Research and Development Tax Incentive: How It Works

The R&D Tax Incentive (R&DTI) is Australia's main support program for company-conducted research and development, delivered as a tax offset under Division 355 of the ITAA 1997. Companies under $20M aggregated turnover receive a refundable offset of the company tax rate plus 18.5 percentage points, commonly 43.5%.

Companies at $20M or more receive a non-refundable offset with a two-tier intensity premium. Registration with AusIndustry is due 10 months after year end (30 April for a 30 June year end) and must happen before the offset is claimed.

What the R&D Tax Incentive is

The R&DTI is a tax offset, not a grant. You spend on eligible R&D activities during the income year, register those activities with AusIndustry, then claim a notional deduction and the corresponding offset through your company tax return.

It is entitlement-based. There is no competitive assessment, no funding round, and no cap on the number of claimants. If your activities and expenditure meet the statutory tests, you are entitled to the offset.

Eligibility is self-assessed. Neither AusIndustry nor the ATO approves your claim in advance (outside the separate Advance Finding process), which means the quality of your registration and your evidence is what carries the claim through any later review.

Who administers it: AusIndustry and the ATO

The R&D Tax Incentive is jointly administered by AusIndustry, which assesses whether your activities are eligible R&D, and the ATO, which assesses your expenditure and the offset amount. Confusing the two is a common source of trouble, because a well-documented cost schedule does not rescue a weak activity description, and vice versa.

In practice, an AusIndustry review reads what you wrote on the registration form and asks for the records behind it: the prior-art search, the hypothesis, the experimental method, the results and how they were evaluated. It is a scientific and technical assessment of the activity as registered.

An ATO review starts from the R&D tax schedule and works backwards to the ledger. It asks whether the amounts claimed relate to the registered activities, whether apportionment is supported by records made at the time, and whether integrity rules such as at-risk, feedstock or recoupment reduce the claim.

External merits review of an eligibility decision goes to the Administrative Review Tribunal (ART), which replaced the Administrative Appeals Tribunal in October 2024.

Who can claim

You must be an R&D entity: broadly, a company incorporated in Australia, or a foreign company that is an Australian resident for tax purposes or operates through a permanent establishment under a tax treaty. Trusts (with the narrow exception of a body corporate acting as trustee of a public trading trust), partnerships and sole traders cannot claim.

Beyond incorporation, two conditions trip up claimants.

• At risk and conducted for. You must bear the financial risk of the R&D and hold the rights to exploit the results. Work done for a foreign parent or under a contract where the customer owns the outcomes usually fails this.
• Refundable offset restriction. The refundable offset is not available to companies controlled by income-tax-exempt entities.

What the offset is worth

The offset rate is a function of your aggregated turnover, your company tax rate, and (for larger claimants) your R&D intensity. The $20M aggregated turnover threshold is the dividing line.

Refundable offset: aggregated turnover under $20M

The rate is your company tax rate plus 18.5 percentage points. For a base-rate entity taxed at 25%, that is 43.5%.

Refundable means the offset is paid as a cash refund to the extent it exceeds tax payable, which is why loss-making startups receive cash rather than a carried-forward benefit.

Worked example: a pre-revenue company with $400,000 of notional R&D deductions and no tax payable receives $174,000 as a cash refund ($400,000 × 43.5%).

Non-refundable offset: aggregated turnover $20M or more

The rate is your company tax rate plus a two-tier R&D intensity premium, applied marginally:

R&D intensity is notional R&D expenditure divided by total expenditure for the year. There is no flat 38.5% non-refundable rate: that ended for income years commencing on or after 1 July 2021.

Worked example: a company with $50M total expenditure, $5M of notional R&D deductions and a 30% tax rate has 10% intensity. The first $1M (2% of $50M) attracts 38.5%, giving $385,000. The remaining $4M attracts 46.5%, giving $1,860,000. Total offset: $2,245,000. Unused non-refundable offset is carried forward.

