I checked my accounts last week the way I do every few days — not obsessively, just a habit — and one line looked wrong. My long-term U.S. Treasury bond fund, the single most “boring” thing I own, was down. Not by a lot. But down, on a week when I hadn’t expected it to move at all.
Bonds are supposed to be the calm one. The fund I bought specifically because I wanted something in my retirement account that wasn’t tied to the stock market’s mood swings. So when it moved anyway, I went looking for why, and found the answer sitting in the financial news: the yield on the 30-year U.S. Treasury bond had just touched about 5.33%, its highest level since 2007 — roughly nineteen years. Long-term borrowing costs for the U.S. government are the highest they’ve been in almost two decades, and my “safe” bond fund felt it immediately.
This post is about why that happened, and why it isn’t actually a contradiction. A bond fund losing value when rates rise isn’t a malfunction. It’s the fund doing exactly what a bond does. I just hadn’t sat down and done the arithmetic before.
What actually happened
A few things lined up at once, according to the coverage I read. The U.S. government has been issuing a large amount of long-term debt to fund a growing deficit, which means more supply of 30-year bonds hitting the market. Inflation has stayed above the Federal Reserve’s target for roughly five years running. And there’s a new Fed chair, Kevin Warsh, who took over earlier this year, which means markets are still recalibrating what his Fed will tolerate (Bloomberg, Aug 17, 2026). Put those three together and long-term yields climbed to a level not seen since 2007.
The ripple effects went beyond bonds. Nasdaq-listed stocks — especially AI and semiconductor names — fell harder than the broader market on the same news, because higher long-term rates mean future profits get discounted more heavily today (Detroit News, Aug 18, 2026). Rising yields don’t stay in their own lane. They touch everything that’s priced off the future.
The part that confused me: why do bond prices fall when yields rise?
Here’s the mechanism, in plain terms. A bond is a promise: lend this money, get a fixed interest payment, get your principal back at maturity. If you already own a bond paying, say, 4% interest, and suddenly new bonds come out paying 5%, your old 4% bond becomes less attractive by comparison. Nobody wants to pay full price for a lower-paying promise when a better one just showed up. So the price of your bond drops until its effective yield lines up with what the market now expects.
The technical term for how much a bond’s price moves per unit of yield change is duration. It isn’t the same as maturity, but it moves in the same direction: the longer a bond’s maturity, the higher its duration, and the more its price swings for a given change in yield. A 30-year bond has a much higher duration than a 2-year note, which is exactly why my long bond fund moved more than a shorter-duration fund would have on the same news.
I want to be careful here: what follows is arithmetic, not a forecast. I’m not predicting where yields go next. I’m just doing the math on what a given move would mean, using round numbers to make the mechanism visible.
The rule of thumb: price change ≈ −duration × change in yield. Say a fund has a duration of 20 years (a rough, illustrative figure for a long-maturity bond fund — not a specific claim about any one product) and yields rise by 0.50 percentage points: price change ≈ −20 × 0.50% = −10.0%. Now compare a fund with a duration of 3 years, hit by the exact same 0.50-point move: price change ≈ −3 × 0.50% = −1.5%.
On a hypothetical $10,000 position: about a $1,000 paper loss for the 20-year fund vs. roughly $150 for the 3-year fund.
Same news. Same size of rate move. A more than six-times-larger price swing, purely because of how long the bond’s payments stretch into the future. Nothing here predicts what yields will actually do — it just shows why “bond fund” isn’t one risk profile. It’s a spectrum, and length is the variable that matters most.
Where this goes wrong: bonds are not automatically “safe”
I think it’s worth naming the uncomfortable part directly, because I got tripped up by it myself. There’s a common assumption that bonds are the safe half of a portfolio — the thing that goes up when stocks go down. That’s true for credit risk (the risk a borrower won’t pay you back; U.S. Treasuries carry essentially none of that). It is not automatically true for price risk from interest rate moves.
In 2022, long-duration Treasury funds went through one of the worst stretches in their modern history, falling alongside stocks rather than cushioning them — the opposite of what a “safe asset” is supposed to do in a downturn. Rate-driven bond losses and stock losses can absolutely happen in the same year, for the same underlying reason: rising rates hurt long-dated assets of every kind, whether they pay a coupon or a dividend.
So the honest version of “bonds are safe” is narrower than the popular version: long-term government bonds are safe from default. They are not safe from price swings when interest rate expectations shift. If you’re holding a 30-year bond fund expecting it to behave like a savings account, the last few years — and the last few days — are a reminder that it won’t.
