Four numbers rarely line up the way they have this year. Growth sits above potential. Unemployment sits below the non-accelerating inflation rate of unemployment (NAIRU). Inflation sits above the target band. And the cash rate sits above the neutral rate, r*, but still short of what a standard Taylor rule would prescribe. Put together, these estimates describe an economy running hot while policy leans against it, but not hard enough.
Growth Has Overtaken Potential
Headline GDP growth through the year is running at 2.14%, although most of that growth happened in the tail end of 2025. Growth in the first half of 2026 was much more modest.
Potential growth - how fast the economy can grow when it is at potential - is now running near 1.99% a year, having drifted down from around 4% in the mid-1990s in a slow, fairly steady decline.
Actual output sits only fractionally above potential: on the log-level comparison, GDP and potential output are running almost on top of each other. But the low rate of potential growth matters for what happens next. With the economy’s sustainable growth rate now around 2%, it does not take especially rapid demand growth to keep output at capacity or push it further above it. An economy that once had room to grow at 3% or 4% without generating much pressure now has considerably less headroom.
The composition of the decline in potential growth is as telling as its size. In the late 1990s, productivity alone contributed roughly 2.4 percentage points a year to potential growth. Through the 2010s that had fallen to around 1 point, and in 2020–2026 its contribution has turned negative. Population growth, by contrast, has consistently contributed around 1.3 to 1.65 points, with participation recently adding a modest lift. Potential growth used to be built on productivity. It is now being sustained largely by population and participation, with productivity working against it.
That framing matters because the growth being generated is not translating cleanly into living standards. GDP per capita growth is only 0.68% a year.
And real net national disposable income per capita, which strips out the terms-of-trade effects that GDP alone misses, is essentially flat at 0.05%.
Productivity helps explain why the aggregate growth numbers have translated so poorly into living standards. GDP per hour worked has compounded at just 0.13% a year over the past decade, near the weakest sustained reading on record.
The output gap is real, and it is doing real work in the inflation story. But it sits on top of an economy whose underlying growth capacity has slowed sharply and is now being sustained largely by population growth and participation rather than productivity. That makes the modest headline growth rate less reassuring than it looks: with potential growth near 2%, it does not take much acceleration in demand to push the economy further beyond capacity.
Labour Markets Run Tighter Than Headlines Suggest
The headline unemployment rate has drifted up to 4.5%, its highest reading of the post-pandemic period. Read in isolation, that looks like slack building. But what matters for inflation is unemployment relative to the economy’s structural unemployment rate. My joint model of potential output and the NAIRU puts that rate at 4.74%, against actual unemployment of 4.35% in Q2 and 4.46% in July. Even after the rise in unemployment, the labour market therefore remains around 25 basis points tighter than neutral.
What stands out is how little that 4.74% estimate has moved. Some of that stability is by design: the model excludes 2020Q2 to 2021Q3 from the fit because the usual concept of an output gap breaks down during lockdowns, when productive capacity was switched off rather than left idle by weak demand. Outside that window, though, the NAIRU continues the gradual downward glide evident since the early 2000s, even as actual unemployment has moved sharply around it. That is a more modest claim than pandemic-proof stability, but it still suggests the structural unemployment rate has changed much less than the cyclical one.
The Phillips curve provides a useful, if noisy, cross-check. Inverting it to ask what inflation implies about the NAIRU quarter by quarter produces an estimate that moves around seventeen times as much as the smoothed joint-model series. Despite that volatility, the two track each other reasonably closely, with a correlation of 0.84, and the latest Phillips-implied estimate of 4.99% is not far from the joint model’s 4.74%. It is not an independent estimate, the two approaches share some of the same economic structure, but it is consistent with current estimate of the NAIRU.
The jobs-flow data tell the same story from a different angle. Employment needs to grow by roughly 21,000 a month just to keep pace with population and hold the unemployment and participation rates steady, the breakeven pace. Trend employment growth fell well below that threshold through the first half of 2025 as unemployment rose. Since then it has recovered to around 28,000 to 29,000 a month, comfortably above breakeven. On current flows, the labour market is re-tightening rather than drifting into slack.
In summary, the labour market remains tight, and in recent months may be tightening further.
