I only have a limited understanding of what went wrong at SVB (if I want to know, I will wait for an account put together after the incident), but the common theme is that it was a rapidly growing bank that badly mishandled its duration and liquidity risk.
As such, I have seen the argument to the effect “other banks have Treasury bonds like SVB, and the same thing will happen to them.” The idea is that there will be contagion to other banks, since they are all (allegedly) in the same boat.
The thing to keep in mind is that the American banking system is fragmented, and has a lot of smaller banks. Although SVB ended up near $210 billion in assets, it is still dwarfed by banks like Bank of America (just over $3 Trillion in assets at the end of 2022). It was small enough to qualify for more relaxed regulatory requirements that the politically potent regional/community banking community has carved out. So it appeared that SVB was able to manage itself pretty much how one would expect Silicon Valley residents would manage a bank.1
My banking primer is aimed at discussing larger banks that are professionally managed, mainly because those are the only banks I have ever come into contact with. Larger banks are divided into bureaucratic fiefdoms that interact with each other via highly constrained risk and liquidity management protocols, and they have risk committees as well as regulators watching over them.
Although banks will obviously have some interest rate risk (commonly called “duration risk,” although that might be a misnomer), banks are in the business of allocating credit risk and buying/selling liquidity. Taking duration risk is not a core business of a bank. Since they need a team of expensive traders to trade bonds as part of their liquidity management, they will take some duration risk, but not enough to overshadow the risk profiles of other units. (I discussed this in my earlier banking primers, such as this one.)
It is not enough to know the face value of bonds a bank owns, you need to know their duration, and how it compares to the interest rate risk profile of their liabilities and derivatives exposure. American banks that keep a lot of mortgages on their balance sheet are going to end up with a lot of expensive-to-hedge optionality that will likely be untouched, but hedging vanilla directional interest rate risk is not rocket science. (You get new DV01s that cancel out your old DV01s, and your friendly local risk manager is shoving a report with your DV01s in your face every morning and/or evening, never mind the live risk analytics.)
In order for there to be “contagion” from interest rate risk, you need a lot of banks thinking that Treasuries under 2% were an amazing investment opportunity, and they were willing to bet the farm on that thesis. Sure, interest rate losses will show up — partly because accounting reporting is somewhat removed from the true economic situation of a bank. (A bank does not follow the accounting convention of a bond fund, where everything would be marked-to-market daily.) But the issue is whether they are significant enough to make a dent in bank capital.
Some people might want to drag out the experience of banks in the late 1970s-early 1980s, where they were whacked by the interest rate spike. (The American Savings & Loan crisis can be partly summarised as one of the initial drivers of excessive gambling at small “Savings & Loan” banks is that they needed to make up for interest rate losses on their mortgage portfolios.) That was an era when nobody knew what duration even was, never mind having regulators watching the interest rate risk profile of banks. After 1994, bank regulators clamped down on interest rate risk.
I obviously have not been poring over bank balance sheets, so I leave the reader to do their own homework. That said, I remain far more concerned about credit risk. For modern banks, credit losses are the major driver of contagion. Defaults in the “real economy” weaken financial players, so markets start to unravel as participants start to doubt the credit-worthiness of their counter-parties. Unlike interest rate risk, credit risk seeps into financial statements as it takes time for borrowers to default and the losses to be recognised, and the ultimate recovery values are foggy.
In summary, if you want to scare me about banks, you need to scare me about credit risk.
As someone who graduated electrical engineering in the very early 1990s, I remember when Silicon Valley was synonymous with producing hardware like oscilloscopes and computers, and not crackpot theories about money and super-intelligent AI that is going to Skynet us.