UST Monthly · October 2026 · Wed 30 Sep close → Fri 30 Oct close · 21 sessions
Month 10 · October 2026 · draft v0.2
October gave back 65% of September’s sell-off, and the Fed wasn’t the reason
A bull month: 10Y −35bp to 4.94%, 2Y −32bp to 4.56%. Real yields did 94% of the 10Y move as AI-financing worries pushed money into Treasuries. Bills moved only −4bp and the Fed held, so this was a haven month, not a Fed repricing. November tests it fast: refunding, payrolls and CPI land in the first eight days.
The read
ShapeThree days did 33 of the 35bp; the other 18 sessions netted −2bp.
VolLong-end vol near its 2026 high (10Y 90th, 30Y 89th pct); 2Y mid-range (50th).
Supply$341bn of coupons sold; dealers kept just 9%. Their long-end holdings still rose $12.4bn, bought in the secondary market.
Month change 30 Sep → 30 Oct; YTD from 2 Jan. Sparklines are October daily closes. Click a tile to open it in the Data lab.
▸
The month in four cards
The story first; open “Check the numbers” on any card, or jump to the chart that proves it
Read this first · each card links to its evidence
Card 1 · What happened
−35bp
10Y rally; the biggest monthly fall of 2026
The 10Y opened the month at its 2026 high (5.31% on 5 Oct) and closed at the month’s low, 4.94%. That undoes 65% of September’s +54bp; the 10Y is still +19bp above end-August.
Real yields were −33bp of the 10Y’s −35bp. AI-financing worries in tech and credit (15, 22 and 29 Oct) drove it, not soft data: PPI ran at 6.3% y/y and Q3 GDP beat. 22 Oct (−17bp) was the biggest one-day fall of 2026.
Check the numbers
10Y real 2.93% → 2.6% · 10Y breakeven −2bp · Treasury real yield curve
The Fed held at 3.75–4.00% on 28 Oct and Warsh signalled “wait and see”. Bills fell just 4bp and a December hike is still about 70% priced. The rally sat in 2Y–30Y, not in next-meeting odds.
Asset managers added 10Y futures longs; dealers ended up holding more long bonds
Asset managers’ net 10Y longs rose 224k contracts to 3.00m, the largest in the 25-week sample; leveraged-fund shorts grew 191k. That pairing is mostly the cash–futures basis trade, not a directional bet. Dealers’ >21y inventory rose $12.4bn.
Check the numbers
CFTC TFF 29 Sep → 27 Oct · NY Fed primary dealers 30 Sep → 21 Oct
Five lines, three paths, and what would change our mind
View in short
The rally is a haven trade that has not yet convinced the front end
What changed: a haven rally that closed at the month’s low (cards 1–2). → 01
Why: fear about AI financing, not softer data or a softer Fed; inflation pricing even rose at the front. → 03
What it means: the rally is exposed. A December hike is ~70% priced while the 2Y is 32bp lower. If tech and credit calm, the haven bid can leave faster than it came. → 02
How we’d hold it: no outright duration call into refunding and payrolls. The 2s10s steepener (W42-2) is at 38bp from a 31bp entry; target 40bp, stop 25bp. → 12
What would change our mind: 10Y back above its 20-day average (5.13%, ▲ 19bp higher) says the haven bid has gone; 2Y back above 4.88% (end-September) says the Fed is back in charge. → 11
Three November paths
Risk sentiment and the 10 Nov CPI decide which one
Haven holdstech/credit stress persists
AI-financing worries keep money in duration; payrolls soft. Curve: bull flattener from the long end.
→ W42-2 steepener stalls; 10Y towards end-August (4.75%).
Haven fadesrisk calms after midterms/refunding
Equities and credit recover; the real-yield rally reverses first. Curve: bear steepener.
→ Helps W42-2 if 2Y stays anchored; 10Y back to the 20-day average.
Hot data, Fed backCPI 10 Nov hot; Dec hike firms
Front end reprices towards the hike; 2Y back above end-September. Curve: bear flattener.
Toggle tenors, window and scale · hover for each day · click a day for the whole curve · drag to zoom in the Data lab
Daily par closes · lettered markers are the days that moved the month
The 10Y peaked at 5.31% on 5 Oct and closed the month at its low, 4.94%; the 2Y fell almost as far (−32bp)
Show
WindowScale
Why it matters2Y and 10Y fell together (−32 vs −35bp), so 2s10s ended only −3bp despite large daily swings. Sections 04–05 break the month into days and weeks.
Trade readDon’t extrapolate the trend. The 20-day average (5.13%) is still above the 50-day (5.00%), so the 2026 uptrend hasn’t formally turned.
WatchThe 20-day crossing below the 50-day (gap 12bp, from 19bp at end-September).
Event notes from Mortgage News Daily and CappNotes recaps. Moves are par close to par close. Shaded: October (wider windows).
