R&D Tax Incentive: How the Australian Scheme Works

The R&D Tax Incentive (R&DTI) is Australia's main innovation support program, delivered as a tax offset under Division 355 of the Income Tax Assessment Act 1997. Companies register eligible R&D activities with AusIndustry within 10 months of year end (30 April for a 30 June year end), then claim the offset on eligible expenditure in their company tax return.

For companies under $20 million aggregated turnover, the offset is refundable at the company tax rate plus 18.5 percentage points, commonly 43.5%. At or above $20 million, it is non-refundable with a two-tier intensity premium.

What the R&D Tax Incentive is

The R&DTI is a two-agency program with one shared statutory test. AusIndustry, acting for Industry Innovation and Science Australia, decides which activities are eligible R&D. The ATO decides which expenditure attracts the offset and administers the claim.

That split matters in practice. A perfect expenditure calculation attached to a weak activity description is still a weak claim, because the activity description is the part AusIndustry can review, and it can review it years after lodgement.

The program is self-assessed. Nobody approves your activities before you register them (unless you seek a binding Advance Finding from AusIndustry). You assert eligibility, and you carry the evidentiary burden if the assertion is later tested.

Who can claim

You need to be an R&D entity: broadly, a company incorporated under Australian law, and in some cases a foreign company with a permanent establishment in Australia or one that qualifies through a tax treaty. Trusts and sole traders cannot claim in their own right.

Beyond entity type, three gates apply:

Two integrity rules sit underneath all of this. You must be genuinely at risk for the expenditure, and the R&D must be conducted for you, meaning you bear the financial risk and hold the rights to the results.

Work done for a foreign parent, or through associated entities, is an active ATO focus area and should be assessed carefully rather than assumed away.

What counts as R&D under Division 355

Division 355 recognises two kinds of activity: core R&D activities (the experiments themselves) and supporting R&D activities (work that directly relates to them). Eligibility is assessed per activity, never per project or per product. A typical registration contains three to seven core activities.

This is the single most common conceptual error in first claims. "We built a new platform this year" is not an eligibility argument. "We could not determine in advance whether approach X would hold latency under 80 ms at 16 concurrent streams" is the beginning of one.

The four criteria for a core R&D activity

Section 355-25(1) defines core R&D activities as experimental activities whose outcome cannot be known or determined in advance, conducted through a systematic progression of work for the purpose of generating new knowledge. Read as four tests:

1. Experimental in nature. A genuine experiment, not routine development or the application of a known solution.
2. Outcome unknown in advance. Tested against a competent professional with access to knowledge publicly available or reasonably accessible anywhere in the world. New to your company is never enough.
3. Systematic progression. Hypothesis, experiment, observation, evaluation, logical conclusions. Not ad hoc trial and error.
4. Purpose of generating new knowledge, including new or improved materials, products, devices, processes or services.

The worldwide knowledge standard

Both limbs of s 355-25(1) must be satisfied, and the worldwide-knowledge standard is where most claims are won or lost.

In Absolute Vision Technologies [2022] AATA 2319, a decisive finding was the absence of any literature review demonstrating that a knowledge gap existed at all. A prior-art search (literature, patents, internet searches, expert consultation) should predate the activity and be kept as evidence.

Supporting activities and the dominant purpose test

Supporting R&D activities are defined in s 355-30 of the same Act as activities directly related to core R&D activities. For software claims this is usually where most of the hours sit: test harnesses, data pipelines, instrumentation, environment setup.

An extra hurdle applies to two categories. If a supporting activity produces goods or services (or is directly related to producing them), or falls into one of the excluded categories, it must also be undertaken for the dominant purpose of supporting the core activity. Dominant means the prevailing or most influential purpose, not merely one purpose among several.

Getting the split right changes both the narrative and the expenditure apportionment. See core vs supporting R&D activities for the drafting mechanics.

