Separating term premium from inflation risk is a distinction that only holds if you believe the supply pressure resolves without a monetary response. It has not, historically.

The obvious read on 24 September is reassuring: yields up, breakeven compensation down 0.020 to 2.33%, Brent off 3.18% to 101.856. Nobody is paying up for inflation protection, so the long end must be charging for duration and issuance. Fine as description. As diagnosis it assumes the two things are independent variables, and they are not. Term premium driven by fiscal supply is compensation for the risk that a sovereign issuing more paper than the private sector wants at the prevailing real rate eventually stops trying to clear the market at that rate. The 30y at 5.47% is the price of that doubt. The path from there to inflation is not immediate, and that is exactly why breakevens do not register it.

The front end is the more interesting number. A 2y at 4.87% against an effective funds rate of 3.88% is not a market pricing cuts, it is a market pricing the Fed as having already overstayed its position on the easy side. That reading fits 4.1% unemployment better than the alternative does. It also fits a balance sheet that stood at 6,747,704 million in assets and grew on the week rather than shrinking. The statement says one thing about restraint. The portfolio says the runoff is over in practice if not in language.

So the curve is steepening for the honest reason: the front end thinks policy is too loose and the long end thinks the debt stock is too large. Both can be right simultaneously. That is the fiscal dominance configuration, and it does not resolve through patience.

Gold at 4284.97 and silver at 64.308 sit awkwardly beside falling breakevens. TIPS compensation measures expected CPI prints under an assumed reaction function. Metals measure doubt about the reaction function itself. When those two disagree, the disagreement is the information.

The mispriced leg here is credit. High yield OAS at 2.8pp is a market that has looked at a 10y at 5.18% and a 30y at 5.47% and concluded the cost of capital has not changed for anyone who borrows. Corporate refinancing happens against the long end, not against the funds rate. Either duration is wrong about the fiscal arithmetic, or spreads are wrong about what a 5.47% thirty year does to leveraged balance sheets over a refinancing cycle. My money is on spreads.

Expect the steepening to continue and the 2y to keep sitting above the funds rate until the Fed either validates it or explains why it is choosing not to. The explanation, when it arrives, will be about the labour market. It will be about the debt service line.

The long end is not the interesting part of this. A 2y at 4.87% against an effective fed funds rate of 3.88% is the market instructing the Fed to hike, and it is doing so in a week when the 10y breakeven fell to 2.33% and Brent dropped 3.18% to 101.856. Two-year notes do not carry meaningful term premium. That gap is a forecast of policy, not a supply story, and the supply story at the 30y end gives the Fed a convenient way to describe the whole curve as something happening to it rather than something being demanded of it.

Ask what would justify the front end. Unemployment at 4.1%, unchanged. Payrolls at 159,075 thousand, up 162. That is a labor market doing exactly what a central bank says it wants: adding jobs without visible wage panic. High yield OAS at 2.8pp says credit sees no stress worth pricing. Nothing in that picture asks for a higher policy rate.

What does argue the other way is the inflation data itself. The August CPI index rose 1.318 points and core rose 0.976, and core adding that much is not an energy artifact. That is the part of this I would not wave away. Prices are still accumulating in the index while the market's inflation compensation falls, which means the bond market has decided the Fed will suppress that gap by force.

Look at where the consumer price actually comes from right now. Crude fell on the day and over the week, US crude stocks excluding SPR built by 2,969 MBBL, and the EIA has the country on track for record production this year. Yet the EIA also says elevated crack spreads and crude prices are pushing pump prices up, and ERCOT load is running near record highs. Refining margins and a strained grid are not interest-rate-sensitive. You can take the funds rate to wherever the 2y says and you will not add a hydrocracker or a transmission line. You will, reliably, take hours off the bottom of the wage distribution.

The bill also goes abroad. Over five days USD/MXN rose 2.63%, USD/CAD 1.11%, USD/PLN 1.01%, USD/HUF 1.06%, and AUD/USD fell 1.42%, with the broad dollar index at 119.5133. Nothing happened in Mexico City or Warsaw. The US 10y went to 5.18% and the 30y to 5.47%, and that is the global funding rate whether or not anyone outside Washington voted on it.

My position: the front end is wrong and I would fade it. The Fed has not moved, DFF is flat, and the data that would license a hike is a core index still grinding higher against an employment picture that is merely adequate. If the Fed ratifies 4.87%, it will be tightening into refinery margins and electricity demand, and paying for it with jobs.

The 10y going 7bp cheaper is the least informative print on the screen. The yen is the one worth paying for.

USDJPY fell 0.98% to 157.295 on a session when US 10y yields rose 0.070 to 5.18% and the 30y rose 0.070 to 5.47%. That is backwards. EURJPY fell 0.89%, GBPJPY 0.79%, while EURUSD rose 0.10% and GBPUSD rose 0.20%. So this is not a dollar move dressed up as a yen move. It is yen-specific, and it arrived on one day: five-day USDJPY is still +0.26%. Something is being taken off, fast, in the funding leg rather than the asset leg.

The term premium reading of the long end is comfortable because nobody has to be wrong for it to be true. It is also unsupported by the corroborating prices. WALCL rose, by 1,156, so runoff mechanics were not the marginal seller this week. High yield OAS at 2.80pp widened 0.07. Seven basis points on a 280 handle is not a risk premium event, it is a drift. Genuine supply indigestion shows up in funding and basis before it shows up in a commentary about the 30y, and it has not.

The falling breakeven leg is mostly an oil print. Crude stocks excluding SPR built 2,969 MBBL on the week, Brent fell 3.18% to 101.856, WTI 2.60% to 95.786, and the EIA is flagging record US crude production for 2026. Inflation compensation at 2.33% dropping 0.020 against that is arithmetic, not a demand signal. Strip it out and the nominal move is almost all real rate.

Which leaves the front end, and I do not have a clean story for it. The 2y at 4.87% sits roughly 99bp above 3.88% effective funds with unemployment unchanged at 4.1% and payrolls up 162 thousand. That is a tightening path priced against a Fed that has not moved and a labour market that has not cracked. Either the 2y knows something the August data does not, or the curve is expressing something that is not a policy forecast at all.

The crowded position is the steepener. 2s10s at +0.31pp, wider by 0.050 on the day, is a trade that has worked and therefore is held by people who did not size it for a front-end repricing. The pain trade is a flattener driven by 2s, not by 10s.

Gold down 2.13% over five days and silver down 2.94% while copper is up 1.07% tells you the same thing as the yen. Crowded length is coming out of the metals that carried. That exit is narrower than the entrance. It always is.