Executive Summary (pg.4)
My office is recommending that this important incentive be retained, and a suite of reforms are made to the way the system is administered.
This report details how small and family business taxpayers have been subjected to review, examination and audit by the 2 agencies responsible for the delivery of the program. In all cases, this compliance activity was retrospective and commenced several years after the relevant research and development (R&D) activity was undertaken and the R&DTI refund received. In almost all cases, the R&DTI claims were rejected in total. This has had a devastating impact on the companies, with some saying they face financial ruin. Others have scaled down their R&D efforts in Australia and reduced their R&D staff.
It is therefore critical that wherever possible the program is improved for small and family businesses and that R&DTI risk and compliance activities are conducted as close as possible to when “they” register their R&D activities and before they claim the benefit with the ATO.
We identified an overall shift in the way the R&DTI legislation has been interpreted over the last three to four years; a narrowing of focus leading to a rejection of claims, which in previous years has been regarded as low risk.
There appears to have been a broad-brush approach to program integrity with a view to recouping Government expenditure on the R&DTI. Valid claimants have been swept up along with those who have been badly advised by unscrupulous R&D “consultants”
The ATO and AusIndustry tell us that they are revising their approach to R&DTI compliance, which we welcome. It is also critical that there is a mechanism by which those businesses adversely affected by past poor processes can seek redress.
We have grave concerns that the administration of the program does not provide sufficient certainty for small business and that compliance activity by the regulators can and does negatively impact the continued viability of small businesses. The potentially devastating impact on the business is primarily due to audit activity coming years after an incentive has been invested back into business – with many having relied on expert professional advice. Small businesses simply do not have the cash-flow or retained earnings to repay the R&DTI together with accumulated interest and penalties.
The program requires a fairer, more consistent, educative and customer-focused approach by both AusIndustry and the ATO embedded consistently throughout both networks. Our recommendations are based on the importance of the characteristics of good governance including transparency, clarity and certainty to ensure that companies are encouraged to undertake R&D activities and claim the R&DTI. The recommendations have the following 4 themes:
· Where compliance examinations / audits are necessary, they should take place as close as possible to the first year of registration of a project and should not be retrospective beyond one year
· Guidance materials needs to be comprehensive, clearer and up-to-date and developed in consultation with small business.
· Substantiation and record-keeping requirements should reflect commercial practicality with regulator personnel fully equipped to understand and collaborate with small business
· Small business must be assisted to help identify and retain professional and responsible R&D consultants.
The recommendations are:
1. ATO and AusIndustry administration of the R&DTI should be seamlessly integrated.
2. Clear joint ATO and AusIndustry guidance must be provided and maintained
3. The approach to compliance should be: (pg.6)
Proactive – assisting a company to comply in the first year of registration of a project
Professional – ensure staff reviewing/auditing/examining companies are properly trained and experienced in both the technical aspects of the legislation and how to apply it to industry sectors (particularly when R&D is underpinned by software using Agile and Rapid development methods).
Collaborative – AusIndustry’s statutory examination process should include as standard a meeting with the company to allow it to explain the registered activities and work through the company’s records.
Proportionate – Both ATO’s and AusIndustry’s requirements for supporting evidence should be proportionate to the size and complexity of the business and the amount of the claim
Advisors should be skilled and responsible for their advice:
Like the company tax return that includes a declaration by the tax agent, the R&DTI schedule should include a declaration by the R&DTI adviser.
Registered tax agents who prepare or lodge R&DTI claims should meet specialised further education criteria that demonstrates understanding of the requirements of the R&DTI.
Additional Continuing Professional Education (CPE) opportunities through structured courses should be provided by the ATO and AusIndustry to support registered R&D consultants achieve their CPE requirements.
Program Governance (pg.10)
The agencies should engage with small business in the same way and within the same time-frame that they expect of small business.
Changes to R&DTI Compliance Approaches by AusIndustry and the ATO
Both AusIndustry and the ATO in the second meeting, in October 2019, and in subsequent correspondence have advised that they have changed their practices in relation to R&DTI compliance
Under AusIndustry’s new Integrity Framework, each company (R&D entity) will be grouped in particular risk categories:
· Those “getting it right” are in a lower risk and support when required
· Those that “are trying to get it right” are medium risk with targeted education and advice
· R&D entities “who do not comply” are considered higher risk would be receive correspondence and meet with AusIndustry to discuss eligibility issues – the company may withdraw their application or AusIndustry may commence a statutory assessment…. after which AusIndustry can either discontinue if evidence is provided by the company. A review process is also available if activities are still found to be ineligible.
