Long-term treasury bonds are fairly volatile. That's how Silicon Valley Bank went under: their long-term bond holdings dropped in value enough to make them insolvent.

They can swing a few percent. So can gold. Either one could make a bank insolvent. In the long term treasury bonds are not volatile, especially if you hold them to maturity.

What does holding to maturity have to do with their current value?

It makes the current market price irrelevant because you're still owed the same amount on the same date.

The current market price is about what it is worth _currently_.

When your deposits are denominated in _current_ dollars, and that's what your customers can demand, then it doesn't matter that your expectation of how many dollars you are going to receive in 2035 is stable. It's about what our assets are worth right now, in case you need to liquidate them to satisfy withdrawal requests.

If you can contrive your deposits to be denominated in 2035 dollars, then long term treasury bonds are 'stable' in that sense.

Similarly, if your deposits are denominated in grams of gold, then gold is a stable backing for those.

If you have a mismatch between what you owe and what you own, then you need a thick equity cushion between your assets and fixed liabilities.

This conversation was not about banks when the comparison came up, and when I talk about long term value I'm not talking about bank reserves. (And even if you argue a stablecoin is like a bank, it's one with an utterly massive reserve ratio.)

If you're worried about short term value then you can use shorter term bonds if you want, whatever. It doesn't make a difference to the reason I brought it up in the first place, because either option is more stable than gold.

What do you mean by long term value? The current market value is typically the best estimate we have for their long term value.

No one is talking about bank reserves. I'm talking about assets.

> The current market value is typically the best estimate we have for their long term value.

It's not. The EMH has been empirically disproven in the 80s.

Could you please link me to the evidence? Which version of the EMH has been disproven?

EMH comes in multiple different strengths. The strongest version would be something comical like 'market prices are omniscient and perfectly predict future prices'. That's almost certainly wrong. Very weak versions are something like 'Don't bother actively trading on the news as a retail investor, because by the time you've heard them, the folks over at Goldman Sachs and the hedge funds and their computers will have traded on them a million times over already', and these are almost certainly true. (But even somewhat stronger versions are probably true.)

See Bob Schiller's work (for which he received the econ Nobel prize in 2013).

The “weak version of EMH” has nothing to do with markets being “efficient”, it's a property of random markets. Assimilating the two just Fama's motte-and-bailey fallacy.

When you say 'random' you probably mean that market prices are a Martingale? See https://en.wikipedia.org/wiki/Martingale_(probability_theory...

That's very, very related to being efficient.

> Assimilating the two just Fama's motte-and-bailey fallacy.

No, not at all.

> That's very, very related to being efficient.

“No, not at all”.

> What do you mean by long term value? The current market value is typically the best estimate we have for their long term value.

In situations where we still care about dollars, so no hyperinflation or total collapse of the United States, the current market value of a Treasury bond can't actually vary that much. And the amount it can reasonably vary is mostly proportional to how many years are left in the bond.

By the time your bonds reach maturity, you always have more dollars than you started with. Long term you always profit. And you get to choose what length of bonds you buy, so if you want to you can guarantee your dollars increase in the medium or short term on top of the long term.

> No one is talking about bank reserves. I'm talking about assets.

I'm saying you're too worried about "withdrawal requests" a normal bank would see.

> In situations where we still care about dollars, so no hyperinflation or total collapse of the United States, the current market value of a Treasury bond can't actually vary that much. And the amount it can reasonably vary is mostly proportional to how many years are left in the bond.

Well, it was enough variance to bring Silicon Valley Bank down.

> By the time your bonds reach maturity, you always have more dollars than you started with. Long term you always profit.

I'd be very happy to have you as my investor in some long term bonds---with terrible below-market-but-barely-positive interest rates.

> I'm saying you're too worried about "withdrawal requests" a normal bank would see.

Silicon Valley Bank saw massive withdrawals, because their liabilities exceeded their assets.

Again, worrying about the long term value is an entirely different problem from worrying about your assets temporarily shrinking 10%. I was talking about the former. SVB, an example of the latter, does not affect my argument. And again, SVB could have used shorter term bonds to avoid that problem.

> I'd be very happy to have you as my investor in some long term bonds---with terrible below-market-but-barely-positive interest rates.

Very funny. Look, that's one feature of Treasury bonds, not the only feature. They get pretty good yields compared to gold in the long term, and you can trust them a lot.