Activities that can never be core R&D

Section 355-25(2) lists categories that can never be a core activity, though some may still be supporting activities if they pass the dominant purpose test:

• Market research, market testing, market development or sales promotion, including consumer surveys.
• Prospecting, exploring or drilling for minerals or petroleum for the purpose of discovering, locating or determining the size or quality of deposits.
• Management studies or efficiency surveys.
• Research in social sciences, arts or humanities.
• Commercial, legal and administrative aspects of patenting, licensing or other activities.
• Activities associated with complying with statutory requirements or standards, including routine testing and analysis.
• Reproducing a commercial product or process from physical examination, plans, specifications or publicly available information.
• Developing, modifying or customising software for the dominant purpose of internal administration by the entity or a connected or affiliate entity.

From income years starting on or after 1 July 2025, activities related to gambling or tobacco cannot be core or supporting activities unless conducted for the sole purpose of harm minimisation.

Boundary cases: compliance work and market research

Two boundary notes are worth holding onto. Being regulated does not trigger the compliance exclusion: work undertaken to demonstrate compliance with a standard is excluded, while genuine experimentation that happens to occur in a regulated setting is not.

The second concerns market research. A structured study measuring a clinical or behavioural endpoint against a pre-stated hypothesis is not market research, even if it runs on A/B infrastructure. The same experiment measured on conversion, churn or revenue is.

Eligible R&D expenditure

Once activities are settled, the ATO side asks which costs are notional R&D deductions. The main categories:

• Salary and wages for time spent on R&D activities, apportioned by actual R&D time, plus on-costs.
• Contractor payments for R&D activities. Payments to associates must be paid in cash by year end, not merely accrued.
• Apportioned overheads directly related to the R&D.
• Decline in value of depreciating assets, to the extent used for R&D.

Several rules reduce or block a claim: feedstock adjustments, recoupment and clawback where a government grant funds the same expenditure, and the requirement for an Overseas Finding before offshore R&D expenditure is claimable. A fuller treatment sits in what R&D expenditure you can claim.

Offset rates

Companies under $20 million aggregated turnover receive a refundable offset at their company tax rate plus 18.5 percentage points; larger companies receive a non-refundable two-tier offset. The old flat 38.5% non-refundable rate ended on 1 July 2021.

The refundable offset is not available to companies controlled by income-tax-exempt entities; they receive the non-refundable offset instead.

How the intensity premium is applied

The intensity premium is applied marginally: the higher rate only touches the slice of notional R&D above 2% of total expenditure. R&D intensity is notional R&D expenditure divided by total expenditure.

Notional R&D deductions above $150 million in an income year attract the offset at the company tax rate only, with no premium. The $20 million turnover threshold is legislated to rise to $50 million from 1 July 2028, which does not affect current income years.

Worked example: refundable versus non-refundable

Refundable: $6M turnover software company

A software company with $6 million aggregated turnover, taxed at 25%, incurs $800,000 of notional R&D deductions. Its offset is 43.5% of $800,000, which is $348,000.

Because the offset is refundable, any amount exceeding tax payable is paid as cash. A pre-revenue company with no tax liability receives the full $348,000 as a refund.

Non-refundable: $30M turnover, 10% intensity

A company with $30 million turnover, taxed at 25%, has total expenditure of $25 million and notional R&D of $2.5 million. R&D intensity is 10%.

• 2% of total expenditure = $500,000, offset at 25% + 8.5 = 33.5%, giving $167,500.
• Remaining $2,000,000, offset at 25% + 16.5 = 41.5%, giving $830,000.
• Total offset: $997,500, applied against tax payable, with any excess carried forward.

Both examples clear the $20,000 minimum comfortably. If your notional R&D deductions come to $15,000, there is no claim unless the spend went to a Registered Research Service Provider or a CRC contribution. Our rebate calculation guide works through the apportionment steps in more detail.

Key dates

Registration is due 10 months after the end of the income year. For a 30 June year end, that is 30 April.

The deadline is statutory, and registration must precede claiming the offset in your tax return. If you have lodged the return already, you register and then amend.

Overseas Finding applications are different: they are due by the end of the income year in which the overseas activity starts, so they cannot be handled retrospectively. See the registration deadline guide.

