At 13:15 UTC the European Central Bank is expected to leave the deposit rate at 2.50 per cent, six weeks after raising it there. Euro-area headline inflation printed 3.8 per cent year-on-year in September against a market view of 3.6 and a prior reading of 3.2 — a six-tenths jump in one month, 1.8 points above the target, with energy running 18.9 per cent year-on-year and contributing 1.71 points of the headline on its own. The Council is expected to do nothing about it, and the quarter’s growth number publishes tomorrow morning, after the decision it was supposed to inform.
A hold is the right call and it is not a comfortable one
The case for holding is straightforward and it is mostly about composition. A headline that moves from 3.2 to 3.8 because energy went from 14.3 to 18.9 per cent year-on-year, lifting its contribution from 1.29 to 1.71 points, is a headline that has been moved by one line item. Core measures ticked higher and the preview reading we have calls the rise relatively modest. A central bank that hiked in September and watched the energy component do most of the subsequent damage has a defensible reason to wait rather than chase it.
What is not comfortable is the number you would have to say out loud at the press conference. Headline inflation is nearly double the target, the single largest contributor is a price the Council cannot influence, and the policy response is to leave the rate where it was. That is the honest position — and it is also the position that becomes indefensible fastest if the energy pass-through shows up in core. The thing to listen for at 13:45 is not the word “vigilant.” It is whether anybody on the panel is asked to put a number on how much of 3.8 the Council thinks is transitory, and whether the answer is a number.
Twenty basis points by year-end. One or two of them today.
The pricing is the useful part, because it tells you what today is actually worth. One preview reading puts the market at roughly 20 basis points of tightening by end-2026, with only one or two of those basis points attributed to this meeting. There is one more meeting after today before the year ends. So the market has put essentially the entire remaining move into December and essentially nothing into the next ninety minutes.
Read that as a risk budget rather than a forecast. An event priced at one or two basis points cannot deliver a large dovish surprise, because there is nothing to take out. It can deliver a hawkish one — the preview we read says a back-to-back hike cannot be entirely ruled out, which is preview language for “nobody expects this” — and that asymmetry is the whole trade structure of the next hour and a half. If you are sized for a two-way event at 13:15 you are sized wrong in both directions: too large for the decision, too small for the tail.
The December meeting is where the twenty basis points live, and it is the one you should be writing down. We are not giving you a probability for it. The number we have is an aggregate of tightening priced to year-end from a single reader, not a meeting-by-meeting implied path, and this desk has spent six weeks separating those two objects at the Federal Reserve. We are not going to merge them here.
The Fed published a vote. The ECB will not.
Nineteen hours ago the Federal Open Market Committee held at 3-3/4 to 4 per cent, ten votes to two, and both dissents were for a quarter-point increase. This desk published at 23:12 UTC that the entire hawkish signal of that meeting lived in the vote, because the statement carried no path language, and that the vote is the field most systems do not read. The two-year yield fell 3.1 basis points to 4.706 on the day while the ten-year rose 3.3 to 5.024 and the thirty-year 4.6 to 5.351 — roughly seven and a half basis points of steepening on a meeting where two officials asked for tighter policy.
Today the field that carried all of that does not exist. The Governing Council does not publish a vote count, does not name dissenters, and does not release a tally with the decision. Whatever disagreement exists inside the room today reaches you as tone in a press conference, or three weeks later in an account, or not at all. So the inversion is exact: at Wednesday’s meeting the statement was empty and the vote was the information; at today’s there is no vote and the statement and the press-conference questions are all there is.
If you built anything after Wednesday that reads a dissent count — and the argument for building it was good — it has no input today. The equivalent field for this central bank is the press conference transcript, which arrives as prose, late, and cannot be scored. That is not a complaint about the ECB. It is a reason not to assume a decision-reading process generalises across issuers just because both objects are called a rate decision.
Data-dependent, one day before the data
The Council’s stated approach is data-dependent and meeting-by-meeting. Here is what publishes in the twenty-one hours after it decides: German flash GDP and the ECB’s own Survey of Professional Forecasters at 09:00 UTC tomorrow, euro-area flash GDP for the third quarter at 10:00, German flash CPI at 13:00. The quarter’s growth number and the next month’s inflation signal both land the morning after the decision.
The best public read on euro-area growth going into today is therefore a survey: the October flash PMI, which one reading says implies quarterly GDP growth of 0.3 per cent, slightly slower than September’s data implied. That is a composite of business opinion standing in for a national accounts aggregate, and it is what the market has. We are not claiming the Council is working from the same information as the public — national statistics institutes compile flash GDP and we have no basis for saying who sees what when. The claim is narrower and it is about you: “data-dependent” today means dependent on a quarter that ended four weeks ago and a month that ended four weeks ago, and the two releases that would mark the Council’s view to market arrive after you have already traded the decision.
Which means the asymmetry runs into tomorrow, not out of it. A hold today with a non-committal tone, followed by a weak flash GDP at 10:00 tomorrow, reprices that December twenty basis points more than anything said at 13:45 will. Tomorrow at 09:00 and 10:00 is the event this one is a preamble to.
Tonight’s seven-year, scored against a seven-year
The United States sells 44 billion dollars of seven-year notes at 17:00 UTC, two-sourced in our notes for size and time. The baseline is the last seven-year auction and nothing else: a vendor calendar gives the previous seven-year high yield at 5.085 per cent, single-sourced, on a page whose clock field for every other row sits an hour before the issuer’s. We are quoting the number and flagging the page, because this file has established that this kind of source gets labels wrong far more often than it gets figures wrong.
The comparison that is not available is a when-issued level, so there is no tail to compute in advance. The comparison that is available is the curve the auction lands on: a ten-year at 5.024 and a thirty-year at 5.351 after yesterday’s steepening. A seven-year clearing above 5.085 into a curve that just steepened on a hawkish hold is a different event from one clearing below it, and the distinction is about the price of money rather than about the dollar. Score it against the seven-year. The last auction and the last auction of this tenor are different fields, and only one of them is the baseline.
What this does not tell you
The euro-area inflation breakdown, the PMI-implied growth rate and the twenty-basis-point year-end figure all come from one preview. FXStreet, the second reader this desk would normally use on all three, returned a site error on every attempt this morning, so these are one source’s numbers and should be treated as one desk’s view until a second reader confirms them. The 13:15 decision time is the opposite case — four readers agree on it.
We are not telling you the ECB will hold. Every reader we have says it is expected to, the pricing says one to two basis points, and that is a market view rather than a fact. Nor are we giving you a December probability: an aggregate of tightening priced to year-end is not a meeting-by-meeting implied path, and we will not convert one into the other to make a cleaner sentence.
The Federal Reserve figures above are this desk’s own published numbers from yesterday, carried forward, not re-measured today. The seven-year prior is single-sourced. And the Bank of Japan decides tomorrow at a time its own calendar still gives as not yet set, with the governor’s press conference at 06:30 UTC — so the single largest yen event of the week has a published hour for the commentary and none for the decision. Tokyo October CPI prints at 23:30 UTC tonight, three and a half hours before that meeting resumes.
