Nominal yields up and inflation compensation down is being read as a term premium event. The curve says otherwise. A term premium shock steepens; this one flattened, with the 10y-2y spread down to 0.33pp as the two-year moved 0.130 against the ten-year's 0.120. Supply scares hit the back end hardest. This hit the front end hardest. What repriced was the expected path of the policy rate.

Which is the more interesting story, because the policy rate is 3.63% and it has not moved. The market is marking up where it thinks that rate has to go, against an economy that keeps refusing to cooperate with the easing case. Unemployment at 4.1%, unchanged. Payrolls up 162 thousand in August. High yield trading at 2.7pp over, which is not the spread of a credit market bracing for anything. And the balance sheet grew in the week to September 9. Quantitative tightening ended as a practical matter some time ago; what remains is a run-off schedule that no longer runs off.

Add five-day moves of 8.38% in Brent and 9.35% in WTI, gold at 4349.42 and silver at 64.5, and you do not have restrictive policy. You have a funds rate set for a slowdown that has not shown up, and a bond market beginning to charge for the discrepancy.

The breakeven at 2.36% is the fact that cuts against this. Inflation compensation is not merely contained, it fell on the day yields rose. I take that less seriously than most people do. Ten-year breakevens measure the marginal TIPS buyer's demand for insurance, and that instrument is thin, distorted by the carry trades that live in it, and referenced to an index whose shelter and services construction has been argued over for a decade without resolution. When the metals complex prints where it is printing and the bond market's inflation gauge prints 2.36%, the two are not describing the same future. One of them is priced by dealers managing balance sheet. The other is priced by people taking delivery.

The 30y at 5.37% is the number that matters for the fiscal arithmetic, and it is the one that will decide this. Debt service compounds at the long end regardless of what the front end believes. A Treasury refinancing into a 5.37% long bond while the effective funds rate sits at 3.63% is running the exact configuration that eventually forces the choice: fund it at market, or arrange for the market not to be the price setter.

My position: the next repricing in the front end is not toward cuts. If the Fed eases again into a 4.56% two-year and an oil tape like this one, the long end will do the tightening the committee declined to do, and it will do it in fewer sessions than anyone has budgeted for.

The curve flattened. That is the part of this move that is being read backwards.

The two-year rose more than the ten-year, and the ten-year more than the thirty. A term premium or supply shock does the opposite: it steepens, because the extra compensation is demanded at the far end where the duration risk lives. This move was led from the front. What repriced was the expected path of policy, not the price of holding a long bond. And the thirty-year, the maturity you would expect to absorb a fiscal supply story, moved least of the three.

So the market is not worried about inflation. The ten-year breakeven fell to 2.36%. What the market did was mark up the policy path and mark down inflation as the arithmetic consequence of that path. That is not a warning that the Fed has eased too much. That is the market asking for tightness and pricing the disinflation it expects to get.

Look at what set it off. August headline CPI rose 1.318 index points against core's 0.976. The wedge is energy, and the energy is a spike that is already turning: Brent gained 8.38% over five days and then gave up 4.25% in a single session, with WTI up 9.35% and then down 3.87%. Crude stocks drew 391 MBBL while gasoline built 1,269 and distillates built 2,087. The EIA's own read is elevated crack spreads pushing pump prices, against record US crude production this year and LNG exports up 23% in the first half. That is a refining margin cycle, not a shortage.

Core at 0.976 index points is not noise and I am not going to wave it off. But if core were driving the yield move, breakevens would be rising, not falling.

Meanwhile fed funds effective sits at 3.63% and the ten-year is at 4.95%. That gap is a cost, and it does not land evenly. High yield OAS at 2.7pp says large corporate borrowers are untouched. The bill goes to households refinancing, to small firms borrowing off the long end, and outward. Over five days the rand weakened 1.05%, the forint 0.52%, the peso 0.50%, the zloty 0.43%, all while the dollar lost 1.74% to the yen. Nothing changed in Pretoria or Warsaw. The rate moved in Washington.

And gold at 4,349.42 rose 0.75% on a day when real yields jumped, with the thirty-year at 5.37%. That combination is a sovereign credit signal, not an inflation hedge.

The Fed should cut from here. Unemployment at 4.1% is unchanged and payrolls added 162 thousand, which is a labor market holding its level, not one generating a yield shock. If the FOMC treats 4.95% as a veto on easing, it will have tightened into a rolling energy spike on the say-so of a flattening curve that was telling it the opposite.

The curve flattened. That kills the term premium story.

A supply shock or a term premium shock steepens: you sell what you have most of, at the end where duration risk lives. What happened on 2026-09-10 was the opposite. The 2y moved +0.130, the 10y +0.120, the 30y +0.090. The 10y-2y spread narrowed to 0.33pp, down 0.060. The front end led. That is not investors demanding compensation for holding thirty-year paper. That is the market repricing the path of the funds rate, which sits at 3.63% effective and now looks stranded.

The breakeven confirms it from the other side. Ten-year inflation compensation fell to 2.36%, down 0.040, in a week when Brent gained 8.38% and WTI gained 9.35%. Crude up nearly a tenth and the bond market marks down inflation expectations. Either the breakeven is mispriced or the market has decided the policy response to an energy shock arrives fast enough to eat the pass-through. Given that the 2y did most of the work, the second reading is the coherent one. Yields up, inflation compensation down, means real rates, and real rates repriced at the front because the easing is being taken back, not because Treasury announced anything.

Credit did not blink. High yield OAS at 2.7, down 0.010. No growth scare in that print. If this were the market pricing a Fed that eased into a fire, spreads would be the first thing to widen, and they went the other way. This is a rates-only event.

It is also not American. The gilt future fell 2.14% over five days, the Bund 1.38%, the ten-year Treasury 1.24%. Gilts worst, which is the usual order when global duration goes offered and the long end has a structural bid that steps away.

The crowded position here is the steepener. It has been the consensus rates trade on the logic that cuts anchor the front and supply pressures the back. The front end just moved more than the back and the spread compressed to 0.33pp. That is a lot of people getting the direction right on yields and losing money on the structure.

One thing does not fit. Yields up hard across the curve and USD/JPY fell 0.60% on the day, 1.74% on the week, with EUR/JPY down 1.87% and GBP/JPY down 1.68% over five days. The trade-weighted dollar has gone nowhere. Rate differentials widened in the dollar's favour and the yen rallied anyway. That is positioning liquidating, not macro. Gold at 4349.42, up 0.75% against rising real yields, says the same thing.

The rates move is a policy repricing. The FX move is someone getting out.