Par curve (log maturity) · bars: change by tenor, bp
A bull move from 2Y out: 2Y −32, 5Y −38, 10Y −35, 30Y −33bp; bills only −4bp
Compare
Why it mattersThe belly (5Y) led, which is where Fed path and risk sentiment meet. Bills staying put means the market did not price a shorter hiking cycle for the next meeting.
Trade read2s10s ended −3bp at 38bp after swinging 29–51bp. Curve trades need a catalyst, not the month’s drift.
WatchRefunding sizes on 4 Nov for the 10Y–30Y sector.
Change in nominal, TIPS real and breakeven (nominal − real), bp
Real yields carried the 10Y (−33 of −35bp); the 5Y breakeven rose +10bp
Month
nominalrealbreakeven
Why it mattersFalling real yields with firm or rising breakevens is the signature of a haven bid, not a growth scare that kills inflation. Inflation risk is still being priced at the front.
Trade read5Y breakeven 2.46% vs 10Y 2.34% and 30Y 2.23%: an inverted breakeven curve says near-term inflation is the worry.
WatchOctober CPI on 10 Nov: a hot print hits the 5Y breakeven and the 2Y together.
September runs 1 Sep → 30 Sep (first real-curve observation in the file).
04
Three days did 33 of the 35bp
Waterfall from close to close · switch to every day sorted · hover for the driver
10Y par yield · contribution of each session
15 Oct (−9), 22 Oct (−17), 29 Oct (−7); the other 18 sessions netted −2bp
Why it mattersThe month wasn’t a steady grind lower. It was three shocks with noise in between, so the rally depends on those shocks not reversing.
Trade readTreat the 10Y as event-driven: size for jumps, not drift.
WatchA +10bp day with no news would say the haven bid is leaving.
05
October, week by week
Heat table, Friday levels or the daily regime strip · click a week to drill in
Weekly change in par yields, bp (Friday to Friday; first row 30 Sep → 2 Oct)
The mid-month weeks did the damage (16 Oct week: 10Y −15bp); the last week turned into a steepener
Why it mattersThe month changed character: a long-end-led rally mid-month, then a front-end-led one after the Fed. Day by day it was choppy: 6 bull-steepener, 6 bull-flattener, 6 bear-flattener and 3 bear-steepener sessions.
Trade readThe final week’s 2Y −13bp vs 30Y −2bp is the steepener story; it needs payrolls and CPI to cooperate.
WatchWhether the 2Y keeps leading if the haven bid fades.
06
The year so far
Month-end change or monthly candles · switch series · click a month bar to open it in the Data lab
Par yields by month, 2026
October −35bp is the largest monthly 10Y fall in the 2026 data; it opened at 5.29%, hit 5.31% and closed at its low
Why it mattersA close at the month’s low after a year high early on is a full reversal candle. Even so, the 10Y is still +75bp on the year.
Trade readSize positions for this volatility: 3 of the last four months moved the 10Y by 30bp or more.
WatchNovember holding below October’s open (5.29%) keeps the reversal intact.
Month-end par closes. Bars start in February: par.csv starts 2 Jan. Candles: open = prior month-end close (January: 2 Jan).
07
Volatility map
Realised vol per tenor against its 2026 range, or its path through the year
20-day realised vol, bp/day (std dev of daily changes)
The long end is near its 2026 high (10Y 90th, 20Y 95th, 30Y 89th percentile); the 2Y is mid-range (50th)
Why it mattersVol is concentrated where the haven trade happened. The front end is calm because the Fed was priced and didn’t move.
Trade readLong-end positions need wider stops than the front end right now.
WatchFront-end vol picking up around payrolls and CPI would mean the Fed is back in play.
Percentile = share of 2026 sessions with lower 20-day vol. Dashed tick = end-September.
Pick a contract or click a row · own-scale lines, weekly flow, share of open interest or the basis compass
CFTC TFF, futures only · leveraged funds vs asset managers · weekly Tuesdays
The basis trade grew in the 2Y and 10Y; asset managers’ 10Y net long rose 224k to 3.00m
ContractView
Why it mattersAsset-manager longs against leveraged-fund shorts are mostly the cash–futures basis trade: plumbing, not a view on rates. The new information is size, which is unwind risk if funding tightens.
Trade readDon’t read the asset-manager build as conviction buying; it is mostly the long leg of the basis.
WatchRefunding (4 Nov) for coupon sizes, and any repo stress around mid-month settlement. A 10Y move into the compass’s top-left quadrant would flag an unwind.
Positions 29 Sep → 27 Oct (CFTC Tuesdays). Contracts are not added across tenors.
