The Long End Has Already Voted. It Voted for Gridlock.

The 30yr-10yr spread closed at 33.0bp on 19 October, from 67.0bp on 2 January and the 2nd percentile of 200 sessions. The 10yr term premium printed 0.9857% on 16 October — the 91st percentile of 2026, 11.1bp off its October peak. The generic ballot is D+7.9 and the House is priced at 92% for a flip. The fiscal premium an election would add is already absent; the 11 December funding deadline, 48 hours after the last FOMC of the year, is not priced at all.

There are two ways to hold a view on an election. You can forecast the result, or you can read what the market has already decided the result means. Only the second one is priced. The 30yr Treasury yield is the rate the government pays to borrow for thirty years; the 10yr is the same for ten. The gap between them, the 30s10s spread, is the part of the curve that prices long-run borrowing rather than the next policy decision — fiscal risk, in one number. On 19 October that gap closed at 33.0bp, against 67.0bp on 2 January. It has closed lower on only 4 sessions out of 200 this year. Meanwhile the 10yr term premium — the extra yield investors demand for committing money for a decade instead of rolling 3-month bills — sat at 0.9857%, the 91st percentile of its own year. The level still charges for duration. The slope has stopped charging for the far end. Those two facts are not a contradiction. They are a buyer base, and this is what it says about 3 November. Update, 26 October: the spread has narrowed further since this was written, to 29.8bp on 23 October off the on-the-run marks, which strengthens the case rather than softening it. The level has not moved much; the slope has.

The flattening was a front-end event, and the front end was policy

The curve did not flatten this year because the long end rallied. It flattened because the front end repriced. From 2 January to 19 October, the 2yr rose 130bp to 4.77% and the 3-month bill rose 62bp to 4.27%. The 30yr rose 62bp to 5.48%. The 10yr rose 96bp to 5.15%.

Work the arithmetic once, because everything after this depends on it. In proportionate terms the 2yr rose 37.5% and the 30yr 12.8%. The 2s10s spread went from 72bp to 38bp, and the 30s10s from 67bp to 33bp. A curve that flattens while the policy-sensitive end rises 130bp is a curve being dragged by the Fed, not repriced by the fiscal authority.

The long end is where an election would show up. A government about to borrow more pays more for thirty years than for ten, because the buyer of thirty-year paper is the one who has to hold through the extra issuance. That premium went the other way. On 16 October the 10yr term premium printed 0.9857%, down 11.1bp from its 1.0965% peak on 2026-10-05. It is still the 91st percentile of 2026 — the level is high — but the direction of travel over the five weeks that contained the final stretch of the campaign was down.

Why the level and the slope disagree: one buyer, two jobs

A term premium at the 91st percentile and a 30s10s at the 2nd are describing two different buyers. The pension fund and the insurer buy 30yr paper because their liabilities are thirty years out; they are not choosing between the 30yr and the 10yr, they are choosing between the 30yr and not being matched. That bid sets the level.

The slope is set by whoever buys the 10yr against the 30yr — the duration-constrained, the benchmarked, the relative-value account. That buyer has been leaving since the summer. The 20yr sits 5bp over the 30yr on the 19 October close (5.53% against 5.48%), and it has sat above the 30yr on 73 of 200 sessions this year. The market charges more for twenty years than for thirty. That is not a duration signal. It is the point on the curve where the cash-like bid is too short and the liability bid is too long, and it is the least liquid coupon printed.

Read together: the marginal buyer of duration is a liability matcher who does not care about the election, and the marginal seller of the belly is a relative-value account that does. An election that changes the fiscal path would have to change the first group's mind. Nothing in the last five weeks suggests it did.

The three channels the vote can actually reach

One: the fiscal path. The mechanism is net coupon supply. A unified government writes a budget resolution with new spending and finances it at the long end, which lifts the term premium and steepens 30s10s. A divided one produces continuing resolutions at existing levels, which does neither. The polls have the generic ballot at D+7.9 as of 21 October (Decision Desk HQ average), a 92% price on a Democratic House (Kalshi) and 68% on the Senate. Against a Republican presidency, that is the divided outcome — the one that adds no coupon supply. So the fiscal channel is priced as a non-event, which is exactly what a 33bp 30s10s looks like.

Two: the funding deadline. The continuing resolution that funds the government expires on 11 December, five weeks after the vote and inside the lame duck that opens 9 November. This is the channel a divided result makes more likely, not less: once the electoral cost is sunk, both chambers have an incentive to fight. The mechanism is mechanical and was measurable this week — a funding fight rebuilds the Treasury general account, and a general-account rebuild drains reserve balances, because the cash does not leave the system, it changes hands from the banking sector to the sovereign. On 21 October the account held $945.2bn and reserves $2,962.6bn, with $0.35bn parked at the reverse repo facility. There is no parking lot left to absorb a call of that size.

Three: the sequencing. The FOMC meets 28 October, six days before the vote, and again 9 December — two days before the continuing resolution expires. The Federal Reserve does not change hands on 3 November; the chair's term runs to 2030. But the institution makes two decisions that straddle the political event, and the second one lands inside the funding window. Whatever the vote does to the fiscal path, it cannot change the fact that the last rate decision of the year is a funding decision by accident of calendar.

What the front end is pricing, and it is not a squeeze

The bill complex is expensive in absolute terms — the 1-month at 4.11% is the 97th percentile of its 2026 range, the 3-month at 4.27% the 98th, the 6-month at 4.37% the 98th. Expensive bills are usually read as collateral scarcity. This time the shape says something else.

Measured against SOFR at 3.88%, the bill curve is monotonic: +23bp at 1 month, +29 at 2, +39 at 3, +49 at 6, +62 at a year. A collateral squeeze inverts the front of the bill curve, because the cash that needs a home bids the shortest paper. A hiking path slopes it the other way, because every tenor has to price the policy rate expected over its own life. This is the second shape. The front end is not pricing a funding problem. It is pricing roughly two more hikes.

Which is the piece's real point about the election. The market has a view on the policy path and a view on the fiscal path, and both of them are already consistent with a divided Congress and an unchanged term premium. The event with no price on it is the one that lands 48 hours after the last rate decision.

Where this breaks

First, the slope. A 30s10s back above 45.0bp — its level on 169 of 200 sessions this year, the last of them 9 September — with the 10yr term premium back above 1.00% would mean the long end is charging the vote as a fiscal event rather than discounting it as gridlock. The compression would then be a duration rally that is about to unwind, not a political repricing.

Second the funding leg: SOFR at or above the 3.90% administered rate on the first print after 11 December would mean the funding window was priced as scarcity after all, and the bill curve's monotonic slope was a hiking path that got run over.

Third, the composition. A single party holding both chambers — the markets price Republicans at roughly 8% to keep the House and 32% to keep the Senate, so a Republican sweep is a few per cent, not a coin flip — with 30s10s still under 40bp a week after the vote would falsify the supply mechanism outright: it would mean new issuance does not steepen the far end in this regime at all, and every fiscal-risk read on this curve is mismeasured.