Long-term US bonds just hit an all-time low: why the 'safe' part of your portfolio can fall harder than stocks
4 min read by Opthest
This week the US 10-year Treasury yield touched 5.2%, its highest level since 2007, while the 30-year climbed to 5.48%, a high last seen in 2004. Over the same stretch, the iShares 20+ Year Treasury Bond ETF (TLT) — one of the most common ways to hold long-dated US government debt — fell to its lowest level since it launched in 2002. Meanwhile, according to Investment Company Institute data, tens of billions of dollars moved out of equity funds and into bond funds in September, but that money went mostly into short-duration government debt: whoever sought shelter at the long end of the curve got hit hardest. The paradox is real — just as the market “rushed to safety” in bonds, the slice of bonds that looks the most conservative, government debt, had one of the worst stretches in its history. The reason is duration.
What is “duration,” and why does it matter more than the rate itself?
Duration measures how much a bond’s price moves when market yields change: as a rough approximation, a duration of 17 years means a 1-point rise in rates pushes the price down roughly 17%. The farther out the maturity, the more a bond’s future cash flows (coupons and principal) get discounted at a rate that just changed, and the more their present value moves. A 3-month government bill has a duration near zero — its price barely budges when rates move. A 20-30 year bond, like the ones inside TLT, has a duration well over 15 years: the same rate move produces a price effect many times larger. It isn’t the level of the rate that makes the difference — it’s how long you have to wait before the principal comes back.
Why can a government-bond ETF fall harder than an equity ETF in a matter of weeks?
Because TLT’s sensitivity (its duration) is a huge number next to that of most stocks or short-term balanced funds. A rise of just a few tenths of a percentage point in the 10-to-30-year yield, multiplied by a duration of 15-20 years, produces a double-digit price decline in the fund — a move you’d normally associate with a stock market correction, not the “cautious” part of a portfolio. That’s exactly what happened to TLT: not a credit event, not a default, simply the mechanics of duration amplifying a small rate move into a large price move.
If half my portfolio is “in bonds,” am I automatically safe?
No — and that’s exactly what the “bonds” label hides. A balanced 60/40 portfolio with its bond sleeve concentrated in long-dated government debt can carry more interest-rate risk than an 80/20 portfolio holding short-term bonds. “Bonds” isn’t a single, uniform risk category: two bond funds can behave completely differently in the face of the same rate move, depending on the duration they carry. The capital flows seen this September — toward short-term government debt and high-quality municipal bonds, not toward the long end of the curve — is proof that professional money draws exactly this distinction, not a blunt “stocks vs. bonds” one.
What this article is NOT
This isn’t a forecast of where rates go from here, nor a suggestion to sell long-dated bonds or rotate into short-term ones. The yield levels and TLT’s all-time low cited above are September 2026 market data points reported by industry sources — they’re here to show that duration is a real, measurable risk, not to suggest a view on where the 10-year will be in six months. Rates sitting at decades-long highs doesn’t imply they have to fall, nor that they have to keep rising.
How do you see this risk on a portfolio you track in Opthest?
You don’t need to compute duration by hand to notice it — you just need to look at how it has actually behaved historically. Opthest’s backtest, volatility, and CVaR analyses work on each tracked position’s real historical returns — including any bonds or bond ETFs you hold — so a fund’s rate sensitivity, like TLT’s, already shows up in full in its past risk numbers, with nothing to estimate separately. It’s the same principle behind risk-adjusted return: looking at how a portfolio actually behaved rather than at its label. A long-dated bond fund that moved like a stock during a month of rising rates will show it in its own volatility and drawdown numbers, whatever it’s called. Like every Opthest tool, it remains an algorithmic output on historical data — educational/informational content, not personalized financial advice, with no guarantee of future results.
“Bond” isn’t a synonym for “safe” — it’s a synonym for a different kind of risk than equity risk, one that can be just as large depending on how far out the maturity sits. The 10-year Treasury at its highest since 2007 just reminded anyone who thought they’d found shelter.
Sources: Benzinga, “TLT ETF Hits Record Low as Treasury Yields Near 5.2%”; Infostream Global, “Bond Market Turmoil: What Changed in 2026”.