Global Bonds Just Suffered Six Straight Weeks of Selling — What's Breaking the Bond Market?

Photo: U.S. National Archives, Wikimedia Commons, Public domain

01 The streak itself

Six straight weeks is a long time to lose money in the asset that is supposed to be the safe one. Global bond indices have declined every week since mid-August, and the 10-year Treasury yield has climbed from around 4.4 percent toward the five percent mark — territory not meaningfully visited in this cycle. The selling is not a U.S. story: gilts and Japanese government bonds are under the same pressure, with U.K. long yields rising in step and the JGB curve steepening as the Bank of Japan's normalization removes the cap that held yields near zero for decades.

What makes a six-week streak more than a number is what it says about the marginal buyer. Bonds fall when there are more sellers than buyers at the last price — and every week, the market re-learns that someone who used to buy no longer does. This article walks the three pressures doing the selling, and then the harder question underneath: in a world where central banks are shrinking their portfolios, who exactly is supposed to absorb record issuance?

Analysis — not prediction. N43 and Hermes AI grounds every scenario in the documented record and verified reporting as of September 21, 2026; where evidence is incomplete we say so.

02 Pressure one: energy inflation repricing

The proximate trigger is the one covered in our companion analyses: oil spent last week above $109 on the Saudi pipeline attack and the Hormuz disruption before retreating toward $100 on September 21, and the Fed hiked on September 16 — its first increase since 2023 — precisely because energy was leaking into headline inflation. A bond market that had spent a year pricing cuts had to reprice for a hiking-or-at-least-higher-for-longer path. That repricing is the fastest kind of bond selling, because it re-prices the entire expected path of short rates, not just one meeting.

The energy channel does something subtler than the headline suggests. Sustained $100-plus oil does not just raise near-term inflation prints — it raises the uncertainty around every future inflation print, and bond investors charge for uncertainty through the term premium. When a central bank is forced to tighten into a supply shock, the risk distribution of policy outcomes widens in both directions: more hikes if the shock persists, deeper cuts if growth cracks. Wider distributions mean riskier duration, and riskier duration is sold.

03 Pressure two: the term premium comes back

For most of the past fifteen years, government bond yields were suppressed by something larger than fundamentals: quantitative easing. Central banks bought trillions in sovereign debt, not because it was a good investment but as policy, and their presence as a price-insensitive buyer compressed the compensation investors demanded for holding long-term risk. That era is over, and not rhetorically — the Federal Reserve, the ECB and the Bank of Japan are all running down their portfolios at a measured pace, and the Bank of Japan's exit from yield-curve control removed the most explicit cap of all.

What the six-week streak measures, in part, is the term premium re-emerging as a live variable. Investors holding 10- and 30-year paper now demand real compensation for inflation risk, fiscal risk and the simple fact that nobody is obligated to buy from them. The academic estimates of term premium turned positive in this cycle after years below zero; each re-estimate upward shows up as exactly what we are watching — a slow, grinding rise in long yields with no macro news required to justify any individual week. A six-week streak is what the unwind looks like in slow motion.

04 Pressure three: the supply deluge

Meanwhile the issuance side has never been heavier. Deficits in the U.S., U.K., euro area and Japan remain far above pre-2008 norms, and the net supply of new government paper reaching private markets each year is at record levels — gross issuance, minus maturing debt the Treasury re-issues, plus the bonds central banks roll off their portfolios and return to private hands. That last term is the structural novelty: quantitative tightening means the private market must absorb not just the deficit but a share of the existing stock.

The absorption math is the uncomfortable part. During the QE era, central banks took roughly a third of new issuance in several markets; domestic banks, pension funds and foreign official reserve managers took much of the rest at prices that reflected a captive buyer base. Today every one of those buyer classes is constrained: banks face capital rules that penalize duration, foreign official reserve managers are diversifying away from dollar and sterling debt, and pension funds that bought bonds as inflation hedges were burned twice by inflation surprises. When the natural buyers are full, yields rise until a price-sensitive buyer — a hedge fund, a household, a foreign private investor — decides the yield is finally worth the risk.

05 Is this breaking, or just repricing?

The word “break” gets used loosely. A market breaks when functioning fails — auctions tail badly, liquidity vanishes, funding markets seize. In September 2022 the gilt market genuinely broke: a leveraged liability-driven-investor base was forced to sell into a market with no buyers, and the Bank of England had to intervene within days. The 2026 U.S. Treasury market shows none of those signs so far: auctions are clearing, bid-to-cover ratios remain ordinary, and the selling is spread across weeks rather than concentrated in forced liquidations.

But proximity matters. Six weeks of one-directional losses is how leverage gets tested. The players to watch are the ones who own duration with borrowed money: hedge-fund basis trades, which are enormous and quietly repo-funded, and regional bank and insurance portfolios that hold long paper against shorter liabilities. A move from 4.4 toward 5.0 percent is survivable for unlevered holders — a mark-to-market loss, nothing structural. For leveraged holders, the same move is a margin call. The streak becomes a break only if that second population has to sell; there is no public evidence it has yet.

06 What would end the streak

Three plausible endings, in rough order of market probability. Valuation capitulation: yields simply get high enough — five percent Treasuries yield more than the S&P dividend yield by a wide margin — that private buyers step in without macro help; streaks like this historically end not with good news but with a price. Energy de-escalation: oil retreating durably toward $90 would let the Fed guide back toward hold-and-watch, removing the repricing pressure. Growth scare: the sharpest bond rallies start when the growth data cracks — the irony of the current setup being that weak jobs or spending data would end the selling instantly, at the cost of confirming the pessimists' growth thesis.

What we should not expect is a return to the QE-era configuration. The buyers who suppressed yields for fifteen years are structurally out of the market, and the fiscal trajectory guarantees the supply keeps coming. The six-week streak may pause; the regime of higher, more volatile term premia is the deeper fact this episode is documenting, and it does not reverse on any single Fed meeting.

07 What to watch on the tape

Watch the auction calendar rather than the Fed calendar. The next 10-year and 30-year Treasury auctions after a six-week selloff are the cleanest read on whether private demand exists at these yields: tailing stop-throughs and soft bid-to-cover would confirm the absorption problem; strong stats would confirm valuation capitulation is underway. Second, watch the gilt and JGB long ends independently — if they stabilize while Treasuries keep selling, the problem is dollar-specific (supply, fiscal); if they sell together, it is the global term-premium story and no single government can fix it. Third, watch SOCRATES-grade plumbing: repo rates, swap spreads and futures basis. Those markets show leverage stress days before the cash market does.

The Liberty Bond thermometer on the Treasury building in our hero photo was built for a simpler era: a government selling debt directly to patriotic households. The 2026 version of that question — who buys when the central bank will not — is the one the next six weeks will answer.