Another Forty Years of Optimism
Treasurer Jim Chalmers has released the 2026 Intergenerational Report, projecting an Australian economy more than twice its current size and per-capita incomes 55% higher by 2066.
Underpinning those projections is an assumption that long-term labour productivity growth will reach 1.2% annually. The Treasurer has identified AI as a pivotal contributor to achieving it.
That sounds encouraging.
Unfortunately, Australia has a rather inconvenient productivity record.
The Numbers Tell a Different Story
The Productivity Commission's September update paints a picture of prolonged stagnation.
Australia's Productivity Reality
Over the twelve years to June 2026, economy-wide labour productivity increased just 3.6%. The non-market sector—dominated by government-funded services—actually recorded a decline of 2.5% over the same period.
The economy has barely recovered from the productivity slump that followed the pandemic.
And this is the starting point from which Treasury expects a sustained recovery.
We've Heard This Before
The first Intergenerational Report, published in 2002, assumed annual productivity growth of 1.8%.
Twenty years later, actual growth had averaged just 1.2%.
Yet aggregate GDP growth came remarkably close to Treasury's original projection. How?
Australia's population grew by approximately three million more people than forecast, largely because net overseas migration was roughly double the original assumption. Workforce participation also exceeded expectations.
The economy grew, but not in the way Treasury anticipated.
Subsequent reports continued to revise productivity expectations downward. The 2021 assumption of 1.5% was reduced to 1.2% in 2023.
The 2026 report retains that figure, despite another three years of disappointing performance.
The historical lesson is not that productivity cannot recover. It is that long-term forecasts have repeatedly overestimated Australia's capacity to deliver it.
Enter the AI Miracle
Artificial intelligence undoubtedly offers enormous opportunities. It can automate administrative work, accelerate research, improve logistics and make skilled workers more productive.
But purchasing AI software does not automatically produce higher national productivity.
Businesses must reorganise their operations, retrain employees, invest in infrastructure and eliminate inefficient practices. Some will succeed. Others may simply add AI subscriptions to their existing costs.
Even Reserve Bank Governor Michele Bullock has cautioned that there is not yet evidence of an Australian productivity improvement attributable to AI. She has also raised the possibility that the investment boom in AI infrastructure could initially contribute to inflation rather than improved productivity.
There is an important difference between the potential of a technology and the capacity of an economy to exploit it.
Australia has been adopting digital technologies for decades. Yet productivity has continued to deteriorate.
Why should AI automatically produce a different result?
The Arithmetic of Wishful Thinking
Treasury's 1.2% assumption might appear modest compared with the productivity growth Australia enjoyed in earlier decades.
But consider the consequences if actual performance falls short.
Illustrative cumulative productivity index, starting at 100. Three constant annual growth scenarios; not forecasts.
Calculated using annual compounding.
After forty years, sustained growth of 1.2% produces approximately 61% more output per hour. At 0.8%, the improvement is only 38%. At 0.3%, it is barely 13%.
The difference is enormous. It affects wages, tax revenue, government spending and the living standards of future generations.
Treasury's projections are scenarios built on assumptions, not guarantees. But when an optimistic assumption becomes the foundation of fiscal planning, its failure has consequences.
Technology Is Not a Substitute for Reform
Australia's productivity problem predates AI.
It reflects a complex combination of weak business investment, poor capital allocation, declining productivity in government-funded services and an economy increasingly concentrated in activities where efficiency improvements can be difficult to achieve. The Productivity Commission has identified investment, skills and competition as important areas for reform.
AI may help address some of these problems. It cannot resolve all of them simply by existing.
The question is whether Australia's businesses and institutions will make the organisational changes needed to convert technological potential into measurable economic gains.
The experience of the past two decades provides ample reason to scrutinise the assumptions.
Treasury has already had to reduce its productivity forecasts once. The new report relies on the same 1.2% figure despite a much weaker recent record.
AI may transform the world. But Australia's economic projections depend on something much less certain: our ability to turn that transformation into productivity.
Further reading: 2026 Intergenerational Report, Productivity Commission's September 2026 update and Treasury's review of previous Intergenerational Reports.
Further Reading from Grappy's Soapbox
- The AI Race Will Be Won With Energy — Explores the electricity, data centres and industrial infrastructure needed to realise AI's potential. It complements the new article's distinction between technological promise and economic results.
- How Much Migration Can Australia Absorb? — Examines the distinction between growth in total GDP and improvements in living standards, including the importance of productivity.
- Climate, Energy and Prosperity: The Debate That Shapes Our Future — Considers how energy costs, infrastructure investment and energy policy affect economic prosperity.
