Two data releases landed ten days apart in August, and they pointed in opposite directions. On 20 August, the ABS confirmed unemployment had climbed to 4.5% - the highest level of the post-COVID era, with employment actually falling by 15,800 people when the market had expected a rise of 12,000. On 26 August, the ABS confirmed underlying inflation was still stuck at 3.6%, unchanged for the eighth straight month and hotter than every economist's forecast.
A weakening jobs market usually argues for the Reserve Bank to hold, or even cut. Sticky inflation argues for a hike. Both landed in the same fortnight, and three of the big four banks have already picked a side.
The ten days that split the big four
NAB was first to move. Within a day of the July inflation print, its economics team abandoned its "next move is down" call and became the first major bank to forecast a hike at the Reserve Bank's 28-29 September meeting, taking the cash rate from 4.35% to 4.60%. The trigger wasn't the headline number - annual CPI actually eased to 3.5% from 3.8% - it was what sat underneath it. The trimmed mean, the Reserve Bank's preferred measure of underlying inflation, rose 0.5% for the month, the biggest monthly increase in a year and well above the 0.3% economists had pencilled in. Annually it held at 3.6%, a figure that hasn't dropped since November last year.
ANZ and CBA followed within days, though both pushed their hike call out to November rather than September. That leaves Westpac as the lone holdout, still forecasting no move for the rest of 2026. Its argument: the hottest components of the July print - durable goods and discretionary services - look like one-off timing quirks rather than broad-based pressure, and the labour market is already softening enough that the Reserve Bank will look through a single hot month rather than react to it.
The number that should have settled this
If any single data point was going to end the hiking conversation, it should have been the jobs figures. Unemployment at 4.5% isn't just a tick up from June's 4.4% - it's the highest reading since the pandemic distortions washed out of the data. Underemployment held at 6.4%, and the participation rate slipped to 66.9%, meaning some of the softness is people quietly stepping back from the workforce rather than actively losing jobs. None of it looks like an economy with room to absorb another rate rise comfortably.
And yet the Reserve Bank's own August statement didn't treat the jobs wobble as decisive. The Board held at 4.35% on 11 August, but kept its hiking bias explicitly on the table - it will "raise interest rates further if that is what is required to bring inflation down in a timely way." Governor Bullock flagged upside risks from global oil prices and the Middle East conflict, plus the possibility of "more persistent domestic capacity pressures." Notably, the Board didn't even discuss a cut. The choice on the table was hike or hold - nothing softer.
Why the Reserve Bank might hike into a weakening labour market anyway
This is the scenario the Reserve Bank's own deputy governor called "a central banker's nightmare" earlier this year: inflation running hot while growth and employment cool at the same time. Standard policy doesn't have a clean answer for it, because the two problems pull in opposite directions and only one interest rate lever exists to respond to both.
The Bank's logic, as best it can be read from the August statement, is that inflation expectations are the thing that can't be allowed to drift. If underlying inflation is still 3.6% eight months running while everyone assumed it would fade, a wage-price spiral becomes a live risk - and that risk compounds the longer it's tolerated. Unemployment ticking up is uncomfortable, but at 4.5% it's still below levels that would typically force the Bank's hand toward relief. Westpac's counter-argument is that this reasoning under-weights how quickly a labour market can turn once it starts softening, and that reacting to a single hot inflation print risks a policy mistake in the other direction.
What a hike - or two - actually costs you
On a $600,000 loan with 25 years remaining, a 25 basis point hike passed on in full adds roughly $91 a month to minimum repayments. If September and November both land, that's closer to $183 a month - over $2,000 a year - on top of everything already absorbed from three hikes earlier in 2026. Roy Morgan's most recent mortgage stress reading puts 30.3% of households, around 1.6 million people, at risk to June 2026, and that was before this month's data. Use our mortgage calculator to see exactly what one or two more hikes would do to your own repayment, rather than working from the averages.
What to actually do before 29 September
The lesson from six months of forecast flip-flopping isn't to guess which bank will be right. It's to build a plan that survives being wrong either way.
- Stress-test at rate-plus-two-hikes, not just today's rate. If your budget only works at the current repayment, a hike shouldn't be a surprise you discover after it happens.
- Revisit fixed versus variable now, not after the decision. Fixed rates are priced off swap markets, not the cash rate directly - by the time the Board actually moves, any repricing may already be baked into fixed offers.
- Don't treat any single bank's call as a timetable. NAB, ANZ and CBA disagree with each other on timing by two months; Westpac disagrees with all three on direction. That spread is the honest picture, not a rounding error.
- Rebuild your buffer before it's tested. An offset or redraw buffer sized for one more hike is cheaper to build now than to find under pressure in October.
- Check your borrowing capacity if you're about to apply. Every hike that's priced in already tightens the buffer banks assess you against. Our borrowing power estimator gives a starting figure before you get a formal pre-approval.
Whichever way the Board moves on 29 September, the more useful fact is the one already sitting in plain sight: the people whose job it is to forecast this couldn't agree in August, and the data released since has only made the disagreement more defensible on both sides.
Frequently Asked Questions
Will the RBA raise interest rates in September 2026?
It's genuinely contested. NAB forecasts a 25 basis point hike to 4.60% at the 28-29 September meeting, based on underlying inflation holding at 3.6% for eight straight months. ANZ and CBA also expect a hike but timed for November instead. Westpac is the only major bank still forecasting a hold for the rest of 2026, pointing to a softening labour market as the offsetting factor.
Why would the RBA hike rates when unemployment is rising?
Because the RBA's mandate weighs inflation and employment together, and its August statement treated inflation as the more urgent risk. Governor Bullock has flagged that inflation expectations becoming unanchored is harder to reverse than a modest rise in unemployment. The Board's August statement kept a hiking bias explicit and did not discuss a rate cut, even after the unemployment data.
What does a 0.25% rate hike cost on an average mortgage?
On a $600,000 loan with 25 years remaining, a 25 basis point rise passed on in full adds around $91 a month to minimum repayments. Two hikes in succession, such as a September and a November move, would add roughly $183 a month, or more than $2,000 a year, on top of the three hikes already absorbed earlier in 2026.
What is the trimmed mean and why does it matter more than headline CPI?
The trimmed mean removes the most extreme price movements each quarter to show the underlying inflation trend, which is why the RBA treats it as its preferred measure. In July 2026, headline CPI eased to 3.5%, but the trimmed mean held at 3.6% for the eighth consecutive month - the reason economists and the RBA read the print as hotter than the headline suggested.
Should I fix my mortgage rate before the September RBA decision?
There's no universal answer, because fixed rates are priced off swap and bond markets rather than the cash rate itself, and can move before or independently of an actual RBA decision. The safer move is to model your own repayments under both a hold and a hike scenario using a mortgage calculator, then compare current fixed offers against that, rather than trying to time the announcement itself.
