Having to fire sale bonds (to meet withdrawal obligations) that are selling at a discount because any potential buyers have better yielding instruments to buy results in solvency concerns from the relative excess losses.
As long as the bonds yield more at maturity than you paid for them (along with the cost of the money you used to pay for them, which is very cheap for most large banks), then holding them to maturity is guaranteed to pay off. Whereas we have no idea how cheap they could ultimately become on the spot market at any given time that the owner may need to fire sale them.
Why should depositors leave more money than they need for short-term working capital in the SVB at 0% when a Treasury bill pays 4% with less risk? When all rates were roughly zero, it was easy to just lazily leave everything in the bank; but when rates increased, the reward for leaving and the risk of staying (because the bank is now mark-to-market insolvent, and thus particularly vulnerable to a run) both increased.
When interest rates increased, the NPV of the SVB's assets decreased. Someone had to take that loss. Their managers and shareholders presumably hoped that would be the depositors, by leaving their money in a risky bank earning interest below the risk-free rate, slowly accepting the loss over time. The depositors had no economic incentive to do that though, and they didn't.
For emphasis, the SVB's problem isn't that their bonds are trading at irrationally low prices due to liquidity problems (i.e., a "fire sale"). The market price for their bonds has behaved exactly like a textbook NPV model, and similar assets trade with normal tight spreads; the price is just lower than they wanted. Hold-to-maturity accounting allowed them to ignore that for regulatory purposes, but accounting doesn't change reality.
> Their managers and shareholders presumably hoped that would be the depositors, by leaving their money in a risky bank earning interest below the risk-free rate, slowly accepting the loss over time. The depositors had no economic incentive to do that though, and they didn't.
That's why the grandparent comment mentioned a large bank buying SVB out. Large banks have diverse, non-business deposits, many of which tend to sit, and not chase yields. People in more precarious positions keep emergency funds in the bank, not in CDs or bonds that they can't easily cash out on a moment's notice, or stocks that are risky. People in even more precarious positions are constantly paying money to the bank at a steep interest to cover their credit card payments. This is really cheap money to the bank (and more easily liquidatable at par than low-yielding bonds), though it has to cover those in the most precarious positions who default.
I guess a lot of people hoped that a different large bank would buy the SVB or its assets at a price that would have maintained the accounting fiction that it was solvent. Such banks had no economic incentive to do that though, and once again they didn't.
Hypothetically because they are martyrs who also want whatever goodwill SVB created in its clients (which now become the clients of the other bank). But yeah, it's a stretch.
But that's pretty risky for a bank.
The mark-to-market hole at the SVB was pretty deep though, and their depositor base had just made itself look particularly unattractive. So maybe the FDIC will eventually manage to find a buyer, but it indeed looks pretty hard.
Quite true.
If you hold that bond to maturity in 5 years the US government will give you your $100 in full though.
The Fed has now started a program that lets banks use their treasury bonds as collateral against cash loans - if this was going a few weeks ago it would have saved SVB.
As in, does the program let you borrow cash from the government at a lower rate than the bonds are paying? How could that work?
The program is undercollateralized though, so if a bank fails with such a loan outstanding then the Fed loses money. That part is a subsidy.
The banks share (and bond) capital is on the line as well.
It's still a subsidy though. For example, if you lend my just-barely-solvent bank $100 unsecured, then I can bet it double or nothing on a fair coin flip. If I win then I keep all the gain, so it's +$50 EV to me. If I lose, then my bank fails and you take the loss, so it's -$50 EV to you. Banks are obviously regulated in many ways to discourage that kind of risk-taking, but this policy change slightly weakens that overall edifice.
I think the program might not work like you think? Beyond the initial decision to create the program, no one at the Fed exercises significant discretion. It's open to essentially all American banks, most of which are private businesses with the usual profit motives, whose managers decide when to borrow and how to use the proceeds. If such a bank gets in trouble, then its managers now have slightly greater incentive and ability to try gambling their way out, since their shareholders get all the upside and the Fed gets some of the downside.
https://www.federalreserve.gov/newsevents/pressreleases/file...
one-year overnight index swap rate plus 10 basis points - so call it (currently) 4.7%
What does this say about the rest of the banks in a similar situation wrt their balance sheet?