Cap and minimum spend

The $20M refundable threshold is legislated to rise to $50M from 1 July 2028. It does not affect current income years. For a full calculation walkthrough, see how to calculate your R&D tax incentive rebate.

Core R&D activities and the four tests

A core R&D activity is the unit you register, and eligibility is assessed per activity, never per project or per product.

Section 355-25(1) defines core R&D activities through two limbs.

The first limb is that the outcome cannot be known or determined in advance on the basis of current knowledge, information or experience. It can only be determined by applying a systematic progression of work based on principles of established science, proceeding from hypothesis to experiment, observation and evaluation, and leading to logical conclusions.

The second limb requires that the activities are conducted for the purpose of generating new knowledge, including new knowledge in the form of new or improved materials, products, devices, processes or services.

Both limbs must be satisfied. That resolves into four tests:

1. Experimental in nature. An experiment or set of related experiments, not routine development or applying a known solution.
2. Outcome unknowable in advance. Tested against a competent professional with access to knowledge publicly available or reasonably accessible anywhere in the world.
3. Systematic progression. Hypothesis, experiment, observation, evaluation, conclusions. Not ad hoc trial and error.
4. Purpose of generating new knowledge. New or improved materials, products, devices, processes or services, beyond what is publicly available.

A typical registration contains three to seven core activities. Over-grouping is a recognised drafting failure: combining several distinct technical hurdles into one activity obscures the experiment and can cost you eligibility.

The competent professional standard

The benchmark is a competent professional in the relevant field, someone qualified, current, and with access to worldwide journals, patents, the internet and other professionals. "New to our company" is never sufficient.

Match the professional to the uncertainty: an ML engineer for a model architecture question, a clinician or health researcher for an efficacy or adherence question.

Information that is not reasonably accessible does not count. A competitor's trade secret is not public knowledge, so a genuine gap can exist even if someone has quietly solved the same problem.

Prior art must predate the activity

AusIndustry expects contemporaneous evidence of a worldwide search conducted before the activity started: literature review, internet searches, patent searches, expert consultation. In Absolute Vision Technologies [2022] AATA 2319, a decisive finding was that there was no evidence of a literature review to confirm any knowledge gap existed.

Supporting R&D activities

Supporting activities are activities directly related to core R&D activities, under s 355-30 ITAA 1997. For software claimants this is usually where most of the hours sit: test harnesses, data pipelines, instrumentation, environment setup, and the surrounding product engineering that the experiment depends on.

A higher bar applies in two situations: where a supporting activity produces goods or services, and where it falls into one of the excluded categories. In either case it must be undertaken for the dominant purpose of supporting the core activity, not merely be related to it.

The split matters for both eligibility and expenditure apportionment. See core vs supporting R&D activities for how to draw the line.

What can never be a core activity

Section 355-25(2) lists categories that can never be core activities. Some may still be supporting activities, but only under 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 or locating deposits, or determining their size or quality
• 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 by physical examination, or from plans, specifications or publicly available information
• Developing, modifying or customising software for the dominant purpose of the entity's own internal administration, including internal administration of its business functions

For 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 that are not excluded

Two boundaries trip up claimants who read the exclusions too broadly. Being regulated does not trigger the compliance exclusion: work done to show compliance with a standard is excluded, while genuine experimentation that happens to be regulated is not.

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

Which expenditure you can claim

The ATO side of the program covers notional R&D deductions. The main categories are:

• Salary and wages for time spent on registered R&D activities, apportioned by R&D time, plus on-costs
• Contractor payments for R&D activities, and associate payments (which 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

Apportionment is where most claims are made or lost. The percentage of an engineer's year spent on core and supporting activities needs a basis in records made at the time, not an estimate written in April. What R&D expenditure you can claim covers each category in detail.