Inflation Remains Stuck Above Target
Annual CPI eased to 3.5% in July, down from 3.8% in June, and trimmed mean inflation held at 3.6%. Both sit above the RBA’s 2–3% band. But the single month matters less than the round trip taken to get here.
Every major measure of Australian inflation (headline CPI, trimmed mean, weighted median, and the household and GNE price deflators) fell steadily from highs near 7% in early 2023 to somewhere inside or around the edge of the 2–3% band by mid-2025. Then the process reversed. By mid-2026, most measures had climbed back into a 3.2% to 4% range. That is not inflation getting stuck on the final mile; it is a re-acceleration after getting close to target. One series barely participated in the disinflation at all: the Wage Price Index has remained in a narrow 3.2% to 3.7% range throughout, providing a persistent floor beneath domestic inflation even as more volatile price measures moved around it.
What is driving the current gap matters as much as its size. The model’s decomposition of inflation into target, expectations, demand, supply and residual components attributes most of the current overshoot to demand. Labour-market tightness, measured by the unemployment gap discussed above, is now the largest source of inflation above target, while the supply-side contribution that dominated in 2022 has largely faded. A smaller but persistent contribution from above-target inflation expectations has also remained since 2022. On this decomposition, the inflation problem has changed. What began largely as a supply shock now looks much more like excess demand operating alongside expectations that have yet to return fully to target.
Policy Is Restrictive, But Not Restrictive Enough
The cash rate has sat at 4.35% since May, after two hikes earlier in 2026. Nominal r*, the neutral rate plus the 2.5% inflation target, sits at 3.74%, putting the cash rate about 0.6 percentage points above neutral. Policy is therefore restrictive in the textbook sense. But restrictive relative to neutral is not the same as restrictive enough to close the current gaps. A Taylor rule built on this model’s estimates, and looking through supply shocks, currently points to a cash rate of 5.51%. That should be treated as a benchmark rather than a precise prescription, but the gap of more than a percentage point is still substantial. It suggests the Board is managing the path of adjustment rather than mechanically following the rule. The risk is that a more gradual path also means inflation takes longer to return to target.
That gap is not unprecedented. The same pattern showed up in 2007–08 and again through the 2022 reopening, both periods when the Taylor prescription ran a point or more above the actual cash rate before eventually converging. What stands out this time is the real rate picture underneath it. Real r* has fallen to just 1.24%, a fraction of where it sat in the 1990s and 2000s. That means a 4.35% cash rate can represent substantial restraint even though the nominal rate looks unremarkable beside earlier tightening cycles. It is another reason the Board might be reluctant to chase the Taylor prescription mechanically, even while the gap to it remains wide.
The market is not treating 4.35% as a settled endpoint either. Pricing implied by the ASX rate tracker has the average expected path rising toward 4.7% in 2027 before easing back toward 4.6% by 2028. That path is a probability-weighted blend of possible outcomes rather than a forecast of a sustained tightening cycle. Even so, its upward tilt matters. Markets are assigning a meaningful probability to further tightening rather than treating the current cash rate as clearly sufficient. The question is whether one more hike is enough to turn inflation decisively lower.
What the Combination Implies
None of these four gaps is alarming on its own. What matters is that they all point in the same direction. Output is modestly above potential, unemployment is modestly below the NAIRU, inflation is above target, and the cash rate remains below the model’s rule-based benchmark. The RBA has tightened enough for policy to be restrictive, but not yet enough to clearly turn the cycle. Moving faster risks overshooting as the labour market shows some signs of fatigue. Moving more slowly risks allowing above-target inflation to persist for longer and ultimately requiring a sharper correction.
The near-term test comes at one of the next two Board meetings. September falls at the very end of the month, and markets are now split over whether another hike comes then or is pushed to November. That is a striking change from the near-unanimous view only a few weeks ago that the RBA was done for the year.
Whether the next hike comes in September or November, the underlying picture is the same. Growth is slightly above potential, unemployment remains below the NAIRU, inflation is above target, and the cash rate is still some distance below the rule-based benchmark. None of those gaps is especially large, but unusually, they all point the same way. The question facing the RBA is no longer whether policy is restrictive. It is whether it is restrictive enough.
No comments:
Post a Comment