How Koreans and Americans reach the same bond differently
One thing I found genuinely interesting while reading about this: Americans and Koreans don’t buy the same 30-year Treasury the same way.
| Access point | United States | Korea |
|---|---|---|
| Direct purchase | TreasuryDirect.gov | Not typical |
| Retail route | Bond ETFs, brokerage | KRX-listed bond ETFs |
| Currency exposure | None (home currency) | Won-hedged or exposed |
| Tax treatment | U.S. domestic rules | Pension/ISA wrapper rules |
In the U.S., an individual investor can buy a 30-year Treasury bond directly from the government through TreasuryDirect, no fund wrapper needed. In Korea, retail access to U.S. Treasuries mostly runs through exchange-listed ETFs — funds that hold a basket of Treasuries and trade on the Korea Exchange like a stock. These funds typically come in two flavors: currency-hedged, marked with an “(H)” in the fund name, and unhedged. A hedged fund tries to strip out won-dollar exchange rate movement so the fund’s return tracks the bond’s dollar return more closely; an unhedged fund lets currency movement add or subtract from the result on top of whatever the bond itself does.
That’s a second layer of variability stacked on top of duration risk, and it’s a genuinely different structure from what a U.S.-based bond investor deals with. It’s also, I’d argue, an underexplained topic for English-speaking readers — most bond-investing content assumes a U.S. audience holding U.S. dollars.
Where this fits in my own accounts
I hold my investments across a small number of accounts: two personal pension savings accounts (연금저축), an Individual Savings Account (ISA), and an Individual Retirement Pension account (IRP). Across those, I split my money into a handful of category buckets — U.S. large-cap (S&P 500), Nasdaq, a dividend-focused bucket, cash and short-term instruments, a small “thematic” sleeve, and one long-duration U.S. Treasury bond fund.
That bond fund is a deliberate minority position — roughly a fifth of my retirement portfolio, not the core of it. I didn’t add it because I expect it to be my best performer. I added it because I wanted one holding in the mix that responds to a different set of forces than stocks do — interest rate expectations and inflation, rather than earnings and sentiment. Weeks like this one are the cost of that diversification, not proof that it failed. A 30-year bond fund losing value when the 30-year yield hits a two-decade high is the fund working exactly as designed, even if “working as designed” doesn’t feel great on the day it happens.
The bigger habit underneath all of this hasn’t changed: money still goes in on a fixed schedule, into the same target allocation, regardless of what the bond market or the stock market did that week. I don’t rebalance because of a headline. I rebalance because a set amount of time passed and the numbers drifted from target. This week’s Treasury news doesn’t change that process — it’s exactly the kind of thing the process is built to not react to.
FAQ
Can bond ETFs actually lose money?
Yes. Bond funds hold assets whose prices move inversely to interest rate expectations. If yields rise after you buy in, the market value of the fund’s holdings falls, and the fund’s price (and your account value) falls with it — even though nothing about the bonds’ credit quality changed.
Why do bond prices fall when interest rates rise?
Because existing bonds pay a fixed rate. When new bonds come out paying more, older, lower-paying bonds become less attractive at their original price, so their price drops until the effective yield matches what the market currently expects.
Is it a bad time to buy long-term bonds right now?
I can’t answer that for anyone else, and I’m not trying to. What I can say is that a bond bought at a higher starting yield locks in more income per dollar than the same bond bought at a lower yield — but it can still lose price value in the short term if yields climb further. Those are two different questions, and it’s easy to conflate them.
What’s the difference between currency-hedged and unhedged bond ETFs?
A hedged fund (usually marked with an “(H)”) uses currency instruments to reduce the effect of exchange rate movement on returns, so the fund tracks the underlying bond’s own performance more closely. An unhedged fund lets currency movement add to or subtract from the bond’s return, for better or worse.
How much of a portfolio should be in long-duration bonds?
There’s no universal answer — it depends on time horizon, how much price volatility someone can tolerate, and what role the bonds are meant to play. My own answer, for now, is “a small, deliberate slice, not the core.” That’s a personal choice based on my own goals, not a rule anyone else should copy.
What I’m taking from this
I don’t think I misjudged this fund. I think I hadn’t actually done the math on what “30-year duration” means in practice until a real headline forced me to. Now I have. The next time this fund moves on a rate headline, I’ll know exactly why, and I won’t mistake a duration effect for something being broken.
This is not investment advice.