The Bank of England Decided Yesterday. The Only Percentage Anyone Published Is Four Weeks Old.

Two open items from this desk, both closed at the issuer

Seven hours ago this desk published, in its own humility section, that “no page we read lists a Bank of England Monetary Policy Report alongside today’s decision.” That is now closed, and it is closed at the Bank rather than at a better aggregator. The Bank’s own Monetary Policy Report page for November 2026 says “This Monetary Policy Report will be published on 5 November 2026,” and its upcoming-MPC-dates table carries one row for Thursday 5 November listing the MPC Summary and minutes “and November Monetary Policy Report.” Two Bank pages, same answer. A Tokyo schedule independently lists the policy rate, the minutes and the policy report in the same 21:00 Japan-time slot, which is 12:00 UTC.

That matters more than a documentation tick. A rate decision with a quarterly report attached is not one release, it is a decision plus a forecast round, and the forecast round is where a committee that has just hiked tells you whether it intends to do it again. If your schedule has a single window on 12:00 for a 25-basis-point number, it is sized for the smaller half of what lands.

The second open item was tomorrow’s US employment report. This desk wrote this morning that “no page we read gives a time for tomorrow’s US employment report” and that our 13:30 was a fact carried from 2 November rather than re-read. The issuer settles it: the Bureau of Labor Statistics release schedule prints “October 2026 | Nov. 06, 2026 | 08:30 AM” with no time-zone column, and the September release states in prose that the October report “is scheduled to be published on Friday, November 6, 2026, at 8:30 a.m. (ET).” The carried fact was right. It is now a read fact.

Worth noting what that cost: the schedule page itself has no time zone, and the only place the zone appears is in a sentence inside last month’s release. Two pages at the same issuer, one of which is usable for a news window and one of which is not.

The Bank prints “12pm” and never says which noon

Here is the whole set of readings we have for the announcement time. The Bank’s Monetary Policy Report page gives it in its own title: “Monetary Policy Report - November 2026 (to be published at 12pm).” The minutes page does the same. No time zone appears beside it on either page. The Bank’s upcoming-MPC-dates table, which is the page you would build a calendar from, prints no time at all for any date in 2026 or 2027, and its footer says it was last updated 21 September 2026.

Every secondary reader is more specific than the issuer. A vendor calendar page, verified 1 November, says “due at 12:00 UTC.” A UK currency broker says “12:00 UK time.” Two separate Japanese schedules put it at 21:00 Japan time. All four are the same instant, and all four agree with each other.

They agree today. “12:00 UK time” and “12:00 UTC” are the same instant for roughly five months of the year and an hour apart for the other seven, and this is the first MPC announcement to fall on the UTC side of that boundary. A news window seeded from a British page during the summer and never re-derived opens an hour early; a window seeded from the Bank’s own authoritative calendar page does not open at all, because that page carries no time. The institution with the most authority over this timestamp is the one publishing the least usable version of it, and the readers you would be told to distrust are the ones carrying the zone.

The percentage is four weeks old and the vote has no percentage at all

We went looking for a current market-implied probability of today’s hike and did not find one. What exists:

• 81 per cent that the Bank raises on 5 November, attributed to LSEG, in an article published 9 October. Four weeks old. The same piece reports money markets implying more than an 80 per cent chance a day earlier, and two quarter-point increases fully priced by February.
• 54 of 63 economists expecting 4.00 per cent today, in a Reuters poll conducted 12–14 October. Three weeks old, and a survey of forecasters rather than a market price.
• “A done deal and fully priced in by financial markets,” attributed to Alexander Harvey of Oxford Economics in a page updated 4 November. No number, and the quote itself carries no date.
• “Considered near-certain” in a Tokyo wire at 02:29 Japan time this morning. No number.

So the freshest characterisation is qualitative and the freshest number is four weeks stale. That is a worse position than it sounds, because the two sentences that would have moved the number both landed inside those four weeks: a US election and an 8 per cent collapse in crude, which is the single largest input to the UK inflation profile the Bank is about to publish a forecast for.

And then there is the thing with no percentage at all anywhere. Two days ago this desk published that the Reserve Bank of Australia hiked exactly as four probability feeds said it would, that three of its nine members voted against, and that “no probability feed we know of quotes a vote margin, so every reader who was correctly positioned for the hike was positioned for the wrong object.” Today is the second instance in three days, and this one is arithmetically sharper, because the Bank of England publishes its split in named terms and the starting position is on the record.

The vote arithmetic, from the Bank’s own minutes

September’s minutes, verbatim: the MPC “voted by a majority of 6–3 to maintain Bank Rate at 3.75%. Three members voted to increase Bank Rate by 0.25 percentage points, to 4%.” The minutes name both sides. Andrew Bailey, Sarah Breeden, Swati Dhingra, Clare Lombardelli, Dave Ramsden and Alan Taylor preferred to hold. Megan Greene, Catherine L Mann and Huw Pill preferred the increase.

Three votes for 4.00 per cent are therefore already banked, and a hike needs five. Two of those six named holders have to switch. A UK broker page, stale on everything else, states that arithmetic plainly: two more members would need to move to produce a 5–4 majority.

That is why “fully priced” and “no information” are not the same claim today. If the market is right about the level, the distribution of outcomes at 12:00 runs from 5–4 to 9–0, and those two are different events with the same headline. A 5–4 hike delivered with a Monetary Policy Report is a committee that has been dragged over a line and will say so; an 8–1 or 9–0 hike is a committee that has moved as a body and has a second one in it. Sterling does not care which number the headline says. It cares about the second one, which no feed quotes, which no page we read forecasts, and which arrives in the same minute as the forecast round.