AusIndustry has released a Service Commitment outlining the level of service their customers can expect. In addition, AusIndustry has told ASBFEO that they will undertake training for staff in stakeholder management and customer experience.
Value of the Program (pg.12)
Claimants increased from 6,475 in 2013 to 13,156 in 2017-18.
A reduction of $700 million (or 13 per cent) in tax offsets paid in 2017-18 from the previous 2016-17 year despite a small increase in the number of claimants.
This reduction in tax offsets paid could be attributed to increased compliance activities, clawing back prior year claims.
Of particular note is that software claims on the program have increased (Software R&D expenditure represents 25-30 per cent to total claimed expenditure) and that SME expenditure claims on the program have increased significantly. Approximately one third of the 13,000 companies registered for the R&DTI are undertaking software development. The other companies are in a wide range of industry sectors.
Approach / Methodology (pg.14)
ASBFEO has investigated the effect of the R&DTI legislation, policies and practices on small businesses or family enterprises as a result of a number of companies approaching ASBFEO about the treatment and outcomes of their R&DTI audit/examinations by the ATO and AusIndustry. They reported that the interpretation of the legislation by AusIndustry and the ATO regarding the eligibility and substantiation of software R&D claims, has shifted to become more rigid. Companies that are not software related, raised concerns with ASBFEO where the ATO had retrospectively ruled that their R&D Incentive activities were not eligible R&D even though AusIndustry advised they were.
In undertaking its investigation the ASBFEO:
1. Examined how small businesses typically use the R&DTI;
2. Determined if there had been a change in what activities were considered eligible by AusIndustry and the ATO pre and post the 2019 software guidance material published by AusIndustry and how this had been conveyed to small businesses.
3. Determined the impact on a small business of the ATO and AusIndustry being able to audit/examine R&DTI claims going back four years.
4. Identified the circumstances in the legislation where the ATO can decide that the registered activities are not eligible activities.
5. Examined the liability a R&D consultant incurs where the ATO and/or AusIndustry has audited/examined their client’s R&DTI claim, ruled the claim ineligible and ordered the benefit be paid back to the ATO.
ASBFEO interviewed and received communications regarding over 20 small business R&DTI claimants who had been reviewed/audited/examined by the ATO and AusIndustry to understand the agencies process and the impact they had on the businesses, with a ‘deep dive’ into eight of these companies. ASBFEO examined the ATO and AusIndustry correspondence/decisions/position papers from the agencies to these companies (pg. 15)
Six small and medium R&D consultants were interviewed, as well as with eight R&D consultants from major firms.
Summary of Findings (pg.16)
· Overall, the program is exceedingly complex.
· The operation of the ATO and AusIndustry are not properly integrated.
· The ATO and AusIndustry have critical roles to play in providing further clarity of operation and administration of the law.
· The ATO and AusIndustry guidance material is fractured and incomplete.
· Compliance activities have been reactive, lack the commercial understanding of how small businesses operate, assume guilt, are resource intensive and result in long lag times for decision making so that the viability of the scrutinised business is threatened.
· Many companies are scaling down their R&D efforts in Australia and have reduced their R&D staff due to experience with both agencies compliance activities and the uncertainty surrounding eligibility and the substantiation of the R&D and its expenditures.
· The program complexity and issues with the administration of the program results in the R&D advisor role being critical to the company. While the R&D consultants we spoke to operate professionally, there are those who are incentivised to increase claims and are not held responsible for their advice. Unscrupulous operators are able to continue to advise businesses for long periods of time.
The Retrospective Nature of Program (pg.17)
AusIndustry has the ability to make a finding that is binding on the ATO for four retrospective years. Statutory examinations undertaken years after the R&D has been undertaken has created an uncertain environment for innovative companies undertaking R&D in Australia.
AusIndustry advise that R&D projects and activities can span across multiple years. Businesses are required to register their activities annually. As such, an examination being undertaken in one year on R&D activities for a project covering previous years may require examination of those earlier years, particularly, if the core activity was conducted in a previous year.
