Despite a modest increase in business R&D expenditure in Australia in the latest reported year (2023/24), Australia’s performance on a GDP basis is still woefully worse than its international economic competitors.

But at least business R&D is still being measured every two years by the ABS. There is no reliable, routine national data on business and industry spend on skills and training.

The last ABS survey “Employer Training Expenditure and Practices, Australia, 2001-02” was released in 2003 – two decades ago. These data (e.g. training purpose, industry-specific sectors, total expenditure; spend per employee; training as % of gross wages/salaries etc) are routinely surveyed by business/industry and by governments in major competing economies e.g. US, UK, EU.

In Australia, there have been a few limited ad hoc surveys (e.g. by CEDA, by Deloitte/RMIT, and others cited in the recent Jobs and Skills Australia report (Our Gen AI Transition, Implications for Work and Skills). The CEDA report (Learning Curve – Why Australia Needs a Training Boost) headlines the proportion of people undertaking work-related training has declined 14% since 2007.

Limited qualitative evidence indicates better resourced major corporates, and tech-savvy businesses, are more nimble tech adopters and adapters (if not employers, then self-motivated employees). Businesses that do invest in skills, digital or otherwise, may use “in-house” formal accredited training. But the majority use informal and non-structured training, being disinclined to engage with tertiary education providers. The numbers of enterprise owned training providers have also declined.

There is no direct account of what kind of “on job” skills are learnt “in house” e.g. technical, AI or ‘enabling’ skills, nor how much businesses invest. The relevant Australian Accounting Standard (AASB 138) requires no distinction between types of training expenditure, whether accredited, non-accredited, in-company, or otherwise. It treats all training-related costs uniformly as expenses.

Worse, there is no national data on industry spend on training with tertiary institutions – it doesn’t exist with either universities or TAFEs. Businesses invest in “earn while you learn” mode, apprenticeships and traineeships, but willingness to do this is mainly dependant on government wage subsidies. The numbers commencing traineeships are especially vulnerable to cutbacks in government subsidies.

Regarding “work integrated learning”, the Australian Government has commenced new “practical payments” addressing “placement poverty” for nursing, care and teaching professions. The Higher Education Accord Review suggested financial support for unpaid work placements be funded by employers for all other fields – which is improbable dreaming.

Commendably, in both the higher education and VET sectors, multiple professional bodies invest resources in informing course accreditation in different professions, e.g. in medicine, dentistry, health, nursing, engineering, veterinary, teaching, accounting, law etc. But that’s not tracked either.

We measure productivity routinely, publish job ads in thousands, and track house prices incessantly – yet don’t measure or incentivise how businesses self-invest in skills. But business is big on skills advocacy.

This advocacy invariably extols governments to “better align training with skill needs”, that is, focus public funding/financing of post-school education and training courses on employers’ “on-job” skills needs. This has reinforced the politics of reframing university purpose and funding from being providers of quality education and research to deliverers of “job-ready quals”, TAFE-like plus research.

Such advocacy is based on an implicit and arguable bargain. Governments’ key task is to fund and finance education, training and skills, and businesses provide jobs and (for profits) pay tax. The caustic view is that some businesses squeal about skills shortages, plan poorly, demand governments invest more in skills, leading to greater reliance, and growing controversy, on migrant labour.

All this has dwarfed any evidence of workforce benefits of those employers that make direct investment or co-invest in training. But without routine evidence of business investment in skills and with governments under fiscal stress, this line of advocacy risks being increasingly impotent, despite imperatives for innovation, workplace upskilling and life-long learning.

It is also increasingly risky and irresponsible for corporates in an era where digital is the key economic driver and where AI, quantum and cyber security are potential disrupters of future productivity, innovation and national security in the 21st century.

Australia has a patchy history of skills incentives and levies. The then government in its 2022/23 budget announced a “technology investment” and a “skills and training boost for small businesses”. This was akin to the R&D Tax incentive. The training boost allowed businesses with an aggregate annual turnover of less than $50 million an additional 20% tax deduction for external training courses delivered to employees by registered training providers. It lasted two years, with no evidence of “boost” outcomes.

Australia had a Training Guarantee Act and Levy from 1990-94. It required all employers to spend a minimum percentage of payroll per year (1.5% in 1992) as a payroll threshold or risk a “shortfall” payment to a Training Guarantee Fund. The scheme had mixed effectiveness and was problematic, especially for SMEs.

Australia is leery of “hypothecation” schemes, such as the TG Levy or the UK Apprenticeship Levy. There are, however, positive examples, like the producer levy on agriproducts, matched by government, funding the long running rural Research and Development Corporations. There are construction project levies in some states that create a funding pool for construction industry training for new industry entrants and safety etc. Governments may also include minimum training or “earn-while-you-learn” obligations in its mega procurements e.g. within big tech contracts.

Limited evidence indicates that some Australian businesses invest significant resources in skills and training (including recently also greater R&D on AI) but this is overwhelmingly only in large corporates with resources and the competitive need to invest. SMEs face resource and time barriers, and non-profits, despite employing thousands, have scant capacity.

The solution is never more detailed skills planning and lists of shortages. It requires greater entity self-investment in structured training and not just exhorting fiscally constrained governments. The issue isn’t simply about underfunding; it’s systemic bad design. Australia has no coherent fiscal mechanism that binds economic growth to the “on-job” development of its human capital.

Coherence here goes well beyond individual SMEs getting a one-off government “skills” subsidy. It’s far more ambitious with e.g. AI/quantum/digital skilling across co-operating entities, ideally companies of fast growth-potential in localised innovation districts and precincts. This builds potential for bulking up both individual and aggregate absorptive capacity via local training networks, shared intellectual and physical infrastructure, and cost shared upskilling. When well executed, this builds performance and reputation for inbound investment by multinationals who expect local skills to be at the global leading edge.

This is the expanded horizon needed in a national integrating architecture that unites the efforts of employers, educators and government. There is no such imaginative government-initiated skilling incentive. When designed, it needs to avoid subsidising what good businesses already do, but subsidise networks of best upskilling practice, fiscally encouraging companies to participate.

The present malaise runs counter to the prime minister’s economic outlook speech (July 2025), expressing the government was “determined to create the right framework and make the right investments in skills and training to ensure artificial intelligence is a contributor, not a competitor”.

Maybe business feels the formal education and training system may not be delivering fast enough, nor in a manner that meets the on-job urgency of business and employee uplift in AI/quantum, cyber and digital skills. But it needs to be recognised that this is not like the COVID disruption where international bioscience ultimately delivered a universal vaccine salvation. Complacency doesn’t AI upskill.

Republished from Action Institute for Policy Research and Innovation, 16 September 2025

The views expressed in this article may or may not reflect those of Pearls and Irritations.