The 1mo Bill Pays 15 Basis Points and None of It Is Policy
A Treasury bill is an IOU with a maturity date printed on it. The shortest one, the 1-month bill, pays out before the Federal Reserve next meets — which makes it the cleanest instrument on the whole curve. It cannot contain an opinion about what the Fed will do, because the Fed will not have done anything by the time the bill is repaid.
So when the 1mo bill yields 15 basis points more than the rate the Fed pays banks on reserves (IORB, or the interest on reserve balances — the administered rate that anchors overnight money), that premium is paying for something else: the plumbing cost of holding government paper across a month-end, when the Treasury rebuilds its cash account and dealers shrink their balance sheets. That cost is where the midterm result gets invoiced, because divided government means no fiscal consolidation and no large new programme — and no coupon increase either. The funding keeps running through the bill line.
The 3mo bill, by contrast, spans the 9 December meeting. It pays +31, which is the same +15 plus 16 basis points of policy. Two numbers that look like one chart, and the whole argument lives in the difference between them.
The instrument that cannot hold a policy view
On 2 November the 1mo bill closed at 4.05%. IORB has been 3.90% since 17 September and does not move again until the 9 December FOMC. A bill bought this week matures inside that window, so the 15 basis point gap cannot be a forecast of policy — there is no policy decision left between purchase and repayment.
That leaves plumbing. The premium compensates the holder for three things that have nothing to do with the economy: settlement dates that cluster at month-end, dealer balance sheet capacity that shrinks on reporting dates, and a Treasury cash account that swings by tens of billions as it rebuilds and releases. Anyone who lends the government money across those dates wants paying for the risk of getting the timing wrong.
The premium has been positive for 32 consecutive sessions, since the current rate setting. The median over those 32 sessions is +16; the range is +6 on 09/21 to +27 on 10/15. Today's +15 sits below the median — the premium has compressed, not built, into the vote.
The ladder is +15, then +31, and the gap is a meeting
The bill curve on 2 November, measured against IORB: 1mo +15, 2mo +26, 3mo +31, 6mo +35, 1yr +48. It is not a flat line and it is not a straight slope. It rises like a staircase, and the first step is the only one made of something other than policy.
From 1mo to 3mo the spread gains 16 basis points. That gain spans the December meeting, so it is partly a rate expectation and cannot be read as fiscal. Everything at the 1mo is fiscal-technical; everything at the 3mo is that plus one meeting. Treating a curve point as a single number is how a technical premium gets mistaken for a policy signal.
The long end tells a different story on the same day. The 2s10s slope is 32.0 basis points and the 1mo-to-1yr slope is 33. The front end is where the term structure is steep; the belly is where it is nearly flat. That shape is characteristic of funding that is being pushed to the shortest possible maturity rather than sold to duration buyers.
Bills are 87.8% of gross, and that is the whole mechanism
Over the last 90 days the Treasury sold $7,931.8bn gross, of which bills were 87.8% and coupons $971.5bn. In the prior 90-day window bills were 84.5% and coupons $1,232.3bn — coupon supply fell $260.8bn while the total barely moved.
This is the mechanism that connects a political outcome to a price. A bill can be upsized at any auction with a few days notice and no refunding announcement. A 10yr or 30yr cannot: it requires a quarterly refunding, a size decision, and a duration buyer willing to take the paper. So when the fiscal path requires more money and the political system will not sanction a coupon increase, the marginal dollar is raised at the bill line — and the price of that decision appears as a spread over the administered rate, not as a term premium at 30yr.
The 4 November refunding is the test. It is the one scheduled event that can rotate the mix. If coupon sizes rise, some of the bill premium should drain; if they hold, the bill line keeps carrying the funding and the front end keeps invoicing it.
The parking lot warns, but read it carefully
Reverse repo take-up (RRP — the facility where money market funds park cash overnight at the Fed) printed $5.087bn on 2 November, up from $4.375bn on the month-end itself. On its face that is a large move and it has been read as reserves getting scarce.
The month-end record argues for caution. Across the prior 6 month-ends the facility spiked and then collapsed the next session: $11.539bn on 2026-09-30 to $0.350bn, $6.726bn on 2026-08-31 to $0.725bn. A month-end spike is a calendar event, not a scarcity signal.
What is unusual is the persistence. The 2 November print is the only one in the window that did not collapse the next session — $4.375bn to $5.087bn is a rise. 1 session is not a trend, and reading a single print as proof of scarcity is exactly the error the month-end pattern punishes. It is a thing to watch, not a thing to trade.
The mechanism, the instrument, the falsifier
Mechanism: divided government removes both fiscal consolidation and a large new programme, so the deficit is financed at the margin by bills, which need no refunding and no duration buyer. That keeps a structural bid for compensation at the shortest maturity, where the only risk being priced is calendar and balance-sheet plumbing.
Instrument: the 1mo bill minus IORB, at +15 basis points on 2 November (4.05% against 3.90%). Positive for 32 sessions, median +16. The 3mo spread at +31 is the same instrument contaminated by one meeting and should not be quoted as the fiscal read.
Falsifier: the 1mo spread closing at or below +6, the low of 09/21. That would say the calendar premium has been arbitraged away and the front end is no longer charging for the funding mix. The second falsifier is a 4 November refunding that lifts coupon share decisively above 84.5% — that would move the fiscal risk to duration and this read would be looking in the wrong place.
Where this is wrong
The cleanest way this piece is wrong is that the 15 basis points is not fiscal at all but ordinary quarter-end and month-end intermediation that would exist under any government. The defence is the persistence: 32 consecutive sessions, not a 2-day spike around a reporting date. If the spread had been positive only at month-ends the objection would win. It has not been.
The second way is a data error. IORB moved from 3.65% to 3.90% on 17 September, so any percentile computed across that date mixes 2 different policy regimes and is meaningless. Every spread in this piece is computed only inside the constant-rate window, 32 sessions from 09/17 to 11/02. Against the full 130-session file the 3mo spread reads as an 85th percentile; against the constant-rate window it reads closer to the median. The first number is arithmetically true and analytically wrong.
The third is that a bill premium is not a forecast. It says what the front end charges today for the funding mix that exists today. It does not say the mix persists, and it does not price the outcome of any specific race.
Notes
Bill and coupon yields from the Treasury daily par curve, 11/02/2026 close. Administered rate from FRED IORB, 3.90% since 17 September 2026. Spreads are computed inside the constant-rate window only (32 sessions). Issuance shares from the Treasury Fiscal Data auctions_query feed, gross settlements over rolling 90-day windows ending 2 November. RRP take-up from FRED RRPONTSYD; month-end peaks identified as the last business day of each month.
Charts are built by a generator from these same source files, not typed by hand. Every figure quoted in the prose is computed in the same script that drew the chart beside it.
The Auction Costs Most at the Front End. It Costs Again at 20 Years. — the concession ladder, which found the front end paying most at auction — the same front-end cost seen here as a standing spread rather than an auction-day event
The Vote Cannot Move the Breakeven. It Moves the Real Yield. — where political risk has been paid this year; this piece locates a second channel at the shortest maturity