It is however unfair to adversely affect a company at such a late stage following the incentive payment, particularly when they had, for several years, self-assessed using the guidance material that was available at the time and which they believed they could rely upon. The majority of the companies spoken to had registered R&D activities which they genuinely believed were eligible, on the information provided to them by AusIndustry and the ATO. Many had reinvested the offset back into the business.
AusIndustry’s contention is that ‘the legislation has not changed’. There is evidence however that the current AusIndustry software guidance, in particular, has an effect of narrowing their interpretation of the legislation
Anecdotal evidence suggests that some companies are reporting their R&DTI estimated refund as a liability in their financial records, to recognise the potential of the company having to repay the funds accessed in previous years. This has an impact on possible investment into the company.
A finding about the eligibility or otherwise that is applied as close as possible to when the R&D is undertaken has the potential to be fairer to a company, especially for first time registrants.
AusIndustry has said that, in practice, this is difficult as a high percentage of registrations occur close to the 30 April cut-off date and it would be very resource intensive to reach a high number of those.
AusIndustry, will have additional resources to assist in their integrity functions early in 2020. A procurement process was advertised in early 2019 to engage a third party provider to assist with a range of R&DTI integrity functions. When questioned, AusIndustry said that this was as part of the reform measures to the R&DTI announced in the 2018-19 Budget, where the Government committed additional resources to support program integrity and administration.
All decision making will be retained by AusIndustry, however the contractor will assist with integrity activities and provide recommendations for review and decision. The expectation is that there will be a contract in place in late 2019 with activities commencing in early 2020. A significant proportion of these new resources should be directed to pre-registration examinations, particularly for first time registrants to ‘help to get it right’.
AusIndustry Integrity Measures (pg.17/18)
Although customer service/customer experience and compliance/integrity activities should not be mutually exclusive, we have received many complaints from companies about the fairness and transparency of the statutory examination process and the cultural mindset of AusIndustry assessors. These include:
1. The summary of key issues in the statutory examination letter is a generic list and provides little specific information relating to the company’s registered activities (a copy of the letter is at Attachment D).
2. The letter contains little guidance regarding the type and volume of contemporaneous documents that would meet the program requirements (a copy of letter is at Attachment D).
3. AusIndustry has been very reluctant to meet with companies during the examination process. Evidence is required in writing and many companies are concerned about the correct interpretation of the documents.
4. Long time-frames for AusIndustry to respond to the company and to make a decision during the statutory assessment process.
5. We have been advised that there have been instances where the AusIndustry assessor’s attitude is that the registration is ineligible and it is up to the company to prove differently.
6. We have been advised that an AusIndustry assessor often says ‘A Google search shows me that this existed [at the time of the project] and therefore the core R&D activities cannot be eligible’, thus trivialising the R&D effort and not recognising or acknowledging the fast moving nature of R&D in many sectors.
Both AusIndustry and the ATO appear to take a ‘one-size-fits-all’ approach to companies in examining and auditing the R&DTI registrations/claims. Their requirements do not differentiate between large and small companies – or the size of the offset.
AusIndustry Guidance Material (pg.19)
AusIndustry has issued a range of guidance material over the period of the program.
Incorrect interpretation of the legislation caused confusion for companies.
The Frascati Manual was not referenced in any AusIndustry or ATO’s guidance material released prior to February 2019.
The AusIndustry 2019 Software Guidance material refers to the Frascati Manual, however it makes selective use of the wording in the manual, which in effect, restricts the interpretation of the legislation.
AusIndustry’s responds to Frascati Manual by saying “The test for what is an R&D activity for the purposes of the R&DTI is not the same as the test for an R&D activity as set out in the Frascati Manual.”
“Software is developed in Agile teams, which try things and discard things, it doesn’t map well to the structure that the Frascati model tries to create.”Daniel Petre, Co-founder of Airtree Ventures, quoted in the AFR article of 3 December, 2018 ( pg. 20)
Panel members say no offer was made about revisiting the current guidance material.
The first Federal Court decision in relation to the eligibility of the R&D activities was made on 25 July 2019 in Moreton Resources Limited v Innovation and Science Australia  FCAFC 120 and confirmed that the process of statutory interpretation must always lead back to the text of the legislation, regardless of the wording in any other extrinsic materials. This finding is critical, given the apparent over-reliance on other materials such as the Explanatory Memorandum in recent decisions by the AAT. An analysis of the Moreton Resources decision is at Attachment C. The matter has been sent back to the AAT for further consideration. AusIndustry has not provided further information regarding their ISA’s position going forward.
