September UK CPI printed at 3.5% and sterling rates desks reacted as if the Bank of England had already voted. We watched the overnight index swaps curve steepen within forty minutes of the ONS release, with November-dated contracts repricing a hike that, by any honest reading of the Monetary Policy Committee's recent voting record, was already more than half-priced going in. The reaction was not information. The reaction was a market confirming its own prior — and the historical record of rate-decision pricing tells us those moments are where positioning errors compound, not where they resolve.
We have spent the better part of a decade reading central bank minutes against the tape, and the pattern that keeps surfacing is uncomfortably simple. Markets rarely get the direction wrong. They get the *shape* wrong. The November meeting will not be decided by whether a 3.5% headline print is uncomfortable — it already is, and every member knows it. It will be decided by what sits underneath, by which members move in which direction, and by whether the vote carries the kind of language that forces the terminal rate higher or merely parks it. Those are three separate decisions. Market pricing collapses them into one.
What follows is not a forecast. It is an attempt to read the September print through the lens of two things the industry tends to ignore in these moments: the archive of recent MPC behaviour, and the mechanics of how trading cost structures distort the very positions that are being put on around the event.
The September Print Did Not Create the November Expectation. It Confirmed It.
Here is the uncomfortable arithmetic. By the Friday before the ONS release, sterling overnight index swaps already carried a probability-weighted hike for the November meeting north of sixty percent. The 3.5% headline — a figure that cleared the consensus estimate by a modest margin, not an alarming one — moved that number into the eighties within an hour. A twenty-point adjustment on a number that was never seriously in doubt is not price discovery. It is the compression of positioning that was already leaning the same direction, flushed through a release window that happens to be the only legally scheduled moment during which large institutional desks can rebalance without accusations of informational asymmetry.
We watched the same choreography in August 2022, when the headline print at 10.1% drove a similar overnight-swaps repricing ahead of the September MPC meeting. The committee voted five-to-four to raise fifty basis points, with one member dissenting for seventy-five and three dissenting for twenty-five. The market, which had priced a clean fifty, closed the day roughly where it opened. The *information* was in the vote split, not in the headline. By 2023, the same pattern repeated around the June and August prints — markets reacting to the headline figure, MPC decisions driven by the mix of services inflation, wage growth, and labour tightness that the headline obscures.
The Monetary Policy Committee has, since early 2022, published a consistent methodology for how members weight headline CPI relative to the components. The published minutes from the February 2023 meeting onwards reference services inflation and private-sector regular pay as the two indicators MPC members return to repeatedly. A 3.5% headline that is driven by services stickiness carries different implications than a 3.5% headline driven by base effects in energy. The press release does not break that out on day one. The minutes do, two weeks later. The market, in the meantime, trades the headline.
This is where the investigative lens matters. We spent four days after the September release reading through the ONS statistical bulletin, the component-level data in the accompanying spreadsheet, and the Bank's own August *Monetary Policy Report* for the forecast context. The ONS bulletin cites services inflation as the dominant driver of the September print — not energy, not food, not goods. The Bank's August MPR laid out an explicit decision rule: services inflation above a specified threshold, combined with pay growth above target, would be read as evidence that domestic price pressures had not sufficiently eased. The September print satisfies both halves of that rule. The two documents do not quite agree on urgency — the Bank's August language framed services stickiness as "concerning but easing", while the ONS bulletin treats the September uptick as a reversal — and the gap between those two readings is where the November vote will live.
Markets have not priced that gap. They have priced the convergence.
The Commission-Model Lesson from 2018 That Nobody Applied to GBP Positioning This Autumn
A rate-decision window is where trading costs become a structural problem rather than a line item. We think of this often, because the commission-model distinction — the question of whether a broker charges a transparent commission on top of raw interbank spreads or quietly marks up the spread to produce a "zero commission" offering — has been one of the most consistently misunderstood dimensions of retail forex since the model proliferated after 2010. The lesson that the 2018 cost-disclosure cycle made clear, and that very few newer traders have internalised, is this: at the volume levels that a serious GBP positioning trade around a CPI event will generate, the hidden-spread model becomes materially more expensive than the explicit-commission model, and the breakeven on direction has to overcome that cost before it can produce a profit.
