The Bank changed which inflation number counts, then kept watching the old one

On 11 August the Reserve Bank left the cash rate at 4.35 per cent. The cash rate is the interest rate the Bank sets, and almost every other rate in the country prices off it. The statement noted three increases since the start of the year, described policy as somewhat restrictive, and said inflation is not expected to return to the middle of the target range until late 2027. It added that the risks sat on the high side of that forecast. The Board said it would raise the cash rate further if those risks materialised.

Read that against something the Bank published nine months earlier, in a technical annex almost nobody outside the market has opened.

The benchmark changed underneath everyone

In November 2025 the Australian Bureau of Statistics replaced its quarterly inflation figure with a monthly one as the country’s main measure of headline inflation. Headline inflation is the change in the price of everything a household buys, taken together. An older monthly figure had existed since 2022, but it only covered part of the shopping basket. It was discontinued in October 2025.

That was not a technical reshuffle. The Reserve Bank stated plainly in its November Statement on Monetary Policy that headline inflation from the complete monthly inflation figure becomes the new target for monetary policy. The mandate — two to three per cent — is now measured against a series that began publishing in late 2025.

And the Bank told us it would not be using it yet

The same publication set out the transition arrangements. The statistician will keep publishing the old quarterly figures for at least eighteen months. The reason is seasonal adjustment: prices swing predictably at certain times of year, and stripping that out takes several years of data before it can be done reliably. For the monthly series, that work will not be finished across every item until the middle of 2027.

During that period, the Bank said, it will keep focusing on the quarterly trimmed mean. That is a measure which throws away the biggest price rises and the biggest price falls, leaving the middle — the idea being to show the underlying trend rather than one-off shocks. The Bank prefers it because its behaviour is well understood.

The number the target is written against and the number the Board actually watches are not the same number, and will not be until 2027.

Why this is defensible and still a problem

Take the Bank’s case seriously first, because it is a good one. Steering off a brand new series would be reckless. Any new measure needs years before anyone can tell a real signal from noise. The Bank said as much: it welcomed the monthly figure while warning it may jump around from month to month. Using the mature series through the changeover is straightforwardly the right call.

The problem is not the choice. It is the gap between where that choice was disclosed and where the public conversation happens.

Every monthly inflation figure print now lands as the inflation number. It is the benchmark for the target, so that is how it gets reported, how the futures curve reprices, and how households read whether their mortgage is going up. Meanwhile the Board is weighting a quarterly series most of that audience has stopped watching, for reasons set out in a technical note attached to a quarterly publication.

The credibility cost

This matters more than it would have three years ago, because of what the Bank has just spent eighteen months doing. Three cuts through 2025, then three increases from the start of 2026, arriving back at the same 4.35 per cent the Board set in November 2023.

Whatever you think of that round trip — and there is a real defence, in that the world genuinely changed — it consumed the Bank’s stock of assumed competence. An institution in that position cannot also afford a public misunderstanding about which number it is watching. When the monthly print moves and the Board does not, the reasonable reader concludes the Bank is ignoring the data. The Bank’s actual position, that it is weighting a different and better-understood series during a transition it announced, is correct and almost entirely unheard.

What I would do

Say it in the decision statement, not the annex. One sentence, every meeting, until mid-2027: which series the Board weighted, and why. The transition arrangements were published honestly. They were simply published where only the market reads.

The counter-case

Two objections worth taking seriously. The Bank has not hidden anything. The arrangements are public, detailed, and were flagged before the switch. There is also a limit to how much method belongs in a statement that has to move markets in the same breath. Overloading it has its own costs.

And the two series may not diverge enough to matter. The ABS has published analysis showing the quarterly data derived from the monthly collection tracking the previously published quarterly series closely. If they stay close, the distinction is academic and I am manufacturing a problem out of a footnote.

How you can tell if this is wrong

Watch the decisions where the two series point differently. If the Board moves with the monthly inflation figure in those months, the target and the decision variable are effectively the same thing and this argument collapses. If it holds against a monthly print while citing quarterly underlying measures, the gap is real, and the communication problem is the one worth fixing before the next turn in the cycle.