What reduces or blocks a claim

Several integrity rules cut across the offset calculation:

• At-risk rule. Amounts you are not genuinely at risk on, because a third party will reimburse them, are excluded.
• Conducted for. R&D conducted for a foreign parent or another entity that holds the results generally cannot be claimed by the Australian company. The ATO has flagged these arrangements in Taxpayer Alerts TA 2023/4 and TA 2023/5.
• Feedstock adjustments. Where R&D activities produce marketable output, feedstock rules claw back part of the benefit.
• Recoupment and clawback. Government grants received in relation to the same expenditure trigger a clawback.
• Overseas activities. Offshore R&D requires an Overseas Finding from AusIndustry, applied for by the end of the income year in which the overseas activity starts.

How to register, end to end

1. Identify candidate activities during the year, at the level of a discrete technical or scientific uncertainty.
2. Run and document the prior-art search before the experiment begins.
3. Record the hypothesis, experiments, observations and evaluations as they happen.
4. After year end, split the work into core and supporting activities and draft the registration.
5. Submit the registration through the AusIndustry customer portal (linked from the business.gov.au R&D Tax Incentive page) within 10 months of year end, and receive a registration number.
6. Calculate notional R&D deductions and lodge the R&D tax schedule with the company tax return through your tax agent.
7. Retain the underlying records for at least five years.

The registration form

The registration form (current version released 15 August 2025) 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 for sources investigated, why a competent professional could not know the outcome, the hypothesis, the experiment, the evaluation, the conclusions, and the new knowledge generated. Each has its own character minimum and should be drafted to its own requirement. See what AusIndustry asks on the registration form.

Rand builds this registration from the engineering evidence you already generate in GitHub and Jira, maps it to those fields, and puts it in front of Expert Reviewers before it goes to AusIndustry.

Key dates

The registration deadline is statutory and hard. Registration must precede claiming the offset, so a missed registration deadline is a missed claim for that year. Full detail on every other date sits in R&D Tax Incentive deadlines.

Evidence expectations

Contemporaneous documentation is not administrative housekeeping. It is part of the systematic progression the statute requires, and the case law is consistent on this.

What the tribunals expect

• GQHC [2024] AATA 409: hypotheses must be scientific, specific, and capable of being validated or invalidated, not commercial aims. Observation and evaluation must be evidenced, not asserted.
• Royal Wins Pty Ltd [2020] AATA 4320: no contemporaneous documentation that a hypothesis was developed and tested means the activities are not eligible, regardless of how much work was done.
• Active Sports Management [2023]: the contemporaneous formation of a hypothesis is the essential starting point. "We tried a few things until it worked" is not a repeatable, verifiable experiment.
• Coal of Queensland [2021] FCAFC 54: applying known methods without a genuine unknown outcome fails, even after Moreton Resources [2019] FCAFC 120 established that novelty of context can support eligibility.

A minimum viable record set

Practically, that means keeping design documents and literature review notes with dates, experiment logs and test results, code repositories and issue trackers, time records that support apportionment, and (for clinical streams) protocols, HREC approvals, statistical analysis plans with pre-registered endpoints, and de-identified outcome data.

For a first-year claimant, the starting set is narrower than it sounds. Date-stamped notes of the prior-art search, a written hypothesis with a measurable threshold recorded before the work started, the raw results of each iteration, and a short evaluation against the threshold gives most claims a defensible evidentiary basis, though the outcome always depends on the specific activities and records.

Add a time-allocation record per person per month that ties hours to named activities, and keep everything in systems that show when it was created. What evidence AusIndustry accepts sets out the full hierarchy.

Where to go next

If you are ready to run a claim end to end, the complete step-by-step guide turns this orientation into a working process.

Software claimants should read software R&D eligibility. Health-tech companies running efficacy or adherence studies should start with clinical trial eligibility, because the uncertainty in a health app often lives in the clinical outcome rather than the code.

This article explains the rules; it is not tax advice, and eligibility always depends on your specific activities and records. If you want the registration built from evidence rather than recollection, that is the problem Rand was designed to solve.