The Curve's Span Fell to a Year Low of 79 Basis Points. The 1yr Took 18 of Them in One Session.
A yield curve has two jobs. It prices the policy rate for the next 2 years, and it prices the risk of lending for 30. When a political event arrives, the useful question is not whether rates rose. It is which of those two jobs did the repricing, because they reach completely different people. A move at the front end lands on anybody who borrows short: a treasury desk rolling 12-month paper, a company funding working capital, a bank pricing a loan off the 1yr. A move at the back end lands on anybody who owns duration. The span between the 1yr and the 30yr, the wedge, is the price of waiting, and it is the cleanest single number for which job the market just did. On 10 November the wedge fell to 79 basis points, the lowest of the year, and the 1yr did almost all of the work. The interesting part is what did not happen: the coupon calendar never changed, and neither did a single credit spread. This piece follows the repricing from the vote to a borrower's all-in cost, and names the level at which the story is wrong.
The whole repricing happened at 12 months
Start with the ladder, because the shape is the argument. Between the 5 November and 10 November par closes, the 1yr rose 28.0 basis points to 4.56%, the 2yr rose 27.0 to 4.73%, the 3yr 21.0, the 5yr 22.0, the 7yr 19.0, the 10yr 14.0, the 20yr 12.0 and the 30yr 10.0.1 The 1mo rose 2.0 and the 3mo 13.0. The peak of the move is not at the very front, where the policy rate sits, and not at the back. It is at 12 months, and it decays monotonically in both directions from there.
Read the 1mo as the control. It rose 2.0 basis points, which is noise, because a 4-week bill cannot express a view about anything except the next Federal Reserve meeting. The 3mo rose 13.0, the 6mo 19.0, and then the 1yr took 28.0. That progression is the signature of a repricing of the expected path of policy over the next 4 quarters rather than a change in the current rate. One session earlier, on 9 November, that whole 3-session move was mostly a single day: the 1yr rose 18.0, the 2yr 17.0, the 10yr 10.0 and the 30yr 9.0.
A 4-week bill cannot hold a policy view and a 30yr bond is not asked to. The instrument that sits exactly where the expected path is priced, and where the Treasury's own funding actually clears, is the 12-month point. That is why the number is 4.56% and not 4.98%.
The span of the curve fell to the low of the file
The 30yr closed at 5.35% and the 1yr at 4.56%, a wedge of 79 basis points. Across the 216 sessions from 2 January to 10 November the wedge has ranged from 79 to 142, with a median of 117 and a mean of 114.3. The 10 November print is the minimum of the series. It is the only session in the file below 80, and exactly 1 other session, 27 October, touched 80. The 25th percentile is 104, so the wedge is 25 basis points under the level beneath which a quarter of the year sat.
Two things have to be separated here. The wedge narrowing is not itself a signal about growth, inflation or credit. It is an arithmetic statement about where the repricing was concentrated: the numerator (the 30yr) moved 10 basis points and the denominator (the 1yr) moved 28, so the difference shrank by 9 after rounding, from 88 on 9 November to 79. The same session flattened the 2s10s to 25.0 basis points from 32.0, the second lowest of the 216 sessions, against a 47 median and a 20 minimum set on 21 September.2
Both spreads are saying one thing. The long end accepted almost none of the repricing. A curve that charges 79 basis points for 29 extra years of duration is not a curve where duration is scarce; it is a curve where the marginal buyer of long paper has stopped demanding compensation for the quantity of it. That is the same conclusion the 20yr-over-30yr inversion has been printing for weeks, reached from the other end of the curve.
The 1yr now sits where it sat for 3 sessions in the whole year
The level matters as much as the change. At 4.56% the 1yr par yield is higher than on 213 of the 216 closes in the file, a rank of 98.6 per cent. The year's high is 4.59% on 28 September; the low is 3.40% on 10 February. Three sessions in 2026 have printed at or above today, and 2 of them were in the last week of September.