In summary, there are a number of issues with the new AusIndustry software guidance material which include:
1. It is written in in a prescriptive way when referencing the principles in the Frascati definition
2. Its focus is largely on why activities are not R&D with little focus why it would be R&D. for example there is no guidance on how software development using the Agile method can demonstrate that the development is, or includes, R&D activities; and
3. When its quotes the Frascati manual it has made changes that affect how the definition of understood. (pg. 21)
ATO Integrity Measures (pg.22)
Throughout our consultations however, companies reported that the ATO staff involved in reviews and audits were aggressive in their interactions, exhibited poor client engagement skills, displayed a lack of respect and disregarded the taxpayer’s representation.
Lack of advice from the ATO about the type and volume of documents demonstrating nexus between R&D activities and notional expenses.
When challenged to outline the exact documentation needed, the ATO has been known to advise they cannot provide confirmation as they may give the wrong advice.
Taxpayers are given strict time frames in which to produce documents or answer questionnaires.
there is no equivalent response time limit applicable to the ATO. Taxpayers are left to wait for months for a response and in cases we have seen, another request for information.
ATO Penalties (pg.23)
The ATO has ruled in some cases that a company did not demonstrate ‘reasonable care’ as the tax advice it received was not considered ‘independent’ because R&D consultant charged a commission based on the value of the tax offset (not a fee-for-service).
ATO Debt Recovery (pg.24)
The ATO continues debt recovery action while a debt is in dispute.
Interest should only be charged once all venues of objection are finalised.
ATO Ruling on Eligibility of R&D Activities
There is no legislative basis for the ATO to determine whether or not the activities are eligible under the legislation.
The Tax Practitioner’s Board (TPB) (pg.25)
Given the complexity of the legislation, the R&D consultant should be an expert in the legislative aspects of the program; however there is no guarantee that a tax agent with the R&D condition understands the rules of the R&DTI and has experience in assessing the validity of a company’s claim. There is also value in the R&D consultant understanding the broader tax landscape. (pg.26)
R&D Tax Consultants (pg.26)
We heard from R&DTI consultants that often client companies elect to pay a percentage of the tax incentive as a fee for the service so they have certainty of the cost and are not required to fund the expense from the current cash flows. (pg.27)
Attachment A – Brief Overview of R&D project methodology using Agile and Rapid (pg.28)
The development of new and improved software can and must be able to be considered R&D where this is appropriate. The objective is to create new knowledge
The first section ( of the eligibility criteria) clearly states the purpose as including new knowledge in the form of new products, processes and services. The corollary of the exclusion in the second part is that developing, modifying or customising can be a core R&D activity if it is for a different dominant purpose. This would include:
Newly developed, modified or customised software for internal use by the R&D entity, a connected entity or affiliate that is for production or any other non-administration purpose, or any newly developed, modified or customised software that is created primarily for external users even if it is internal administration software.
This eligibility would be limited to the degree the development, modification or customisation includes, or is reliant on, the new knowledge created by the core R&D activities conducted for the R&D entity.
The R&D Tax Incentive is not a scientific research program. However, it does require that the experimental development activities needed (if any) to create, or enable the creation of new, modified or customised software for most purposes must be conducted in a scientific and verifiable way.
Application to Agile and Rapid based software development processes(pg.29)
A common error by R&D entities is that the creation of a new piece of software or a new function within an existing software product is R&D because it is a systematic progression of work to create a new or improved product.
Methodology Does Not Determine Eligibility (pg.30)
The experimental activity must be necessary because the outcome could not be determined in advance based on existing knowledge or experience. There needs to a more than insubstantial purpose to resolve uncertainty in outcomes in the conduct of the activity.
This uncertainty must be being resolved by answering the hypothesis in the systematic progression work. That is, the development cycle must be being applied to conduct the experiments and they must measure the outcomes so the R&D entity can determine the logical conclusion to the questions raised in the hypothesis.
