The Bank of England raised Bank Rate by 0.25 percentage points to 4 per cent at 12:00 UTC, and the vote was 8–1. The one member against was Catherine L Mann, who wanted half a point and 4.25 per cent. So the headline split understates it: on the question “should Bank Rate be at least 4 per cent today,” this committee voted nine–nil. Seven weeks ago it voted 6–3 the other way. And the forecast round published in the same minute, the one that takes CPI inflation to around 5 per cent by the middle of next year, has Brent crude at $108 a barrel in it — a number the market left behind yesterday.
Six of six switched, and the dissent stayed hawkish
September’s minutes named both sides. Andrew Bailey, Sarah Breeden, Swati Dhingra, Clare Lombardelli, Dave Ramsden and Alan Taylor voted to hold at 3.75 per cent; Megan Greene, Catherine L Mann and Huw Pill voted for 4 per cent. This desk published that arithmetic five and a half hours before the announcement and said two of those six holders had to switch for a hike, with the outcome distribution running from 5–4 to 9–0.
All six switched. Bailey, Breeden, Dhingra, Greene, Lombardelli, Pill, Ramsden and Taylor are the eight; Mann is the one, and she is the one because she wanted more. A committee that was two votes short of a hike in September delivered it with a single dissent seven weeks later, and the dissent is on the same side it was in September. The majority did not win an argument against the hawks. It moved to where the hawks were, and the most hawkish member moved further out.
We also published, in the same piece, what we thought the split would mean: that a 5–4 hike is “a committee that has been dragged over a line and will say so” and an 8–1 or 9–0 is “a committee that has moved as a body and has a second one in it.” That reading scored. Four of the eight said so in terms. Lombardelli: absent evidence of weaker inflation pressures, there is likely to be a case for a further increase in the near term. Pill: achieving the necessary level of Bank Rate is likely to require more tightening, including in the near term. Greene: additional hikes will likely be required. Taylor, alone among the four, pushed the other way — this is not the start of a hiking cycle, policy is not on a pre-set path.
What our range could not express is the thing that actually happened. A distribution written as 5–4 through 9–0 silently assumes the residual votes are holds. They were not. Two days ago we published that the Reserve Bank of Australia hiked exactly as four probability feeds said, that three of nine members voted against, and that nobody quotes a vote margin — so a reader correctly positioned for the hike was positioned for the wrong object. This is the next layer of the same error, and it is ours: we asked for the margin and then wrote a range that could not carry the sign of the dissent. The margin and the direction of the minority are two fields, and we collapsed them into one.
The path is 4.25 per cent in February, and two independent readings agree
The guidance is the standard pair: monetary policy is not on a pre-set path, and the Committee will adjust Bank Rate as necessary to ensure CPI inflation returns to the 2 per cent target in the medium term. Neither sentence tells you anything. The minutes do, in a place that is easy to skip: the Market Participants Survey median expected Bank Rate to reach 4 per cent in November and 4.25 per cent in February.
Set that against something this desk flagged this morning as stale. The only market-implied probability we could find for today’s hike was 81 per cent, dated 9 October and four weeks old — and the same 9 October piece reported two quarter-point increases fully priced by February. That February pricing, which we treated as part of a stale observation, is the same place the Committee’s own November survey of market participants lands. Two readings four weeks apart, one a press report of money-market pricing and one a survey printed in today’s minutes, give the same next step at the same meeting. That is a more useful agreement than any single probability.
And the stale percentage is now closed, at a source we had not read. A prediction-market tracker records the 25-basis-point bucket at 98.5 cents on a 3 November snapshot taken at 15:13 UTC, against a 1.1 to 1.8 cent no-change book, on $451,830 of cumulative volume. Its trajectory is the interesting part: 80.5 cents on 25 September, 92.5 on 15 and 16 October, 94.0 on 17 October, 96.9 on 22 October, 98.5 on 3 November. The four weeks we complained about this morning were worth about seventeen points of repricing. We label this as a prediction-market last trade and not a swaps-implied probability, because it is one, and the two are not interchangeable.
The 5 per cent peak has $108 Brent in it, and Brent settled near $94.6
Here is the number to take away. UK CPI inflation rose to 3.5 per cent in September — which also closes an open item of ours, since this morning we could not tell you which month the Bank’s own 3.5 per cent referred to. It is now expected to rise to around 5 per cent by the middle of next year, with the near-term staff projection at around 4.5 per cent in the first quarter of 2027 and around 5 per cent in the second. Mann, arguing for the larger move, put projected CPI at 4.9 per cent by mid-2027 absent the measures announced at the Budget.
