, they had to look elsewhere.  Over the last few years the bank balances they were providing grew very quickly, they approximately tripled.  That meant SVB had to find assets to invest in.  SVB offered a few mortgages to tech company founders, mostly though they bought bonds.  They bought treasury bonds and also mortgage backed securities.

This was their downfall.  They forgot about the interest rate risk associated with bonds.  They bought large amounts of bonds in 2020 and 2021 when interest rates were very low.  Throughout 2022 the Fed increased interest rates from about 0% to just above 4.5%.  This made the coupon payments of existing bonds small in comparison.  As a result, the bonds SVB owned fell in value.  SVB had to admit this by law and did that in a footnote of their 10-K form[1].  The footnote essentially said that the bank's liabilities were roughly equal to its assets - with virtually no buffer.  Eventually people noticed.  Customers started transferring funds to other banks.  When a transfer like that happens the originating bank (in this case SVB) must pay the other bank in reserves.  This meant that they had to sell bonds to obtain reserves.  That meant selling bonds at low prices and realizing losses.  This process continued until the bank was bankrupt.  Its local regulator then put it into receivership and appointed the Federal Deposit Insurance Corporation (FDIC) as the receiver.

Matt Levine put it like this:

"Basically it was as reckless as it is possible to be with a business model of 'take deposits and invest them in US Treasury bonds.' Which, until recently, might not have seemed that reckless!"

SVB were told about their recklessness.  Bloomberg tell us:

"In late 2020, the firm’s asset-liability committee received an internal recommendation to buy shorter-term bonds as more deposits flowed in, according to documents viewed by Bloomberg. That shift would reduce the risk of sizable losses if interest rates quickly rose."  

They were also threatened with a ratings downgrade by Moody's.  SVB's chief risk officer left the company in April 2022 and was not replaced until January 2023.

Few commentators have discussed index-linked bonds, the TIPS bonds that I mentioned earlier.  If SVB had bought those rather than regular bonds it would not have got into so much trouble.  SVB's problem wasn't that it bought treasury bonds, it was that it bought the wrong sort of treasury bonds.

SVB did invest in mortgage backed securities.  However, the situation here was not like 2008.  In this case the mortgage backed securities were of a good quality.  The borrowers are still paying back their home loans.  The problem with the mortgage backed securities was the same as the problem with the treasury bonds.  They pay a fairly fixed coupon so when interest rates rise they become worth less.

Blaming the Runners

There has been plenty of finger pointing in the last week.  Amongst this, the claim that the customers are to blame is the silliest.  As a matter of common sense nobody should keep a balance with a dodgy bank.  The holders of bank accounts were the victims of the reckless behaviour of SVB.  We are told that Silicon Valley startups and VCs are "herd animals''.  The problem here was supposedly that they all decided to do the same thing at the same time.

Each customer who went to SVB to withdraw was perfectly within their rights.  They were protecting themselves and I don't see how they can be blamed for that.  Their behaviour was not irrational or malevolent.  Nobody can be expected to keep their wealth in a bank teetering on the brink of collapse.

It is true that they could have chosen not to withdraw.  But, that option was never feasible.  The VCs and founders could have met and agreed to keep their mouths shut and keep their balances at SVB.  How could such an agreement be policed?  Of course, it couldn't be policed, so each individual company would take the decision to withdraw its own money.  It was said that this run was "orchestrated" through a discussion on the Slack platform.  This was not so much collusion as dissemination of information.  The information itself was already public in the footnote of SVB's report.

We can imagine the situation where a bank is bankrupt for many years.  However, this bank continues to trade because the customers don't notice.  It then recovers and nothing happens.  We can imagine this, but I have never read of a time where it has actually happened.  Bankrupt banks are always found out sooner or later - and usually sooner.

What really was disgraceful was the cries for a bailout afterwards.  The silicon valley millionaires got their begging bowls together and went to politicians.  They lamented the death of the startup on twitter.  They told everyone who would listen that by not bailing out the customers of SVB they would destroy the technological edge of the US.  Allegedly, modern mainstream economic policy is "data driven" these days.  It supposedly rests on modern econometrics.  I never saw a single empirical study or even a simple regression analysis to lend weight to these claims. In my view, nobody should have listened to all of this.  Sadly many people did and SVB's customers were bailed out.

Insolvency and Illiquidity

Was SVB “insolvent” or "illiquid"?  I have now read many articles about this.  Mainstream economists often say things about this that sound superficially wise and mysterious.  For example, they will say that there is a grey area between insolvency and illiquidity.  We don't have to be so mysterious, the real issue here is that it depends on how you define the terms.

Fractional reserve banks are *always* illiquid to some extent.  They don't carry enough reserves to pay out every customer if they all arrive at once and want to withdraw.  I'm not going to get into the various debates about fractional-reserve banking in Austrian economics.  I will say this though, everyone agrees that fractional-reserve banks are not perfectly liquid. SVB was no different in this regard.

A bank can survive a bank-run if it can sell its assets during the run.  By those sales it can obtain reserves to pay out customers.  It could sell those assets on open markets or to other banks.  However, that relies on the value of those assets on the market!  If the assets are not worth enough then this strategy fails.  For the same reason putting up those assets as collateral for a loan would also fail.

Whether SVB was insolvent depends on time.  I think that it is best to look at the present time.  By that definition a business is insolvent if its assets are worth less than its liabilities now.  By that definition SVB was insolvent, its bond portfolio and other assets were worth less than the amount it owed its customers.  Others define insolvency over a period of time.  This is the (hopeful) idea that although assets are not worth enough now they will become worth more in the future.  This is what created the grey area that many economists love.

Matt Levine takes the view that a bank becomes insolvent when it puts out a balance sheet that admits it's insolvent.  I think that's endowing balance sheets with magical power.  If a bank is insolvent then that's true no matter if a balance sheet displays the fact.

I usually agree with Levine, but I find his view on this strange.  He writes:

"One way to think of this is that US banks — especially SVB, but not only SVB — have had huge mark-to-market losses on their bond portfolios as interest rates go up, but it is traditional for banks to ignore those losses."

This is the worst appeal to traditionalism I've seen in years.

The Future

This is an enormous mess.  The future doesn't look very bright either.  The government have promised huge funds to cover bank losses.  Despite their claims otherwise, these funds are likely to come from ordinary people.  I'll explain more about this and related messes such as Silvergate Bank and Signature Bank in future articles.  To keep things in perspective though, nothing that has happened so far is comparable to the group of bank failures that happened in 2008.

——

[1] - Technically the US treasury call only the 20 year and 30 year ones "Bonds". The call the ones with shorter durations "Notes". They also sell securities with even shorter durations called "Bills" but those are not so important here. This area is full of terminological problems. "US Bonds" are not the same as "Treasury Bonds" which are not the same as "I-Bonds". Also, the UK government call their bonds "Gilts".

Other Articles Like This One

Savings the Overlooked Economic Function

The Speculator: Villain or Productive Member of Society?

Discover New Opportunities

Click here for the Austrian Economics Discord Server.

Click here for the Austrian Economics Discord YouTube Channel.