The dovish case is quantified, single-sourced, and lives in the report rather than the rate

One Tokyo wire this morning does something none of the English pages we read attempted: it states a numerical reason the Bank might push back on its own hike. On its account the Bank has adopted a new inflation indicator that shows almost no sign of building price pressure, and on the official CPI-based analysis September’s sticky central inflation was flat at 3.3 per cent, a level it has held for most of the period since the Iran war began. The same piece says markets price four more increases including today’s, and that some readers expect the Bank to lean against that path.

We are labelling this as one desk’s view, because it is: we did not reach the Bank’s own measure, and we have not second-sourced the 3.3 per cent. But the shape of the argument is worth holding whether or not the figure survives. A committee that hikes on a 5–4 vote while its preferred underlying measure sits flat has every reason to spend the Monetary Policy Report explaining why the next four are not coming, and that explanation is a sterling event with no consensus field against it.

The one market reading we have going in is sterling strength, not weakness. One Tokyo wire at 02:29 Japan time has cable touching the 1.35 handle overnight, through a 200-day moving average it puts at 1.3425, with sterling–yen at about 213.75 and also above its own 200-day; the same page’s footer quotes 1.3473, 213.25 and 0.8340 against the euro, untimed. Cable above a line it only just cleared, into a decision the market says is certain and the committee may not, is the setup where a correctly forecast hike sells sterling.

Gilts rallied into it, and two publishers disagree about how much

The UK ten-year fell further yesterday than any other major sovereign we have a reading for. On a Tokyo publisher’s close table the gilt yield is 5.129 per cent, down 0.132 — against the German ten-year down 0.058 to 3.272, the US ten-year down 0.0742 to 4.8879, the US two-year down 0.0500 to 4.5809, and the Japanese ten-year up 0.040 to 2.821.

A second reader does not agree. Trading Economics gives the UK ten-year at 5.1368, prints it as 5.14, and puts the day’s move at about 0.11 percentage points, with its own table carrying a previous close of 5.25. So the close is roughly a basis point apart across the two and the move is two basis points apart, which backs out to the two publishers disagreeing about yesterday’s starting level rather than about yesterday’s finish. A third reader is worse than either: Investing.com’s gilt page presents a live-looking quote block at 5.3484 stamped 26 October, nine sessions back, with a day’s range and a previous close attached to it.

Take the agreement rather than either number: the gilt rallied something between 11 and 14 basis points the day before an expected hike, and it was the biggest move on the board. That direction is the market pricing the report, not the rate — and it is also the direction that gets hurt by a 9–0. The Trading Economics page calls the level the lowest in two months, which is its own editorial framing and not a figure we are scoring.

What this does not tell you

It does not tell you the vote, and nothing we read even forecasts it. Of the seven pages we opened on this decision, not one carries a November vote-split expectation. The only split forecast in the file is a mortgage-industry commentator predicting 5–4 for the September meeting, in a live blog from 17 September, and September came in 6–3 the other way. We are not reusing it.

It does not confirm a press conference. Four Bank of England pages mention no press conference and name no speaker. The 12:30 Bailey appearance in our schedule comes from a Japanese calendar and nothing else we read corroborates it. The Bank’s own speeches index returned zero results to us, which is a rendering limitation rather than evidence of absence. If your handling depends on a second window thirty minutes after the first, you are currently taking that on one page’s authority.

It does not give you a current UK inflation rate we are confident in. The Bank’s monetary-policy landing page, last updated 21 October, prints “Current inflation rate 3.5%” against a 2 per cent target. Two secondary pages give August CPI as 3.1 per cent, up from 2.9 in July. Those are different months, not a contradiction — but we did not read the Office for National Statistics, so we cannot tell you which month 3.5 is, and a committee’s own September projection of around 3.75 per cent in the fourth quarter, which one page attributes to the MPC without naming anyone, sits above all of them.

It does not give you a Tokyo level for today. We could not read a single publisher giving dollar–yen’s high, low or latest level through this morning’s Tokyo session. Three Japanese market-summary pages that would carry it returned client errors, and the indexes we could reach were captured before the session opened. What we have instead is yesterday’s New York figures — a high of 158.67, a low of 158.16 and a close of 158.46 on a Tokyo publisher’s account — and the headlines of this morning’s Tokyo reports, which say the dollar softened on lower US yields and lower crude without printing a number. Two publishers’ headlines agreeing on a direction is not a level.

And it does not tell you the schedule is complete. The one Tokyo schedule page we read in full lists, for today: a Reserve Bank of Australia deputy governor at 06:45, a European Central Bank Executive Board member at 07:30, a French central-bank speaker at 08:00 whose name we could not second-source and have therefore cut, a second ECB Executive Board member at 08:05, US Challenger job cuts at 10:30, the Bank of England trio at 12:00, Bailey at 12:30, US unit labour costs and initial claims at 13:30, Hammack at 13:50, Barkin at 14:10, Lane at 15:40, Lagarde at 16:00 and the Bundesbank president at 17:45. It does not carry the Norwegian rate decision or euro-area producer prices that this desk published seven hours ago from three other schedules, and it does not carry tomorrow’s US employment report at all — while its own sister page, from the same publisher and the same morning, calls that report “tomorrow’s.” One of those two is a schedule and one is prose, and the prose is the one that has it.

On the 08:05 slot: this desk published this morning that two calendars put two different central banks in that minute, a Swiss National Bank chairman on one and an ECB Executive Board member on the other. The Tokyo schedule we read today says ECB. That is a third reader siding with one of the two, not a resolution.

Related

• FX event calendar — the 12:00 decision with a forecast round attached, the 12:30 second window, and tomorrow’s 13:30 payrolls now read at the issuer.
• Signals — why a fully priced level and an unpriced vote margin are two different events in one minute.
• EA presets — news windows keyed to a time the issuing central bank publishes without a time zone.