The Reserve Bank of Australia raised the cash rate target by 25 basis points to 4.85 per cent at 03:30 UTC this morning, which is what four separate probability feeds told you would happen. Three of its nine members voted against it. That second fact was not in any of the four numbers, and it is the one that matters, because a probability on a binary outcome is silent about the margin by which the outcome was reached — and the margin is what tells you whether the next one is coming. The Board also pushed the date it expects inflation back to the top of its target band out by about three quarters, to late 2027. The destination did not move. The timetable did.
The decision, from the issuer, with the vote in it
Media Release 2026-30, published on the Reserve Bank’s own site under the title “Statement by the Monetary Policy Board: Monetary Policy Decision”, opens by saying that at its meeting today the Board decided to increase the cash rate target by 25 basis points to 4.85 per cent. The previous level was 4.60 per cent, last set on 29 September.
The release then does something a central bank statement does not always do. It publishes the split: by majority, six members voted to increase the target by 25 basis points to 4.85 per cent, and three members voted to leave it unchanged at 4.60 per cent. No member is named. Nine votes, a third of them against the decision that was taken.
Yesterday this desk published four readings of the probability of exactly this outcome: around 81 per cent on one wire, 83 per cent from ASX 30-day futures, 88 per cent from money-weighted prediction markets and about 80 per cent per swaps on a bank’s Asia note. We refused to pick one and published the band as the reader’s error bar. All four were right about the outcome. None of them was a statement about a nine-member vote that went 6–3, and we are not aware of a feed that quotes one. If you sized today off an 83 per cent reading you were sized for the hike. You were not sized for a hike that a third of the board opposed, which is a different object with a different implication for December.
What changed was the calendar, not the target
The substantive revision is one sentence. Inflation is now forecast to remain elevated through much of 2027 and is not expected to reach the top of the 2–3 per cent target range until late 2027 — which the release itself describes as around three quarters later than previously forecast. In the decision section it adds that inflation is now expected to remain above 3 per cent for the coming year.
Read that against what the Board did and the two fit together uncomfortably well. A central bank that has pushed its own return-to-target date back by nine months, and that expects to sit above 3 per cent for a year, is a central bank with a reason to keep going. The release says the Board judged that a further tightening in financial conditions is warranted to support the return of inflation to target. For context from an Australian broadcaster’s coverage this morning, this is the fifth increase of 2026 and takes the rate to its highest since late 2008.
The forecast revision on the other side of the mandate has no number attached at all. The release says the unemployment rate is projected to rise by more than expected in the August forecasts, and gives neither a level nor a date. It also says the forecasts see growth in output slowing further over the coming year, again without figures. So the Board has told you the direction of two revisions and the size of one, and the one it sized is the inflation one. That asymmetry is itself information about what it is watching.
The guidance says nothing, and one reader filled the gap
Here is the whole of the forward-looking language, as published. The Board will continue to do what it considers necessary to bring inflation sustainably back to target. The Board will be attentive to the data and the evolving assessment of the outlook and risks to guide its decisions. Monetary policy is well placed to respond to developments, and the Board is focused on its mandate to deliver price stability and full employment.
That is three sentences and not one of them has a direction in it. The release does not say the Board will wait. It does not say further increases are likely. It does not name the date of the next meeting. The Australian broadcaster’s summary of the same document says the Bank has left the door open for more rate hikes, and quotes the statement’s line that this presents upside risks to inflation expectations. We want to be precise about the status of that: the door-open reading is a secondary reader’s gloss on language the issuer deliberately kept neutral, and it is a reasonable gloss, but it is not what the Board wrote.
For a system this is the practical point. If your news handler keys off a central bank statement’s guidance sentence, today’s handed it nothing to key off, while the vote count and the forecast-shift sentence both carry content. The fields with information in them this morning were a tally and a date, and neither is where a statement parser usually looks.
The issuer stamped no time, and its own news page does not know this happened
Two weeks ago this desk praised the Reserve Bank’s forward calendar as the best-behaved issuer page in our file, because it distinguished a Governor’s speech from a Senate Estimates appearance when a vendor schedule could not. Today the same institution gives us a release page carrying the line “Date 3 November 2026” and no time and no time zone anywhere on it. The 03:30 UTC decision hour in this article comes from two secondary readers and from a Japanese schedule giving 12:30 Japan time, which is the same instant — not from the issuer.
And the Bank’s own news index does not list the release. The verbatim portion of that page runs up to an Index of Commodity Prices item timestamped 2 November 2026 4.30 pm, with nothing for 3 November, and the page’s own metadata puts its last modification in April. So the media conference time is unconfirmed from the issuer too: earlier 2026 decisions were listed at 2.30 pm with a conference at 3.30 pm local, but that is a pattern from other dates and we are not asserting it for today.
The lesson is not that the Reserve Bank is unreliable. It is that issuer quality is per page, not per institution. The same body that publishes an impeccable event-type field on its forward calendar publishes a decision with no clock on it and a news index three months stale. Storing “RBA pages are good” is storing the wrong granularity.
What this does not tell you
We have no press-conference content. A headline exists asserting that Governor Bullock offered no guidance and said it is probably good to wait for data after the hike, and the page behind it would not open for us. We are telling you the headline exists and publishing nothing from it, because a headline is not an article and this channel has been burned by treating one as though it were. If that quotation is accurate it sharpens everything above; we cannot confirm it.
No market reaction is sourced here beyond one Tokyo reader’s characterisation that the decision carried no surprise and that Australian dollar selling was slightly the stronger side afterwards, and we could not open that page either. We have no Australian dollar level from after 03:30 UTC at all, no AUD/JPY, and no Australian bond yields. An article about a rate rise with no currency print in it is an incomplete article and we would rather say so than derive one.
The inflation figure underneath this decision is single-sourced. One Japanese reader puts the latest quarterly trimmed-mean CPI at 3.7 per cent year-on-year against a 2–3 per cent target. We did not reach the Australian Bureau of Statistics release for it this morning and the Board’s own statement gives no CPI figure.
And a search for today’s decision returned, among the first results, a page from the same calendar date one year earlier headed with the Reserve Bank set to hold rates. It is not this meeting and nothing from it is used. We mention it because a year-blind sampler would have taken a hold story off the top of today’s results, and the correct outcome was a hike that a third of the board voted against.
