Proposed changes to Australia’s R&D tax system would expand eligibility beyond genuine research, concentrate benefits among a narrow group of firms, and proceed without clear costings.

The centrepiece proposals in the Strategic Examination of Research and Development (SERD) report, _Ambitious Australia_, would redefine what counts as Research and Development (R&D), leave the fiscal cost unexamined, and concentrate benefits on a very small proportion of Australian innovative businesses.

SERD has delivered a thorough diagnosis of what ails Australia’s R&D system. Chapters three and four prescribe a treatment built around a substantially expanded R&D Tax Incentive (RDTI) and a new venture capital architecture. The diagnosis is sound, but the prescription leaves three questions unanswered.

Since 2011, the RDTI has rested on a clear principle: eligible expenditure must involve genuine technical uncertainty that can only be resolved through systematic investigation based on established science. This definition follows the OECD’s Frascati Manual and is consistent with R&D tax incentive design across every comparable economy. It is the primary integrity mechanism in the program.

The SERD’s premium startup stream (Recommendation 5b) would extend eligibility to development and deployment, early commercialisation, and user testing. These activities occur after the core technical uncertainty has been resolved. The remaining uncertainty is commercial, not technical.

In a typical technology startup, the experimental phase may account for 15 to 25 per cent of total expenditure. Expanding eligibility to product development and commercialisation opens the offset to the remaining 75 to 85 per cent. No OECD country subsidises deployment and early commercialisation through its R&D tax incentive. Countries that support these activities do so through separate instruments with their own eligibility criteria and fiscal controls.

The SERD presents the change as an “implementation pathway,” when it is, in substance, a conversion of the startup stream from an R&D incentive into an early stage business subsidy delivered through the tax system. The report should have acknowledged this significant departure from a long held OECD benchmark and openly debated the implications.

Once development and commercialisation become eligible in the startup stream, the pressure to extend the same treatment to SMEs and large firms will be immediate. The expanded expenditure definition would also create strong incentives for the tax advisory industry to leverage the vaguer boundary to construct aggressive claims, increasing both the cost to the budget and the compliance burden on the ATO.

The minister’s Innovation and Technology Roundtable at CSIRO Eveleigh in August 2025 set a clear standard: reform opportunities should be “budget positive or neutral.” He repeated the expectation at the National Press Club on 17 March.

The SERD report includes several potentially costly tax expenditures: higher offset rates, expanded refundability thresholds, removal of the $150 million expenditure cap, a production tax credit, and expanded ESIC and ESVCLP concessions. None has been costed.

A tax expenditure is revenue forgone through concessional tax treatment. Unlike direct spending, tax expenditures do not appear as line items in the May budget. They are reported in Treasury’s annual Tax Expenditures and Insights Statement, typically released the following January, and receive a fraction of the scrutiny that budget night generates. This means expansions of the kind proposed in the SERD can proceed without triggering the immediate fiscal accountability that would accompany equivalent direct spending.

In public finance terms, a dollar of revenue forgone through an RDTI offset has exactly the same effect on the fiscal balance as a dollar spent on a research grant. Without forward estimates, it is impossible to judge whether the proposed benefits justify the commitment of public revenue. On any reasonable calculation, the cumulative cost of the SERD tax expenditure recommendations would be considerable and well above fiscal neutrality.

The SERD panel, supported by a secretariat and presumably with Treasury consultation, should have been able to produce forward costings. Their absence is possibly the single greatest weakness of the report.

The premium startup stream’s eligibility is determined by a 100 point test that privileges venture capital backing, accelerator participation, IP holdings, and university collaboration. Venture capital investment in Australia is less than 0.2 per cent of GDP. Within that small pool, NSW and Victoria account for roughly three quarters of funding, concentrated in AI, fintech, and biotech.

Under the expanded expenditure definition, a VC backed software startup spending $5 million a year on adoption and user research could plausibly claim $3 million to $4 million in eligible expenditure. At the proposed premium offset rate of 48.5 per cent, this would generate a refundable cash offset of approximately $1.5 million to $1.9 million from the Commonwealth.

Meanwhile, the 46 per cent of Australian businesses that are innovation active (ABS, 2024) but do not conduct formal R&D receive nothing from any element of the RDTI, current or reformed. The ecosystem also continues to face a significant growth gap in late stage capital for physical technology such as robotics, renewables, and medical devices, which requires longer term, capital intensive scale up support that the SERD’s premium stream, anchored to venture capital eligibility criteria, does not address.

These questions address instrument design, fiscal transparency, and distributional equity. Together they point to a conclusion: the expanded RDTI and associated venture capital architecture would have required more work, or perhaps a different design premise altogether.

An alternative would be to leave the RDTI to do what it does well, reduce the cost of genuine experimental research, and create a separate commercialisation instrument for the transition from lab to market. Such an instrument should be open to firms regardless of their investor profile and funded through a capped direct appropriation at $300 million to $500 million per year.

The simpler RDTI reforms proposed in the SERD, including deemed rates, reduced documentation, and quarterly cash advances, could proceed within the existing legislative framework, subject to budgetary conditions. They would make a tangible difference to firms already in the system.

The expanded expenditure definition should not proceed without costings, distributional analysis, and the public debate that a commitment of this scale demands. All of this at a time when Commonwealth departments are operating under fiscal austerity and the Cabinet Expenditure Review Committee is looking for savings, not new open ended claims on revenue.