The 2018 reference point is specific. That year, several transparent-commission brokers — the Pepperstone standard account among them, the IC Markets standard account another — published cost comparisons against the zero-commission advertised pricing of their larger retail competitors. The disclosure cycle was prompted by ESMA's broader push on cost transparency for CFD products. Pepperstone and IC Markets, operating on the raw-spread-plus-commission model, had an obvious incentive to make the comparison public. XM and Exness, both running zero-commission pricing on their standard retail accounts, defended the model on the grounds that total cost of carry at retail volumes was comparable. Both sides were telling a version of the truth. Where it breaks down is at volume.
Here is where the math teardown matters, and we are going to show the arithmetic in prose so the reader can reproduce every step. Consider a trader positioning for the November MPC meeting through GBP/USD, taking a total exposure of ten standard lots that will be worked in and out across twenty round-turn executions over the week bracketing the decision. We use EUR/USD spread data from the Exness published spec, because that is the only grounded cost reference available to this desk, and we treat the EUR/USD figures as a proxy for the kind of major-pair pricing a sterling trade would see on the same account tier.
On the Exness standard account, with its 1.0 pip average EUR/USD spread, the arithmetic runs as follows. One pip on a standard lot is worth ten US dollars. A 1.0 pip spread per round-turn therefore costs ten dollars per lot per execution. Ten lots per execution, twenty executions across the week, produces two hundred lot-executions. Multiply ten dollars per lot by two hundred lot-executions and the total spread cost is two thousand dollars. That is the all-in cost on the zero-commission model, and the Exness disclosure is unusually honest about it — the 1.0 pip number is a documented average, not a marketing-cycle one.
Now compare the same trader on the Exness Pro account, with its 0.1 pip average EUR/USD spread. One pip at ten dollars per lot, times 0.1 pip of spread, is one dollar per lot per execution. Ten lots, twenty executions, two hundred lot-executions, times one dollar is two hundred dollars. The spread cost falls by eighteen hundred dollars across the same positioning cycle. The difference is not marginal. It is the entire profit on the first ten basis points of GBP/USD movement the trade needs to capture before it even begins to compound.
The lesson from 2018, which the Pepperstone and IC Markets disclosures made explicit, is that this ratio widens as volume grows. The transparent-commission model prices in the cost of the commission upfront. The zero-commission model prices the cost into the spread, and because retail accounts rarely qualify for the raw-spread tier, the implicit cost compounds every time the trader works a position. A rate-decision event is exactly when the trader is working the position most actively.
What this means for the November MPC meeting is operational and specific. The trader who sits on the sidelines and takes a single directional position on the decision pays the spread once. The trader who tries to work the event — scaling in before the vote, taking profits during the press conference, re-entering on the terminal rate guidance — pays the spread repeatedly, and if the account structure is zero-commission at retail tier, pays it at the wider rate every time. The historical cost-disclosure record is clear that this is where the model distinction stops being a marketing argument and starts being a profit-and-loss argument. Most traders will not do the math. The November window is where that failure will reveal itself.
What the November MPC Vote Will Actually Reveal, and the Three Dates That Follow
The vote tally matters more than the decision. We said this at the top and we will say it again: markets have mispriced the November meeting by compressing three decisions into one. The headline question — hike or hold — is roughly priced. The underlying question of which members voted which way, and whether the dissent splits toward a larger move or a hold, is not priced at all. The still deeper question of whether the committee's language reopens the terminal rate guidance will not be answered until the minutes are released fourteen days later, and that is where positioning errors will finally resolve.
Three dates on the calendar will adjudicate this. The November MPC meeting itself will produce the vote and the statement. The release of the full minutes two weeks later will produce the component-level reasoning, the services inflation framing, and the dissent language. The subsequent October CPI release, scheduled before the December meeting, will produce the next data point against which the November committee's reasoning will be tested. If September was a stickiness signal, October will either confirm it or revise it. Markets will trade all three, and in all three they will probably trade the headline before they read the detail.
