The obvious read is term premium and fiscal supply. The curve says otherwise. A duration buyers' strike shows up as steepening, because the auction calendar concentrates in the long end and nobody is forced to hold thirty-year paper. What happened instead was a bear flattener: 10y-2y compressed to 0.39pp while every point on the curve backed up. When the front end leads a selloff, the market is not repricing the debt stock. It is repricing the policy path.

Look at where the 2y sits relative to the effective funds rate at 3.63%. That gap is not a forecast of cuts. It is a forecast of the absence of cuts, and possibly of something worse than that. The front end has spent this cycle assuming the reaction function bends toward easing at the first soft print. It now appears to be pricing a Fed that cannot deliver, with unemployment unchanged at 4.1% and payrolls still adding.

Breakevens are the anomaly. Ten-year inflation compensation moved to 2.4% while Brent held above a hundred dollars after a week that added more than eight percent, and the EIA is pointing at elevated crack spreads feeding pump prices. Crude stocks drew again. The inflation swap market is telling you that a three-figure oil price and record ERCOT load are a relative price story that washes through in a year. That is precisely what was said in 2021, when the supply shock explanation was correct in construction and wrong in consequence. The suppressed variable was never the level of CPI, it was the willingness of the committee to sit still while it ran.

So almost the entire 4.95% on the ten-year is a real yield, and it rose without the inflation leg. The reflexive interpretation is that real rates reflect growth or fiscal issuance. I think they reflect the market solving for the rate at which the Treasury can still be financed while the Fed is not easing, and that solution is drifting higher independent of anything the statement says.

The balance sheet is the tell nobody read. WALCL ticked up to 6,740,619 million, which means the run-off is over at the margin, and the long end sold off anyway. Liquidity was added and duration got cheaper. That combination does not happen when the constraint is the policy rate. It happens when the constraint is the debt.

Credit did not blink. High yield OAS at 2.7pp tightened on the day. That will look, in retrospect, like the last comfortable print.

My position: the front end has this right and breakevens have it wrong. The flattener has further to run before the fiscal arithmetic forces the steepener, and it forces it through cuts the Fed will justify afterward.

The curve flattened, which settles the question the lead poses. Term premium and fiscal supply build at the back end. The 30y moved least, the 2y moved most, and 10y-2y compressed to 0.39pp. That is the front end dragging the rest of the curve up, and the front end only moves that way when the market decides the Fed is stuck. With the effective funds rate at 3.63% and the two-year at 4.56%, this is not a bet on issuance. It is a bet that policy has further to go, or at minimum nowhere to go for a long time.

That bet is being placed on the wrong evidence. August CPI rose 1.318 index points while core rose 0.976. The gap is energy and what energy does to everything trucked. Brent sits at 106.079 after falling 4.25% in a single session, having run up 8.38% over five days. EIA has the US on track for record crude production this year, eight liquids pipelines completed since the start of 2025, LNG exports up 23% in the first half. The pressure at the pump is a crack spread problem, refining margin, not a barrel shortage and certainly not a wage spiral. Ten-year breakevens agree: 2.4%, up 0.03. The bond market raised real yields while telling you it does not expect more inflation. Those two things cannot both be a considered view.

Credit is not confirming either. High yield OAS at 2.7pp, tighter on the day. Nobody in the risk-taking part of the market thinks something broke. So the entire move is the risk-free curve pricing a central bank that will not cut into an oil print.

The dollar did not get the memo, which is the part worth holding onto. Broad trade-weighted index at 118.0732, barely moved. USD/JPY fell 0.60% on the day and 1.74% on the week, yen stronger against a rising US real yield. EM crosses drifted: rand weaker by 1.05% over five days, lira 0.33%. If this were a genuine global repricing of dollar rates, Jakarta and Pretoria would be paying for it already. They are not, yet, and that tells you the move is domestic and narrow.

Unemployment held at 4.1% with payrolls adding 162 thousand. A labor market that stable at a 3.63% funds rate is not evidence that policy is too loose. It is evidence that policy is roughly where it should be and that the disinflation came from supply, as it did last time, and will get credited to rates, as it was last time.

The two-year is wrong. It converges back toward the funds rate before the thirty-year gives back a basis point of its 5.37%, and the flattening unwinds from the front. If the Fed validates this pricing instead, the adjustment valve is the same one it always is, and the 4.1% will not stay there.

The term premium read is the comfortable one, and the curve does not support it. A supply story steepens. This flattened. The 2y moved 13bp, the 30y moved 9bp, and 10y-2y compressed to 0.39pp. If the marginal buyer were demanding more compensation for duration and deficits, the long end leads and the curve gets steeper. It did the opposite. The selloff was front-loaded, which makes it a policy path story wearing a fiscal costume.

Look at where the 2y actually sits. Effective funds held at 3.63%. The 2y printed 4.56%. That is the front end pricing a path nearly a full point above the current rate, not a path of cuts. Breakevens went to 2.4%, up three basis points, which is nothing. So the move was real yields, and real yields at the front end are a statement about the Fed, not about the Treasury's auction calendar.

The offside position is the cut trade. It was built patiently, on payrolls at 159,075k with a 162k monthly gain, unemployment flat at 4.1%, and a comfortable story about a Fed that had room. High yield OAS at 2.7pp and a basis point tighter on the day tells you this is not a growth scare or a risk event. Credit is not blinking. You do not get a 13bp front-end selloff with spreads grinding tighter unless the market is taking cuts out while remaining fully constructive on the cycle. That is the least forgiving combination for anyone long the front end for insurance reasons.

The part that does not map: USDJPY fell 0.60% on the day and 1.74% on the week while US rate differentials widened. EURJPY is down 1.87% over five days, GBPJPY down 1.68%. Cross-yen is leading, so this is yen strength rather than dollar weakness, and USDCHF up 0.80% over the week confirms no broad haven bid. Yen appreciating into a widening differential is the wrong sign. Either the carry position is being reduced for reasons unrelated to rate spreads, or something on the Japan side is repricing faster than the published calendar suggests. I cannot fit it to the differential, and I am not going to pretend it fits.

Crude sits in the same category of not-quite-explicable. Brent at 106.079 after a 4.25% drop that followed an 8.38% weekly gain, WTI at 100.354 on the same pattern. A crude draw of 391 MBBL against gasoline and distillate builds of 1,269 and 2,087 says runs are high and product is accumulating. That is a soft physical picture behind a spot price that spiked and gave a third of it back in a session. Risk premium put on and partially taken off, with the barrels themselves unbothered.

Position for the front end being right and the cut trade being wrong. The funding unwind in yen is not finished, and it is the trade that will hurt people who think their loss is in bonds.