The bond market is not being complacent about oil. It is being right about oil and wrong about the Fed.

A 4,450 MBBL draw against a stock of 424,460 does not tell you anything about the average price level over the next decade, and a ten-year breakeven that moved on it would be a broken instrument. Look at what EIA has been publishing around the edges: LNG exports up 23% in the first half on new capacity, uranium output more than tripled in 2025, longer laterals lifting Permian crude and gas, eight liquids pipelines finished since the start of last year, Dangote pulling barrels out of Nigeria. That is not a scarcity market. That is a logistics and refining-margin market, which is exactly what elevated crack spreads describe. Brent at 97.88 on a week like that is a crack story wearing a crude costume. The 2.35 breakeven has nothing to apologise for.

The fragility sits somewhere else. In July, core CPI rose 0.724 index points while headline rose 0.245. The entire gap was energy holding the top line down, and energy has now stopped holding the top line down. Headline converges up to core, not the reverse. Central banks cut off headline and explain themselves off core, or the other way around, depending on which one is lower that quarter.

What actually interests me is the front end. Fed funds effective sits at 3.63 and the two-year yields 4.34. The market is not pricing complacency, it is pricing tightening, roughly seventy basis points of it embedded above the current policy rate. That is a hawkish position, and I do not think it survives contact with the reaction function.

Consider what the Fed would need to hike into. Unemployment at 4.1, payrolls up 162 thousand, high yield spreads at 2.65 with no sign of stress anywhere in credit. No victim, no crisis, no cover. Just an oil-driven headline print that every model on the twentieth floor will classify as supply-side and therefore ignorable, because that classification was validated in 2021 and never revisited. Meanwhile the balance sheet grew again last week, by 6,292 million to 6,737,204 million. Small, and in the direction that matters.

So the trade is not to fade the breakeven. It is to fade the two-year. The market has priced a Fed that will show up with a hike into a soft-landing labour market and tight credit, and that Fed does not exist. The instrument that has this right is gold at 4,429.825, which is not pricing inflation expectations at all. It is pricing the probability that expectations stop being the binding constraint.

Thirty-year at 5.25 against a ten-year at 4.77 is the same message from the other end of the curve. Own the steepener.

The bond market has this right, and for the wrong reason.

Breakevens sat at 2.35% while WTI added 9.09% in five days to 91.777 and Brent went to 97.88. The 10y drifted down to 4.77%, the 2y fell to 4.34%. That is a market saying an oil move does not change the Fed's reaction function. Correct. But read the rest of the EIA release list before deciding it was luck. Eight petroleum liquids pipeline projects completed since the start of 2025. LNG exports up 23% in the first half of 2026 on higher capacity. Longer Permian wells lifting crude and gas output. Dangote pulling more shipments out of Nigeria. That is a supply system adding capacity in every direction at once. A single-week draw of 4,450 MBBL to 424,460 does not overturn it, and neither does a crack spread, which is a refining margin, not a barrel shortage. Gasoline stocks fell 1,173 while distillates built 796. If this were a genuine product squeeze, distillates would not be growing.

Someone is capturing that margin, and it is not the household filling the tank. Elevated crack spreads mean the refining layer is taking a cut of a price move that consumers pay in full. A central bank that tightens in response to it taxes the same household a second time, on behalf of a shortage that does not exist.

So do not tighten for the oil. Tighten for the thing the oil is distracting from.

Core CPI in July moved 0.724 index points to 336.789 while headline moved 0.245 to 332.813. Headline is running slower than core. That is the inverse of an energy pass-through. Whatever is pushing prices is domestic and it is not crude, and it has been doing it long enough that calling it a residual has stopped being credible. Unemployment is 4.1% and unchanged. Payrolls added again in August. High yield OAS at 2.65 shows no credit stress worth naming. This is not an economy in the process of loosening, and the case for cuts, which I would normally be making, is not there in this data.

Fed funds effective is 3.63% against a 2y at 4.34%. The front end is already telling you it does not expect much easing. The 30y at 5.25% with the 10y-2y spread at 0.41 is where the actual inflation worry lives, and it is a worry about the persistence of the core number, not about barrels.

My call: the Fed holds, and holds longer than the consensus expects, and the oil rally will have faded from the story by the time anyone writes it up. The thing that eventually forces a cut will be the labor market cracking, and by then the argument will be about how fast, not whether. Watch participation and the bottom quartile of wages, not the pump.

Breakevens at 2.35 are the right price, and the ten-year at 4.77 is not asleep. The thing that is mispriced this week is the yen.

Start with what the oil move actually is. Crude stocks ex-SPR drew 4,450 MBBL. Gasoline drew 1,173, distillate built 796. That is an inventory and product-mix print, and EIA has already told you the pump story runs through crack spreads, which is a refining margin, not a demand impulse. Ten-year breakevens price average CPI over ten years. Nothing in a single weekly draw touches that horizon, and a market that repriced 2.35 on it would be the one making the error.

The July CPI data cuts the same way. Core rose 0.724 on the index against 0.245 for all items. Core outrunning headline means the pressure is not energy. If you want to be long the inflation trade, oil is the wrong leg to stand on.

And the front end is already carrying the hawkish weight. Two-year at 4.34 with effective funds at 3.63. That is not a curve pricing a cutting cycle. 2s10s at 0.41, two basis points flatter on the day, with the thirty-year at 5.25 doing the term premium work further out. Oil rallies nine percent and the curve flattens. That is a market pricing the Fed to lean against it, not a market ignoring it.

Now the yen. USD/JPY down 2.40% over five days, EUR/JPY down 2.14%, GBP/JPY down 2.52%. The crosses moved more than the dollar pair, so this is yen strength, not dollar weakness. Japan imports its energy. A nine percent move in WTI is a terms-of-trade hit and the currency rallied through it anyway. That is the anomaly, not the flat breakeven.

It is also not a risk-off unwind. High yield OAS at 2.65, a basis point tighter. MXN, ZAR, PLN, HUF all firmer on the week. EM carry bid while the funding currency appreciates is not the usual pairing, and when it happens the yen leg is being driven by something domestic. The BoJ has a market operations meeting on the calendar and current account projections out for September.

The exposure sits in one book: short yen, long EM carry, both working. That is the setup, not the confirmation. The door on that trade is narrower than the entrance because everyone walked in through the same one.

Gold is the tell against me. XAU off 0.96% on the day, silver off 1.15%, both softer on the week with crude up nine. Gold selling into an oil rally with breakevens pinned means real yields are holding and the bond market's read is intact. If gold turns while 2.35 stays put, the front end is wrong and so am I.

The combination of yen bid, EM bid, oil bid and gold offered does not map onto anything I can name. I am flagging it rather than explaining it.