The front end tells the cleaner story. The UST 2Y NOTE net short shrank by 82,028 contracts on the week, and the UST 5Y NOTE by 23,512, both flagged as short-covering rather than a stalled short. SOFR-3M futures, the cleanest read on near-term policy rate expectations, saw its net short shrink by 122,152 contracts, its own flow tagged short-covering. All three sit at a positioning extreme by the desk's own long-window measure. The UST 5Y NOTE net short has not been more stretched since 7 July 2026. SOFR-3M's z-score against three years of history sits at negative 2.43, one of the most lopsided books on the board.
Set that against the UST 10Y NOTE, where the net short grew by 75,016 contracts on the same report. That extends a position that has not been more stretched since 13 January 2026. The UST 10Y NOTE sits at the 93rd percentile of its own one-year positioning range even as the front end's covering stretches sit closer to their least-crowded points of the year. Two cohorts trading the same asset class, moving in opposite directions in the same week.
The data offers a plausible reason for the split. Canadian CPI landed soft on 20 July 2026: the month-on-month print came in at negative 0.4% against a forecast of negative 0.2%, a sharp deceleration from the prior month's 1.0%. The year-on-year read missed too, at 2.8% versus a 2.9% forecast. The median CPI miss, at 1.9% against a 2.1% forecast, marked the most forecast-relevant undershoot, flagged as disinflationary. The front-end futures cohort's covering fits a market pricing central banks closer to done hiking, or closer to cutting, than the long end will concede.
The front end is covering into softer inflation data as the long end keeps shorting into a fiscal-supply story. That gap is the market's real disagreement on 20 July 2026, not the headline rate path.
The long end's resistance has its own grounding. The fiscal-pricing curve still shows a firmer path: the market prices the policy rate 36.3 basis points higher in twelve months than the current implied front rate of 3.627%, a delta the desk's own pricing data explicitly flags as a firmer path, not cuts. The UST 10Y NOTE yield sits at 4.57%, in the 98th percentile of its trailing year, with a z-score of 1.93. A short this stretched, last seen in January, is a short built on the view that long-duration supply, not near-term disinflation, sets the price at the long end. The 2s10s curve, at 0.37, sits in just the 6.7th percentile of its own year. That steepening pressure is itself unusual.
This reading has a real gap. COT data lags a week, and the UST 10Y NOTE short could simply be catching up to a fiscal story that has been running longer than the front end's disinflation trade. There is also no way to separate genuine long-duration hedging from outright directional conviction inside a single net-short number; the cohort mixes both. The upcoming US Bond 19-Year 10-Month auction (reopening) on 22 July and the US Note 10-Year auction on 23 July will test whether demand at the long end matches the short's implied skepticism, or whether the auctions go smoothly and leave that short looking early rather than right.
The falsifier is specific: if the next COT report shows the UST 10Y NOTE short beginning to cover alongside the front end, rather than extending further, the curve-split reading fails. The more likely explanation then becomes a single, delayed repricing working its way down the curve, not two genuinely separate views of the data.