The minutes put the energy inputs on the record: Brent crude and UK gas futures averaged $108 a barrel and 216 pence a therm over the fifteen working days ending 27 October. That window closed nine days before publication. Yesterday Brent settled at about $94.6 on two settlements agreeing to fifteen cents, down between 8.3 and 8.8 per cent in a day, with West Texas near $88.1 and trading into the $86s overnight. Against the figure in the minutes, Brent is roughly $13 lower — about an eighth.
We are not going to claim the Report’s projections are mechanically conditioned on that exact average; the Bank’s conditioning conventions are not something we read today, and the minutes state the average rather than labelling it an assumption. What we can say is narrower and still matters: the energy price this Committee had in front of it when it signed off a 5 per cent inflation peak is a number the market has since marked down by an eighth, and the single largest input to a UK inflation profile moved 8 per cent in one session inside the gap between the data cut-off and the press release. If you are trading the forecast rather than the rate, you are trading a forecast whose biggest exogenous input is nine days stale and wrong in the disinflationary direction.
The Budget is in there too, and quantified. The Committee was briefed on 28 October, the projections and scenarios are conditioned on announced Government policy, and the Budget measures are expected to have a downward impact of a quarter of a percentage point on CPI inflation by the second quarter of 2027. So the two largest identified disinflationary forces on this forecast are a fiscal measure worth 0.25 points by mid-2027 and an oil move the forecast does not contain at all.
Four central banks in four days, and only one of them is finished
Today was not one decision. Norges Bank held at 4.5 per cent at 09:30 UTC, two and a half hours before the Bank of England hiked, and held in the most conditional way available: the committee had not received information indicating the outlook had changed materially since September, it will likely be necessary to keep the policy rate elevated for a time, and it is prepared to raise the policy rate if needed. Norwegian core inflation was 3.0 per cent in September against a target of around 2. September’s hike there was the second of a cycle that began in May. All 26 economists in a 29 October to 2 November poll called today’s hold, and the majority of them expect no further hikes — which is a forecasting community disagreeing with a central bank’s own stated readiness, on one reader’s account.
Behind those two: the Reserve Bank of Australia to 4.85 per cent on Tuesday, six–three, the Riksbank to 2.0 per cent yesterday, and a European Central Bank widely expected to tighten in December. Four decisions in four days, three of them tightenings, and the one hold came with a hike still on the table.
For a system trader the consequence is not directional. It is that the correlated-exposure assumption most books carry — that a long dollar position and a short sterling position are different trades — is weaker this week than usual, because what is moving is the global policy direction rather than any one currency’s. Sizing two legs independently through a week like this is how a book discovers it had one position all along.
What this does not tell you
It does not tell you what sterling did. We are publishing within half an hour of the announcement and we have no post-decision level in cable, euro–sterling, sterling–yen or gilts. What we have is the London morning before it: cable made 1.3501 and 1.3465 with a 1.3481 ticker at 08:32 UTC, having touched the 1.35 handle overnight through a 200-day moving average one publisher puts at 1.3425 — a single-sourced line. The ten-year gilt rallied something between 11 and 14 basis points yesterday, the biggest move on our board, on two readers who disagree about yesterday’s starting level. An 8–1 hike with a hawkish dissent and a 4.25 per cent February path is not the dovish outcome that rally was positioned for, but we have not watched the repricing and we are not going to describe one we did not see.
It does not give you the new forecast table. We read the summary and the minutes. We did not read the Monetary Policy Report itself, so every projection above reaches you through the minutes rather than from the Report’s own fan charts, and we have no GDP, unemployment or market-path numbers from it. The Report page was still serving its pre-publication placeholder when we looked.
It does not tell you Mann’s vote means a second hike is coming faster. One member wanting 50 basis points is one member. She was already in the minority in September and in July. The four near-term-tightening quotes above are the better evidence, and even those are individual member views recorded in minutes, not a Committee position — the Committee’s position is the sentence saying policy is not on a pre-set path.
It does not second-source Norway. Norges Bank’s own press release returned a client error on the URL we had. Every Norwegian figure above, including the 4.5 per cent level, the Wolden Bache quotes, the 3.0 per cent core rate and the 26-economist poll, is a secondary reading of a release we could not open. The Riksbank and European Central Bank lines come from the same single page.
And it does not tell you the oil gap is tradeable. A stale conditioning input is a reason the published forecast is probably too high, not a reason the next print is lower. UK CPI at 3.5 per cent and a committee that has just gone nine–nil on “at least 4 per cent” can carry a weaker oil path and still hike in February, and the Brent average in the minutes may not be what the projections are conditioned on at all.
Related
- FX event calendar — the 12:00 decision with a forecast round attached, 17 December as the next announcement, and tomorrow’s 13:30 payrolls.
- Signals — why a vote margin and the sign of the dissent are two fields, and why a range that collapses them is not a forecast.
- EA presets — sizing two legs independently in a week when four central banks move in the same direction.
