The Research and Development Tax Incentive (RDTI) is the single biggest government program supporting startups in Australia. The program as a whole accounts for around $3 billion in Federal Government expenditure each year, of which about two-thirds is spent on companies with less than $20m in annual turnover. Support for startups under the RDTI dwarfs expenditure on any other startup program. The RDTI is open to any business undertaking eligible research and development (R&D). For most startups developing new products is at the core of the business. Since its introduction in 2011 the RDTI has been a key reason for startups to employ Australians to undertake R&D, and a major driver for companies to keep R&D efforts based in Australia even when the business expands overseas.

Taking a lead from companies like Atlassian and 99designs, an increasing number of young Australian tech companies are choosing to ‘build local, sell global’ by keeping R&D teams in Australia - even when the business has shifted the other parts of its operations, or even its headquarters, overseas to meet the expectations of customers or investors. The RDTI is a major reason why this model is attractive. Federal political leaders have, for some time, been rightly proud of the program’s ability to support the rapid growth of new, high growth companies creating high value jobs in cutting edge technology fields.

There is a serious risk that a depleted RDTI could begin to cripple Australia’s burgeoning startup sector at a critical time.


The R&D Tax Incentive is set up in such a way that, despite its importance to startups, it has always been awkward for software firms to meet the requirements. Under the scheme, eligible R&D must follow a strict and extensively-documented scientific model (called the ‘Frascati model’) of hypothesis, experimentation, observation and evaluation, and logical conclusions. The results of the experiment must not be able to be predetermined, and the process must be carried out with the intention of generating ‘new knowledge’ (though it’s worth noting that the legislation specifically identifies developing new or improved products or services as ‘new knowledge’).

This process is highly suitable for lab experiments and scientific research (the ‘R’ in ‘R&D’). It is not as readily applicable to software development or other kinds of technology commercialisation. Nevertheless, the current iteration of the RDTI was designed explicitly to include most software development as eligible R&D. When introducing the current legislation in 2010, the Minister said:

Recognising the pervasive nature of information technology in a modern economy, the new R&D Tax Incentive will ensure most software R&D is treated consistently with R&D occurring in other sectors.

As a result, it became established industry practice for software companies to claim activities related to the development of new, innovative software products under the scheme. That practice was encouraged by specialist R&D tax advisers from reputable global accounting firms (alongside a growing number of boutique R&D tax firms). Politicians spoke openly about the positive impact the scheme was having helping exciting young technology companies grow. Implementing agencies also clearly supported the idea that software development was a core part of the scheme - in 2012, the year after the scheme was introduced, AusIndustry released formal advice to the ICT sector which said:

As a key enabler of business and innovation, ICT R&D is central to services, processes and products across virtually all sectors of the economy. Based on experiences gained from the R&D Tax Concession, AusIndustry anticipates ICT involvement in the new R&D Tax Incentive will be high.

In line with this advice, claims made by software companies were paid consistently for many years. The OECD updated the Frascati Manual in 2015 to explicitly acknowledge the validity of software development as a form of R&D.


The R&D Tax Incentive has been the focus of some concern over the last few years. The cost of the scheme has been rising, leaving policy makers worried that it could spiral out of control. The biggest growth in cost has been among small companies (less than $20m in turnover), which includes almost all startups. Partly in response to this concern, the Turnbull Government commissioned a review of the program, which submitted its findings to government in April 2016. The review made six recommendations to help strengthen the scheme’s ability to encourage ‘additionality’ (R&D that wouldn’t have been undertaken if the scheme didn’t exist) and to reduce cost by increasing integrity.

The Turnbull Government did not formally respond to the 2016 review until more than two years later, when it introduced changes to the R&D Tax Incentive as part of the Federal Budget in May 2018.


In the meantime, concerns about the cost of the scheme continued, and the program came under pressure. In February 2017 the ATO and AusIndustry issued a Taxpayer Alert seeking to clarify the position on the eligibility of software claims under the RDTI. The Alert noted that the ATO and AusIndustry were ‘reviewing the arrangements’ for software companies claiming the R&D Tax Incentive. It referenced concerns held by the ATO and AusIndustry that software companies were not conforming to the ‘stringent requirements of the laws that govern the R&D Tax Incentive’.

Whether linked to this Alert or not, over the last two years the RDTI has been slipping away from startups. Companies who had been claiming, on sound advice from seasoned R&D Tax Incentive advisers, similar activities for multiple years in a row have been told their software development activities aren’t eligible. Audits are up, and software audits now come back negative more than 50% of the time.

Perhaps most significantly at a system level, companies have begun to see R&D claims as risky, and started to self-censor. A common experience right across the sector is that R&D claims for genuine software development are being pared back substantially to try to reduce the risk of facing a potentially catastrophic clawback. Investors, too, have begun to see RDTI as a risk - software companies with significant historical RDTI payments are seen as exposed.

This has been compounded by an overall reduction in the scheme for all claimants. In September 2016 entitlements under the scheme were reduced by 1.5%, to 43.5% for refundable claimants and 38.5% for offset-eligible claimants.


All of this amounts to a significant cut to the R&D Tax Incentive. In fact, we know the quantum: This comes at a time when Australia’s total spend on r&d is already well below the OECD average at around 1.87% of GDP (OECD average around 2.3%) and is falling. Australian businesses spent $16.7 billion on R&D in 2015–16 compared to $18.9 billion in 2013–14, a decrease of 12%.


There is a serious risk that a depleted RDTI could begin to cripple Australia’s burgeoning startup sector at a critical time. The program has become a vital piece of infrastructure in the startup landscape. Reduced access for startups is likely to result in:

  • A reduction in working capital for high growth technology startups
  • Lower growth rates for affected companies
  • Lower hiring rates for software development firms
  • Decreased R&D output as support diminishes
  • Failure of some companies, including particularly R&D intensive businesses counting on a reliable source of R&D support
  • A reduction in Australia-based R&D by global companies
  • A contraction in the number and quality of Australia’s new technology businesses


In the immediate term, there needs to be some reassurance given to vulnerable software companies. Firms with turnover of less than $20m that have been claiming the incentive in good faith and on credible professional advice need to be confident that they are not going to be subject to audit processes unless their claims are manifestly unreasonable or have had sharp unfounded increases. A moratorium on reviews within these parameters should remain in place until the introduction of a clear legislative fix to the way the R&D Tax Incentive operates or a new scheme that directly supports software development is implemented. This would help address uncertainty and reduce existential risk for good-faith claimants.

Recent anecdotal evidence suggests the program’s administrators may have begun adopting a less hard-line approach to software audits, giving some companies a warning and advising against future claims rather than moving straight to recoup funds. Such an approach would accomplish some of the aims of a moratorium, although publicly formalising this position would help with certainty and confidence.

Reducing clawbacks helps reduce the immediate harm caused by software falling out of the RDTI, but it doesn’t solve the larger problem: the outflow of support and capital from an area of activity vital for Australia’s long term economic prosperity.

To address this, a clear path to supporting technology companies doing genuine software development work in Australia should be formulated. While this goal might be able to be achieved by adjusting the language of the legislation, it is becoming increasingly apparent that a separate program, focused specifically on software and driven by job creation as a primary metric, is needed. Such a scheme would help limit the scope and cost of the support provided, and would support growing tech businesses based on the additive economic value of their development activities.