The headline reads as solid for the year but weak for the quarter. The AI/data centre investment boom is largely discounted because the equipment was imported rather than produced in Australia. Once the imported equipment is netted out, the boom adds nothing to GDP. What remains is an economy still running above its annual speed limit, inflation that has mostly returned to the band but is not all the way home, and a productivity trend that continues to disappoint.
The quarter itself was quite soft. GDP rose just 0.27% in the March quarter, well below the 0.87% of the quarter before and below my own nowcast of around 0.5%. A single quarter's figure is an unreliable guide at the best of times. The through-the-year figure is the one that usually matters, and at 2.52% it sits uncomfortably above the RBA's potential growth estimate of roughly 2%. And that 2% is itself flattered by strong population growth feeding the labour input in the production function, with productivity adding almost nothing. The economy is running hot in the least healthy way: adding bodies and hours rather than output per hour. The fear is that another weak quarter will see a substantial reduction in the through the year figure because of base effects (when the 1pp contribution from Q2 in 2025 drops off).
Of note, and related to the productivity question in the previous paragraph, quarterly per capita GDP growth was negative in Q1-2026. Hopefully this is not a return to the per capita GDP slump of 2023-24.
The Capex Boom That Isn’t
New private business investment jumped 5.72% in the quarter and 10.4% through the year, a sharp acceleration from the flat readings of a year ago. This is the AI and data-centre story made visible in the accounts, and it is the single largest positive contributor to the quarter at roughly 0.6 percentage points.
The trouble is what sits on the other side of the ledger. A server bought from abroad enters the accounts twice: once as private investment, a positive, and once as an import, a negative. The two very nearly cancel, and the domestic value-add from the act of buying a foreign machine is close to zero. That is not an accounting quirk to be argued away. It is exactly what GDP is meant to capture. The boom is real in dollars and very nearly invisible in growth.
Imports Cancel the Investment
The mirror image is right there in the data. Net exports subtracted about 0.8 percentage points from growth this quarter, almost perfectly offsetting the investment contribution. The external sector’s volume share of GDP has fallen to just 0.6%, the lowest since the import surge of the last great capex cycle a decade ago. Then it was mining. Now it is equipment for data centres. The shape is the same: a domestic investment spike pulls in imports, and the trade account absorbs the blow.
This is why the two should be read as one transaction, not two separate forces that happened to net out. The investment and the imports are the same servers, counted at two doors.
Where the Growth Actually Came From
With investment, imports and inventories washing toward zero between them, the quarter’s modest growth rested entirely on household consumption. Note, in the next chart (as with the previous three) the contributions have been rounded to the nearest 0.1 percentage points.
Inflation only slightly elevated
Although the Q/Q household inflation print is below the pace consistent with a 2.5% target, the annual print remains just above the target 2-3 per cent band. Again, another print like this for Q2 will see base effects push inflation back within the band.
Unit Labour Costs Are Easing, Not Solved
The channel from the labour market into prices is unit labour costs, and the calculated measure is running at about 3.3% through the year. That is down from the highs near 7% of the past couple of years.
The Productivity Collapse Underneath It All
This is the number that should worry everyone, and it is worse than “good but not good enough.” Labour productivity, measured as GDP per hour worked, grew just 0.3% through the year and fell 0.6% in the quarter itself. The annual pace has rolled over from around 1% a year ago to essentially flat.
The trend view is starker still. The decade-average growth rate of labour productivity has collapsed to 0.16% for the 2020s, against roughly 1% in the 2010s, about 1.4% in the 2000s and well over 2% in the 1990s. The ten-year moving average has fallen to 0.29%, the lowest reading for decades. This is not a soft patch. It is a structural step down to a rate of productivity growth Australia has not seen sustained in the entire post-war record.
It also ties the whole report together. Productivity is what allows an economy to grow faster without generating inflation. With productivity growth near zero, an economy running above its speed limit cannot expect cost pressures to fade on their own, which is precisely why the household deflator and unit labour costs remain elevated even as the broad measures ease. The capex surge was supposed to be part of the answer to this. So far, in the accounts, it is mostly imports.
What It Means for the RBA
Today's print suggests the RBA is unlikely to hike further in the near term. The Q1 inflation print was good, and short of a catastrophe, should get within target in Q2. The weak Q/Q GDP growth print for Q1 and the base-effects for Q2 means that the RBA will wait and see how that data plays out.
Bottom Line
- A soft quarter, an above-potential year. GDP rose just 0.27% in the quarter but 2.52% through the year. The annual figure is the signal, and it has the economy running a little hot.
- The AI capex boom is largely a GDP non-event. Business investment surged, but the imported equipment that drove it cancels most of the contribution.
- Growth came from household consumption.
- Broad inflation is almost back in the band.
- Unit labour costs are easing.
- Productivity was negative for the quarter, and over the long run remains close to zero for the decade.
- The RBA has no reason to ease and little to tighten. Patience is the path: hold, and wait for the household measures to follow the broad ones home.
No comments:
Post a Comment