There's a lot to digest with the downfall of Silicon Valley Bank.
SVB's Failure is Different From Most
First (and contrary to my original assumption), bad loans were NOT the cause of the problem.
Dating back decades, every other bank failure I'm aware of has happened because of bad loans (or in the case of some small community bank failures, embezzlement). SVB is different.
SVB failed due to an old-fashioned bank run, exacerbated by its own deficient Asset - Liability management. So what caused the bank run?
The answer is really interesting, and ironically has its roots in interest rates held artificially low for over a decade.
Tech Startups Run on Venture Capital (VC), and VC Runs on Low Interest Rates
SVB's customer base was heavily concentrated in tech startups. Which are typically cash hogs of an exponential order until they really take off. Which means their cash for daily operations -- payroll, rent, keeping the lights on, etc., plus the cash for all the R&D -- comes from equity, often in the form of Venture Capital.
Venture Capital (VC) feeds on low interest rates. If Mr. Rich Guy is sitting on a pile of cash, and neither banks nor bonds are paying much of anything, and the stock market is priced for perfection, VC's potential returns start to look awfully attractive.
Yeah, you hit on only 1 out of every 4-5 investments, but that one is a gusher, because.....you typically get your investment return when the company you invested in is acquired by a bigger fish. Often funded by...you guessed it...another VC fund.
Problem is, when interest rates start to climb, Mr. Rich Guy's tolerance for VC's low hit rate decreases significantly. That's because he can invest his whole pile of cash at a more palatable rate of return.....which means it's no longer going into VC funds, which means that VC's source of cash to invest in tech startups begins to dry up.
So the tech startups are still running through significantly more cash than they can generate internally, but no longer have a sugar daddy to help them out. They have to start drawing down their deposits (which deposits were the result of VC funding) to keep the doors open. Trouble is, there's now a lot less VC capital coming in.
So SVB started to see its deposit base erode, slowly at first. Then it all hit the fan last week. The FDIC / Federal Reserve will receive bids on SVB today, and the new owner will open the doors perhaps as early as Monday.
Bad A/L Management Caused a Loss of Confidence, Which Caused the Run
So what is Asset / Liability Management? It's matching up the maturity of your loans and investments (assets for a bank) with the maturity of its deposits (liabilities for a bank).
At its core a bank takes in deposits and uses that money mostly to lend out. What it doesn't lend, it invests in bonds, and there are strict regulations on what bonds the bank can buy.
To do that profitably, the bank pays less for the deposits than it collects on the loans.
The problem with low interest rates is that the bank had to buy longer-term bonds to get any yield worth talking about.
The problem there is that longer-term bonds are MUCH more subject to fluctuation in market value, even if the ability of the bond issuer (the US Government, a large corporation, whatever) to make good on its obligation is unquestioned.
So SVB was chasing yield, and given the eroding deposit base, held longer-term bonds than was prudent. When they had to sell the bonds to raise cash to meet the slow drain on deposits, they took a big loss on market value -- roughly 10% of what they had the bonds on the books for.
That spooked some VC funds, who are large depositors many many times more than the $250K cap on FDIC insurance.
So they took their money out, advised their clients to do so as well, and the run was on.
SVB tried to issue more equity and/or debt, but ran into a loss of confidence in the investment community, and could find no one willing to buy, which led to the FDIC takeover.
In fairness to SVB, no bank in the world can come up with the cash to pay everyone if all depositors showed up at the same time.
But SVB was more vulnerable than most due to (1) its bond portfolio being longer term than it should have been, and (2) it didn't have a super-strong capital base to begin with, and (3) its depositors don't fit the typical profile -- roughly 85% of all SVB's deposits are in excess of the $250K FDIC cap.
Now What?
For reasons stated in my post above, I think all depositors will be made whole, regardless of the balances in their accounts. If the FDIC allowed depositors to take a hit here, you would see nastiness flow through the world financial system that makes 2008 look like Mardi Gras in the French Quarter.
Keep in mind, this is bailing out depositors, not the bank itself. Make the depositors whole. Let the bank fail.
Side Note on SVB Culture
SVB would appear to have had a really arrogant internal bro culture.
The CEO sold $3.6 million in shares just a few days before SVB sold the bonds that tipped over the first domino. The CFO sold about $575K at about the same time, and the Chief Marketing Officer sold a couple hundred thousand worth.
Those transactions are public record the instant they happen, and the fact that they would do that so close to what they knew would at the very least be a major hit to the stock price is just stupid.
Also, in 2021 a former VP in their own VC arm was fined about $50K for illegally tipping a friend to a pending transaction his area was working on. Rather than take his medicine, he started lobbying the judge for leniency, including submitting letters of reference and recommendation from several high profile residents.
Trouble is, those letters were fabricated. The moron lied to the judge, got busted, and is now serving 15 months for trying to mislead a federal judge. Turned a $50K mild embarrassment into a felony sentence that will change his life forever. It doesn’t get more stupidly arrogant than that.
Now, all of this is salacious and grabs lots of headlines. But none of it is material to SVB’s failure. It does, however, speak to a risk-taking culture that you just can't have at a bank.
Why Long-Term Bonds Are More Volatile Than Short-Term
OK, this gets geeky, and is rooted in exponential math. Skip this if you don't care why LT bonds are more volatile, and are happy just knowing that they are.
A bond pays interest and principal in fixed amounts on a fixed schedule. Typically, it's quarterly or semi-annual interest, with principal at maturity.
The market value of the bond is those cash flows, discounted to present value at today's interest rate, not necessarily the interest rate that prevailed when the bond was issued.
Here's a greatly simplified example. To keep things easy, we're making several assumptions:
- One bond at a 2 year maturity, and another at 10 years.
- Both are $10,000 face value, payable in full at maturity
- But there are no interest payments.
- We're assuming a flat yield curve -- 2% at issuance, instantaneously rising to 5% today.
The day it was issued, the 2-year bond had a value of $9,612
When market rates rose to 5%, the value fell to $9,070 -- a decline of about 6.7%.
The day the 10-year bond was issued, it had a value of $8,203
When market rates rose to 5%, the value fell to $6,139 -- a decline of about 25%
If it were a 30 year bond, the decline would be roughly 60%.
If you’re interested in the underlying mathematical formulas, I can detail. But I figured we were geeked-out enough already.