“Everything costs more than it did three years ago and the margin is thinner, but I can't put on another admin person to chase it. If AI is the productivity story, I want to know where it actually lands in a business like mine.”
— WA drilling contractor, $24M revenue, 90 workers
Australia vs the US on Economic Complexity: The Productivity Gap and Where AI Actually Pays
A rich country that makes narrow things
Harvard's Growth Lab ranks every country on an Economic Complexity Index — a measure of how diverse and sophisticated a nation's exports are. The United States sits near the top at about 14th. Australia sits about 74th, the second-lowest of any OECD member, and Harvard's own verdict is blunt: Australia is “less complex than expected for its income level” and has slipped eight places over the past decade.
A recent methodology recalibration nudged Australia's rank up from about 105th to 74th, but that is an accounting change, not a real shift in what the country can make. The underlying picture is unchanged: Australia is wealthy because it digs up and ships a handful of commodities at good prices, while the US earns its living from a broad base of hard-to-copy goods and services. That is the gap the rest of the numbers flow from.
The scoreboard
Sources: Harvard Growth Lab (Atlas of Economic Complexity), e61 Institute, Reserve Bank of Australia (Statement on Monetary Policy, Aug 2025), US Bureau of Labor Statistics, IMF and OECD. Figures are latest-available and approximate.
Why complexity ends up in the pay packet
Complexity is not an abstraction — it is the chain that runs from what a country makes to what its people earn. Sophisticated industries build deep know-how, which lifts productivity, which lifts real wages and incomes. When complexity stalls, productivity stalls, and living standards follow.
- Australia’s market-sector labour productivity trails the US by about 12%, and the Reserve Bank has cut its trend growth assumption to just 0.7% a year.
- The e61 Institute finds the gap is driven by weak investment in knowledge capital — research, development and software — not a lack of access to the technology itself.
- Real household disposable income per capita has fallen roughly 8–10% from its 2022 peak — the largest decline in the developed world in 2024 — and is barely above where it sat before the pandemic while the OECD rose about 9%.
The leverage is enormous: the Treasury's Intergenerational Report shows that lifting productivity growth by just 0.3 of a percentage point would add roughly 9.5% to GDP by the 2060s. Small, sustained productivity differences compound into very different standards of living.
The one lever nearly everyone agrees on: AI
Australia's Productivity Commission put a number on the opportunity. AI could lift multifactor productivity by more than 2.3% over the decade — worth over $116 billion in extra economic activity, about $4,400 per person, and a boost potentially as large as the internet and mobile phones delivered twenty years ago.
But there is a catch the Commission is explicit about: the dividend only lands if AI is adopted broadly. The risk is a two-speed economy where the big and foreign-owned firms adopt and the long tail of small and mid-sized businesses does not. Poor digital skills are the most-cited barrier. In other words, the national number is really the sum of thousands of individual businesses deciding whether to put AI to work — or not.
What this means for a WA mining supplier
Two things follow for a $10M–$50M drilling, labour-hire, fabrication or mechanical-services business selling into BHP, Rio, FMG, Roy Hill and FQM.
The squeeze is yours too
- The national productivity squeeze is your cost squeeze — rising input costs, thinner margins.
- You cannot lift output by adding admin heads: the office is already the bottleneck, not the field.
The lever is yours too
- The AI dividend lands where repetitive coordination piles up — the back office.
- FIFO mobilisation, MA000011 timesheets, compliance packs, tenders and invoicing are exactly that work.
You do not have to wait for Canberra to reform the economy. The productivity gain the national story is asking for is available in one business at a time, and for a mining supplier it is sitting in the admin. The distinction that matters is whether AI is bolted on as a chatbot beside your coordinator, or whether the coordination itself runs in the system. We wrote about that difference in From Hierarchy to Intelligence.
The national number, made concrete
Recovered admin labour, full 17-agent suite
$555K–$1.1M+/year — recovered admin labour, not a subscription cost
The $116B dividend, in one business
Whether the next site adds admin cost or adds margin — that decision is your slice of the national productivity number
The structural difference
Take the next contract and 50 workers without the next admin hire — same back office, higher output per person
National productivity is the sum of individual businesses. Yours is one of them.
Frequently Asked Questions
What is the Economic Complexity Index?
The Economic Complexity Index (ECI), published by Harvard’s Growth Lab, ranks countries by the diversity and sophistication of what they can competitively export. A country that can make many hard-to-copy products — pharmaceuticals, precision machinery, semiconductors — scores high; one whose exports are concentrated in a few raw commodities scores low. It matters because complexity closely tracks a country’s income today and predicts its growth tomorrow. The US ranks about 14th; Australia about 74th, the second-lowest of any OECD member and lower than its wealth would suggest.
Why is Australia’s productivity growth so low?
Australia’s market-sector labour productivity now trails the US by roughly 12%, and the Reserve Bank has cut its trend assumption to just 0.7% a year. The biggest single driver, per the e61 Institute, is a deficit in knowledge capital — research, development and software — rather than a lack of access to technology. Other factors include a larger low-productivity public sector, harder-to-mine ore bodies, and Australia’s remoteness from major markets. Weak productivity is why real household disposable income per capita has been broadly flat for a decade.
Is AI really worth $116 billion to Australia?
That is the Productivity Commission’s central estimate: AI could lift multifactor productivity by more than 2.3% over the decade, worth over $116 billion in extra economic activity — about $4,400 per person — and a boost potentially as large as the internet and mobile phones delivered. The range is wide and the figure is not guaranteed. Crucially, the gains only materialise if AI is adopted broadly, not just by large and foreign-owned firms. The risk is a two-speed economy where the long tail of small and mid-sized businesses is left behind.
What can a WA mining supplier do about national productivity?
You cannot move the national number, but the $116B AI dividend is made of individual business decisions — and yours is one of them. For a $10M–$50M supplier the productivity lever you actually control is the back office: FIFO mobilisation, MA000011 timesheets, compliance packs, tenders and invoicing. That is where repetitive coordination piles up, and it is exactly where AI adoption pays. Move that work into the system and you take the next site without adding the next admin hire — the productivity gain the national story is asking for, realised in one business.