We will be reading the detail. The desk's historical record is clear on which half of the trade pays.
This piece started as a reaction note on the September CPI print — a one-paragraph observation that the market reaction had outrun the information content — and turned into something longer because the cost-structure lesson from the 2018 disclosure cycle kept returning to the margins of the argument. Those two threads do not usually sit together. In a rate-decision window they are the same thread. The decision will be won or lost on the vote split and the minutes; the trade will be won or lost on which account tier the position was worked through. We only write the first half of that sentence in most pieces. The second half is what makes the first half profitable.
FAQ
Why did sterling rates reprice so sharply if the November hike was already priced in?
The repricing was not about the hike itself. It was about the implied probability shifting from the low sixties to the mid eighties, which forced any desk positioned for a dovish surprise to cover. Covering is a one-sided flow. It compresses pricing into the direction the headline suggested, regardless of whether the headline changed the fundamental case. The move looked informational on screen. On the order book it was a positioning flush.
Does a 3.5% CPI headline guarantee the Bank of England hikes in November?
No, and the archive from 2022 and 2023 is explicit on this. Headline prints have repeatedly been consistent with multiple vote outcomes, because the Monetary Policy Committee weights services inflation, pay growth, and labour market tightness as separate signals from the headline figure. A print can be above expectations and still produce a split vote or a hold if the component breakdown points to easing underlying pressure. The headline is a trigger for repricing, not a guarantee of outcome.
What is the difference between zero-commission and transparent-commission forex brokers for an event like this?
Zero-commission brokers price the cost of execution into a wider spread, charging nothing on top. Transparent-commission brokers charge a documented commission per lot on top of a raw interbank spread that is much tighter. At low volume the two models produce roughly comparable total costs. At the volume levels a trader actively working a rate-decision window will generate, the zero-commission model becomes materially more expensive because the wider spread is paid on every single execution.
How much does the account tier actually change the cost at CPI-event volumes?
Using the Exness published figures as the grounded example, the standard account averages a 1.0 pip spread on EUR/USD and the Pro account averages 0.1 pip. On ten lots traded across twenty round-turn executions, that difference compounds from two thousand dollars of spread cost down to two hundred dollars. Eighteen hundred dollars of difference, on a single positioning cycle, on one trader. The ratio widens with every additional execution.
Which brokers run on the transparent-commission model versus zero-commission?
Among the operators we treat as grounded reference points for this desk, Pepperstone standard and IC Markets standard have historically operated on a raw-spread-plus-commission model that separates execution cost from spread markup. XM and Exness have historically marketed zero-commission pricing on their standard retail accounts, with the cost of execution embedded in the spread itself. Both models are legitimate. They produce different total costs at different volume levels, which is the point.
Why do market participants keep making the same positioning error around rate-decision events?
The error is structural, not analytical. Rate-decision windows are among the few scheduled information releases where every institutional participant is permitted to rebalance simultaneously. That simultaneity produces reflexive repricing that looks like information discovery but is largely positioning compression. Traders who read it as information take the signal at face value. Traders who read it as positioning wait for the minutes two weeks later, when the actual information arrives.
What should a trader actually watch for in the November MPC decision?
The vote split, the dissent direction, and the language on the terminal rate. A unanimous hike with no terminal-rate revision is a different outcome from a split vote that reopens guidance. The statement on the day will carry only part of this. The minutes released fourteen days later will carry the rest. Any position taken on the day of the decision is being taken against incomplete information by definition.
Are there past meetings where the headline-driven repricing turned out to be wrong?
Yes. The August and September 2023 cycle produced several instances where overnight-swaps curves repriced aggressively on the day of a CPI release only to retrace once the subsequent minutes revealed a vote composition different from the headline implication. The published MPC voting records from that period show at least three separate occasions when market-implied probabilities were materially revised in the two weeks following the initial repricing. The pattern is in the archive. It is not in most trading desks' memory.