Everyone will say they are “unrealized” losses. Some people (let’s call them “busters”) will argue this means real losses that already happened but banks are allowed to pretend it hasn’t. Other people (let’s call them “holders”) will argue these losses aren’t real and will only become real if the bank is forced to sell them, while if they manage to hold onto them for the full ten years then the losses never become real and vanish.
The truth is roughly that both sides are right, and the sum of their claims is much weirder than either subset.
The actual instrument in question is a little tough to get your head around. The first part is the basic bond mechanism: you give the government a thousand bucks, and ten years later they give you back that thousand dollars (guaranteed: they can print money so they will never default, only risk is they have to print so much money to pay you back that the economy explodes, and everyone has bigger problems at that point). Why would you do this? You wouldn’t, there is no upside, you only get back what you put in and it’ll be worth a little less because of inflation by then as well. Nobody does it, so right now what we have is not a real financial instrument.
So let’s make a first attempt at offering some upside: if you let them hold a thousand bucks for ten years, they’ll give you 10 bucks twice a year. Now you’ll take it - unless you think you can make more than 20 dollars out of your 1000 dollars each year by putting it somewhere else. What controls how much you can make on your 1000 dollars elsewhere? A lot of factors that all ultimately rest on the interest rate. Okay, so the government can’t just offer a flat 10 bucks twice a month, they have to offer something competitive with the current interest rate. But there’s the kicker: the current interest rate. Once you buy the bond, that amount is fixed, even if the interest rate later changes.
If the interest rates go up after you buy a bond, next years bonds will be offering higher per-year payments, so your bonds are inferior by comparison (they pay the same at the end, but less on the way, and they’ve already paid out some of the payments to you) and thus are worth a lot less. This is important because aside from holding them you can also sell them to someone else for any price you agree on, and whoever buys them from you gets the rest of the yearly payments and the final payout instead of you. They’re a hard thing to sell if interest rates have gone up, because you’re offering 20 bucks a year for five years while the government is offering 50 bucks a year for 10 years.
So: the money you get back eventually is worth less than when you handed it over because of inflation, but you’re getting small payments all along the way based on the interest rate at the time of the agreement. You care about interest and about inflation.
We have to stop for a moment and talk about interest rates and inflation. It is generally accepted that an increase in interest rates will cause a decrease in inflation a few years later. Confusingly, people will also say that interest rates move in the same direction as inflation but with a lag. It’s not that confusing though:
an increase in inflation at time t=0…
…will cause an increase in interest rates at time t=1…
…which will cause a decrease in inflation at time t=2…
(…which will cause a decrease in interest rates at time t=3…)
(…which will cause an increase in inflation at time t=4, and we’re back to step 1)
And so the cycle goes.
So in effect, you’re betting on this tension between interest and inflation resolving in your favor. In practice I believe the effect of inflation is smaller than the interest payments, so it’s also generally believed you always have a way out of the bet: just hold for the full ten years and the interest payments over that time will more than cover the inflation loss.
Except you can’t just hold on to the bet, because you’re a bank, and that thousand dollars you gave to the government is not your thousand dollars - it is some customer’s deposit, and they might want it back. So you better plan to have another thousand dollars somewhere else that you can give that customer, because the only way you can turn this bond back into money before the 10 years is up is selling it. And as mentioned before, if interest rates have recently gone up, your bond is not going to sell for anywhere close to breaking even.
That’s what that unrealized loss figure of 600 billion is: if you sold them for market value today, how much would you lose? As pointed out in the article, because interest rates were extremely low when these bonds were made, their yearly payout is very low. Because interest rates have risen rapidly, new bonds have much higher yearly payouts. And because interest rates have risen recently, we’re still in the lag period before inflation falls, so the final payout is also worth less (once again, I believe the effect of inflation differential is smaller here, and the price is I think mostly driven by the interest rate differential).
Concretely, right now, you could probably sell those bonds for no more than 75 cents on the dollar, and likely closer to 65 cents. If the Fed keeps raising interest rates like they’re doing now until the end of the current Presidential term (where someone else will get to tell the Fed what to do), we might get below 50 cents on the dollar, or even lower.
Banks bought 2 trillion dollars of an asset and right now that asset is only worth 1.4 trillion. That is, objectively, a huge loss. Point to the busters.
…But if we just hold the asset long enough, it is worth about 2 trillion again. Point to the holders, and this is why we call them “unrealized” losses.
…But if we can’t hold the asset (because, say, everyone withdraws at the same time), we have to sell at the current market rate, and those losses are forced to be realized. Point once more to the busters.
…But if we can rely on the FDIC or the government or other banks to step in and cover our withdrawals, we aren’t forced to sell - we can hold until the value returns, and pay back the FDIC or the government or the other banks then. Point once more to the holders.
And so it goes, back and forth between the busters and the holders. Who is right? In aggregate it’s both and neither, but at specific times for specific banks it could very visibly be one or the other. No wonder it’s so confusing!
There’s a lot more of these back and forths at every level that further complicate things. The government is jacking up interest rates so they’re causing the pressure… but banks know this dynamic exists and didn’t prepare for it so they made themselves vulnerable to this pressure… but the government regulations make these investments much more attractive to banks (very roughly: regulations say you only have to put up 0-20% of the value of these investments as collateral, for other investments it could be 100% or even 400% collateral) so the government pushed the banks in this direction… and so this cycle goes, too.
The fundamental dynamic is these bonds were an easy investment that turned into a giant Sword of Damocles over your head that’s growing by the day. In ten years you can step out from under the Sword, but any day now your depositors might panic and drop it on you, but if they do drop it the government might catch it before it kills you.
Thus, finally, some insight into the title of the post: very uncertain times indeed.