09
Where October’s supply went
Flow of each auction to its buyers · by auction, by sector, or both
Coupon & TIPS auctions, October 2026 · ≈ $bn awarded by buyer group
Investors took 91% of $341bn; dealers kept 9%, and only 2.5% of the 10Y and 6.8% of the 30Y
Why it mattersNo October auction showed stress (the five with published tails were 0.0–0.9bp). Indirect bidders took 60% overall and 80% of the 10Y: end demand for duration was firm.
Trade readWeak auctions aren’t a risk to price in yet.
WatchThe 4 Nov refunding sizes and the 10Y/30Y auctions that follow. A long-end dealer share above about 15% would be the first crack.
≈ offering × competitive share by bidder class. 3Y, 10Y and 30Y from FiscalData accepted amounts; others from Helious result pages. SOMA add-ons and non-competitive bids excluded. Indirect bidders include foreign official and private accounts and some domestic funds.
10
Who ended up holding it
Dealer reshuffle flow, dealers by week, change by maturity, or foreign official custody
NY Fed primary dealers 30 Sep → 21 Oct · Fed H.4.1 custody to 28 Oct
Dealers cut $28.1bn of bills and short coupons and added $24.7bn further out, led by >21y (+$12.4bn)
Why it mattersDealers kept only about $3.8bn at the 10Y, 20Y and 30Y auctions, yet their >21y holdings rose $12.4bn (+$12.2bn in the week to 14 Oct). The build came from the secondary market: clients sold long bonds to dealers into the rally, or dealers chose to hold more. These data can’t say which.
Trade readNot a weak-demand signal, since auctions were strong. It is extra long-end inventory that dealers will want to distribute before refunding supply lands. Foreign official custody −13.8bn: no sign official buyers drove the rally.
WatchDealer >21y holdings in the next two NY Fed releases: falling back means clients took the bonds; rising into refunding is a pressure point.
11
November: calendar and levels
Click an event to see the levels it threatens
What’s coming (ET dates) · level ladders
Refunding, payrolls and CPI land within eight days; a December hike is still mostly priced
Date
Event
Threatens
Mon 2 Nov
ISM manufacturing (Oct); Treasury borrowing estimates
Supply
Tue 3 Nov
US midterm elections; JOLTS (Sep)
Term premium
Wed 4 Nov
Quarterly refunding statement (coupon sizes); ISM services
Long end
Fri 6 Nov
October payrolls (consensus +55–65k, unemployment 4.2%)
W42-2 · 2Y
Tue 10 Nov
October CPI
W42-2 · breakevens
8–9 Dec
FOMC (hike ~70% priced on 2 Nov)
Front end
Ladder
Why it mattersThis is where the haven rally meets the Fed.
Trade readLet the first week decide. Fade nothing before refunding sizes are known.
WatchPayrolls (6 Nov) and CPI (10 Nov) move the 2Y first; switch the ladder to 2Y to see how far it is from each level.
12
Setups: trend since the signal
Each card charts its setup from the day it was issued to month-end · hover for values
Setups open at month-end · signal shaded from 16 Oct (W42 edition)
The steepener is +7bp in the money; the fly has drifted towards its stop; the trend flip hasn’t fired
Risk: hot payrolls or CPI; hawkish Fed talk. S = issued, T = triggered, F = Fed day.
W42-1LIVE
5s10s30s fly, receive 10Y
On the month-end fit: 0.99bp at the signal → 1.93bp now
Target 0bp · stop 3.2bp (▲ 1.27bp away)
Risk: a 10Y-only sell-off if the haven bid unwinds. Weights 0.57×5Y and 0.47×30Y on this fit; the weekly’s own refit governs the stop.
W42-3MONITOR
10Y trend flip
20-day 5.13% vs 50-day 5.00% · gap 12bp
Not fired · fires on a cross below
Signal: the 20-day average crosses below the 50-day.
Why it mattersThe steepener has made 7bp of its 9bp target in six sessions, mostly on the Fed day. The fly moved away from target as the 10Y lagged the wings.
Trade readHold the steepener with its 25bp stop; the fly needs the haven bid to fade before it works. No new setup at month-end.
WatchPayrolls (6 Nov) and CPI (10 Nov) for the steepener; the next weekly PCA refit for the fly.
Fly = PCA-neutral 10Y residual (2 factors, 120 days to 30 Oct), so earlier values are restated on today’s fit. Education only, not investment advice.
13
Data lab
Build your own chart: up to four series, any window, level or change, overlays and statistics · drag to zoom · click for the whole curve
Every daily close in 2026 · Treasury par and real curves
Pick series and a window; the table recomputes
Series
Window
ShowOverlays
Overlays apply to a single series in level view. Daily vol = standard deviation of daily changes in the window. Beta is the regression slope of the second series’ daily changes on the first’s.
UST Monthly by Jasmeet Gujral. For education only; not investment advice. Intraday levels in news reports differ from the Treasury par closes used here. Positioning figures carry their as-of dates.