That is the fact that makes this a midterm piece rather than a rates note. For 9 months the 1yr traded below 4.40%: the whole first half of the year sat near the low as the market priced a policy path that cuts. The 5 November close was 4.28%, which is where it had been for most of October. Three sessions later it is 4.56%, and the level is back at the top of the year's range, on no change to the calendar, no data release between 5 and 10 November beyond the CPI print on the 10th, and no change to the coupon sizes.3
The 1yr-minus-3mo spread makes the same point in the term structure's own language. It printed 26 basis points on 10 November against a median of 11 across the file, and only 22 of 216 sessions have printed at or above it. A 12-month bill paying 26 basis points over a 3-month bill is a market that wants to be paid for committing money past the next 2 meetings, and until this week it was paying 11.
Supply did not change, and that is the point
The obvious explanation for a 28 basis point move at the front end is more paper. The auction record does not support it. Gross coupon offering, notes plus bonds, has printed between 360bn and 371bn in each of the 9 complete months of 2026: 366bn in January, 361 in February, 362 in March, 371 in April, 371 in May, 367 in June, 366 in July, 360 in August and 362 in September. October shows 158bn, but that is 3 of the 5 planned coupon sales having settled by the 22nd, not a cut.
Bills carried 85.8% of gross settlement across January to September. In September the split was 86.2% bills against 362bn of coupons. The 90-day windows agree: 971.5bn of coupons settled in the 90 days to 12 November against 1077.6bn in the prior 90, and bills took 85.5% of gross against 86.8%. Coupon supply is 106.1bn lower year-on-year on the two windows, not higher.
So the quantity of duration being sold fell and the price of 12-month money rose 28 basis points. Those two facts cannot both be explained by supply. What is left is the other side of the ledger: the price at which the marginal holder is willing to take the paper, which is the term premium, and the composition of who is left buying.
The mechanism: the vote closed the long end and left the bill book open
Stated as a chain, with each link a series this desk holds. First, the 3 November result produced divided government, which closes the legislative fiscal channel: neither party can pass a budget resolution alone, so the deficit path stops being the thing that reprices. Second, the term premium responded exactly as that predicts. The ACM 10yr term premium printed 0.8738% on 6 November against 0.9354% on 22 October, minus 6.2 basis points, in the 78.0th percentile of its 2026 prints against a year range of 0.4627% to 1.0965%.4 The long end lost the thing it had been pricing.
Third, with the long end no longer repricing, the political fact has to reach a price somewhere, and the only place left is where the funding actually happens. Bills are 85.8% of gross settlement. The 1yr point is the longest bill in the regular cycle and the shortest coupon, which makes it the exact joint where the money-market book meets the coupon calendar. Fourth, the 1yr repriced 28 basis points and the wedge collapsed to its year low.
Fifth, and this is the part that reaches a balance sheet rather than a screen: the spread did not move with it. IG OAS printed 0.91% on 9 November against a 1yr range of 0.73% to 0.94%, which is 3 basis points off the widest print of the year. HY OAS printed 3.34% against a range of 2.60% to 3.48%, 14 basis points inside its own high. Neither series has a print after 9 November, so the credit leg of this is 3 days stale by construction and is dated as such throughout.
That combination is the whole finding. When the risk-free rate at 12 months rises 28 basis points and the spread over it does not move, the all-in cost of a 12-month obligation rises by 28 basis points and nobody's credit assessment changed. On the 10 November levels an IG borrower at the 12-month point is at 5.47% all-in, 0.91 of spread plus 4.56 of Treasury, against 5.19% on the 5 November 1yr. The same borrower at 30 years, if it could issue there at the index spread, is at 6.26%. Twenty-nine extra years of duration buys 79 basis points.
What confirms this, and what breaks it
The confirming state is the one the tape is already in. The wedge holds under 90 basis points, a level only 12 of 216 sessions have printed below, while IG OAS stays inside 1.00% and the 1yr stays above 4.40%. That is a front-loaded repricing with an unmoved credit complex, which is what a political event that closes the fiscal channel and touches nothing else should look like.