How to claim, step by step

1. Identify the experiments. Work backwards from technical or clinical uncertainties you could not resolve from existing knowledge, not from project budgets.
2. Document the knowledge gap before you start: literature, patents, competitor documentation, expert input, recorded as you go.
3. Frame each core activity as a hypothesis with a measurable, pre-stated threshold that can be confirmed or disproven.
4. Track R&D time against activities as work happens, not in a reconstruction exercise in March.
5. Register with AusIndustry through the customer portal within 10 months of year end.
6. Calculate notional R&D deductions and apply the relevant offset rate.
7. Lodge the R&D tax schedule with your company tax return through your tax agent.
8. Retain the evidence for at least five years after lodgement, and longer if a review, objection or amendment period is still open.

What the registration form asks for

The registration form itself is a field-by-field questionnaire, structured as Projects, then Core activities, then Supporting activities. A core activity is not one narrative: it is split across separate fields (sources investigated, why a competent professional could not know the outcome, hypothesis, experiment, evaluation, conclusions, new knowledge), each with its own character minimum.

Drafting each field to its own requirement is a different exercise from writing a project summary. The registration form field guide covers each one.

Evidence and record keeping

The expected standard is contemporaneous evidence: records created as the work happened. For technical claims that means commit history, pull requests, issue trackers, design documents, test results, experiment logs and time records.

For clinical claims it means protocols, HREC approvals, statistical analysis plans with pre-registered endpoints, de-identified outcome data and literature review notes. For a full breakdown, see what evidence AusIndustry accepts.

What contemporaneous means in practice

Retrospective documentation carries little weight. In Absolute Vision Technologies [2022] AATA 2319, documents prepared after the fact for an internal review were not accepted as contemporaneous, and there was marked discord between the records and the activities as registered.

In Royal Wins Pty Ltd [2020] AATA 4320, the absence of contemporaneous documentation that a hypothesis was developed and tested defeated the claim regardless of how much work had been done.

This is the problem Rand was built for: it assembles the registration from engineering evidence you already generate in GitHub and Jira, so the record predates the claim rather than following it.

Common reasons claims fail review

Claims fail for the same handful of reasons across every sector.

• A commercial goal dressed as a hypothesis. In GQHC [2024] AATA 409, vague or "clumsy" commercial hypotheses failed. A hypothesis must be scientific, specific and capable of being invalidated.
• No documented worldwide knowledge gap. Novelty to the company is not novelty to the field.
• Missing observation and evaluation. Results must be analysed against pre-stated thresholds, not asserted.
• Trial and error without isolable variables. In Active Sports Management [2023], the contemporaneous formation of a hypothesis was held to be the essential starting point of the activities.
• Records that do not match the registration.

Note also that applying a known method in a new context can be eligible: Moreton Resources Ltd v ISA [2019] FCAFC 120 construed "experimental activities" broadly. But Coal of Queensland [2021] FCAFC 54 is the counterweight: without a genuine unknown outcome and a documented systematic progression, the four-part test still fails.

Where software and health-tech claims sit

Software claims

Software is domain-neutral under Division 355, and most software development is not R&D. Configuring documented options, integrating third-party APIs from vendor docs, building dashboards from established patterns, migrations, refactoring and version upgrades are routine.

The eligible core is usually a narrow algorithmic or systems question with a measurable target no published result demonstrates, with product engineering as supporting activities. See the software eligibility criteria.

Health-tech and clinical claims

Health-tech claims often work differently, because the uncertainty may not be technical at all. Whether a six-week app-delivered intervention measurably improves an IBS symptom score is a genuine knowledge gap even when the app itself is routine to build.

A 2022 determination deems certain clinical trials to be core activities: phase 0 to III trials, and pre-market pilot and pivotal stage trials, of a therapeutic good not yet on the ARTG. Phase IV post-market trials, trials of generic products, and any market research, market testing or sales promotion component are not covered.

Wellness apps outside TGA regulation cannot use that shortcut and must meet the ordinary four-part test. Clinical trial eligibility unpacks the distinction.

Where to go next

The R&DTI rewards precision: per-activity assessment, a documented worldwide knowledge gap, hypotheses with thresholds, and records made at the time. Get those four right and the offset calculation is arithmetic. For a fuller process view, read the end-to-end R&D Tax Incentive claim walkthrough.

Rand builds the AusIndustry registration from the engineering and clinical evidence your team already produces, with Expert Reviewers approving it before lodgement.

This article explains the rules; it is not tax advice, and eligibility always depends on your specific activities and circumstances.