The falsifier is the wedge back at or above 104 basis points, the 25th percentile of the 216 sessions, inside 10 sessions. That is 25 basis points of re-widening from 79, and it can only happen one of 2 ways: the long end sells off, or the 1yr rallies back. Either one kills the claim that the repricing was concentrated at 12 months, because either one means the long end resumed pricing supply or the front end was a positioning move that unwound.
The second falsifier isolates attribution rather than direction. If IG OAS prints through 1.00% while the wedge stays under 85 basis points, credit is doing the work and this piece has the wrong actor. One point zero zero is 9 basis points wider than the 9 November print and 6 wider than the year's high of 0.94%, so it is a clean break rather than a drift, and it is checkable on every daily FRED print once the series catches up.
The third is arithmetic and it is the one that would make the whole framing irrelevant. If the long end rallies while the front end holds, the wedge widens without any fiscal news at all, and the level stops being informative about who priced what. The wedge is a difference; a difference can move because either term moves, and a piece that reads it as a single signal is only valid while the numerator is quiet. It was quiet for the 3 sessions covered here, by 10 basis points against 28.
Notes
1. Par yields: data/treasury_par.csv, US Treasury daily par yield curve, 216 sessions from 2 January to 10 November 2026. The 10 November close is the newest row; all levels and changes in this piece are par closes, not intraday marks. Session changes are close to close. The 3-session window is 5 November to 10 November; the 1-session window is 9 November to 10 November.
2. The 30yr-minus-1yr wedge is the 30yr par yield minus the 1yr par yield in basis points, computed on every close in the ledger. Minimum 79 on 10 November, maximum 142, median 117, mean 114.3. Sessions at or below 80: 10 November (79) and 27 October (80). The 2s10s is the 10yr minus the 2yr on the same basis; minimum 20 on 21 September, maximum 74, median 47. The 1yr-minus-3mo spread uses the 3mo column, not the 3-month bill auction high rate.
3. The CPI release on 10 November 2026 is the only scheduled data event inside the window, per data/events.json (kind cpi, 08:30 ET, confirmed). No Treasury refunding or coupon-size change falls in the window; the coupon sizes quoted are the settled auction amounts in tmp/all2026.json.
4. Term premium: FRED THREEFYTP10, the ACM 10yr term premium, 0.8738% on 6 November against 0.9354% on 22 October, minus 6.2 basis points. The series has no print after 6 November, so it is 6 days stale against the par curve and is used here only for direction, not for a same-day attribution.
5. Credit: FRED BAMLH0A0HYM2 (high-yield OAS) and BAMLC0A0CM (investment-grade OAS), both last printing 9 November 2026: 3.34% and 0.91%. The 1yr windows run from 9 November 2025. The all-in figures add the index OAS to the named Treasury par yield; this is a spread-plus-rate construction, not a matched-maturity yield, because the index OAS is measured on a basket whose average maturity is longer than 12 months. It is used to size the direction of a cost change, not to price a bond.
6. Issuance: tmp/all2026.json, 346 rows from the Treasury Fiscal Data auctions endpoint, gross offering amounts by security type and settlement month. October is partial, 22 of 31 days, and is excluded from every band and share quoted. The 90-day window figures are data/capital_demand.json, regenerated 12 November: coupons 971.5bn in the 90 days to 12 November against 1077.6bn prior.
7. Levels line: 2yr 4.73, 3yr 4.74, 5yr 4.81, 7yr 4.89, 10yr 4.98, 20yr 5.39, 30yr 5.35, 2s10s 25bp, all from the 10 November par close. Cross-checks in data/latest.json carry an 11 November mark for the constant-maturity series (UST10Y 4.998, UST30Y 5.356); those are a different series and are not mixed in.
The 1mo Bill Pays 15 Basis Points and None of It Is Policy — why the very front of the bill curve cannot carry a view
The Curve Peaks at 20 Years, and the Dealer Is the Buyer There — the coupon calendar this piece checks the supply claim against
The Stress Gauge Fell 0.652 in Four Sessions. The All-In Cost of Borrowing Rose 4.4 Basis Points. — the spread side, measured to 9 November, from the same snapshot