How Agile and Rapid Can Demonstrate R&D Eligibility (pg.30,31)
The philosophy behind the development of modern software development processes is to optimise …. processes so they are a complete development management system, but the process does not get in the way of, or needless delay, the deployment of good and tested products. It has allowed the compression of the cycles of development to the extent that at the extreme the time between public release sprints can be weekly instead of the traditional once a year release or less. This has resulted in a sharp decline in the creation, and need for, systematic development documentation. The traditional stages of:
Set requirements → Design → Develop → Test → Maintain
are replaced by a process of many short sprints managed through Scrums or alike:
* Grouped into User Stories that may be grouped at a higher level into Epics for larger or complex overall goals.
* The Backlog features are ranked and weighted in order of importance.
It would be up to the project manager to document the discussions of the obstacles and the development progress, performance and results.
If a software developer is following this process and documenting the critical uncertainties and their conclusion, then there is no reason to argue that the use of Agile means that the business is not able to demonstrate that the development was, or included, R&D activities within the systematic progression of work.
The use of Agile and equivalent methods can be an effective tool to manage R&D software development activities. So long as the ATO compliance reviews seek to merely readily verify the calculation of estimates, determinations and calculations, then these styles of methodologies should not prevent the R&D entity from providing documentation and the particulars showing the basis and method of these allocations between R&D and Non-R&D activities. This would meet the requirements of s 262A.
Equally, a review by AusIndustry should be approached from an understanding of how the R&D entity applies the Agile process. This would include a judgement of d whether the project manager and development team properly group backlog features into stories that separate the core,
supporting and non-R&D activities and how the conduct of development tasks, processes of managing progress and obstacles and determinations of results and conclusions demonstrates the R&D. They should not expect each ticket against each individual backlog feature to be individually demonstrable as an R&D experiment.
To focus on each individual development ticket and expect each invoice to individually demonstrate the eligibility of the activity and the nexus of the expenditure to that experiment is missing the point of seeking to encourage more businesses to do more R&D. It takes a business’ focus off the R&D to create documentation that is not required by the law and will fail to encourage businesses to better manage their new knowledge creation processes. It is not needed to readily verify what R&D a business did and whether it is reasonably calculated.
Attachment B (pg. 34)
The current R&D legislation and application of the 2015 Frascati Definition to R&D in the development of software.
The R&D legislation was passed into law in 2011.
The legislation enacted a very significant broadening of what is eligible R&D compared to the old definition.
The new definition needs to be considered as a whole and in how it applies to software development in a conceptual way. That is, it should be applied based on the principles it expresses and not as prescriptive list.
Case Study – the Company, the ATO and AusIndustry (pg.38)
Where the project is an entirely new application/system and it is prior to the creation of a Minimum Viable Product (MVP) then the majority of the project will likely be core R&D developments and the supporting setups, frameworks, libraries etc. It is possible that a small proportion of the activities will be neither core R&D experiments or supporting R&D tasks directly required to enable these experiments to be conducted. Typically, this will require a very high percentage of the development activities to be either core or supporting R&D activities
Where the project is an upgrade or improvement to an existing system
Typically, the Core and Supporting R&D activities will drop to around 30 per cent of the development, maintenance and support costs.
Examinations/Reviews by AusIndustry and the ATO (pg.38)
Once the company is selected for a review by AusIndustry or the ATO, the current processes will seek to exclude 100 per cent of the activities for the maximum time allowed (typically 4 years) as the starting point. This will be either because:
a. AusIndustry has rejected the whole of the claim because it considers the claim to be the whole of the project. Examples have been given of cases where:
· the claimant has already demonstrated that the claim is limited to only specific areas of improvement or new capabilities.
· AusIndustry then seeks to exclude any experiments on aggregation, implementation or integration and then breaks down the functional improvements to such a granular level that the R&D is missed.
b. The competent professional test is often misapplied:
· would a competent professional consider the outcome should be possible
– instead it is applied as –
· would a competent professional consider that the method to reach the outcome can be known in advance – so no experimental development activity is required.
c. The ATO will reject the totality of the expenditure claim because the contract developer costs or the Jira time and task tracking data does not demonstrate the nexus between the expenditure and the activities they are for.
d. The ATO will expect a level of documentation that is not available or reasonable to expect. This has seen a progressive ramp up in previously acceptable
documentation based on reasonable estimates and allocation methodologies based on previous guidance from the Commissioner of Taxation. Today there is an expectation that all businesses will have highly accurate and separately verifiable time-sheets (whilst admitting that the AAT has stated for internal employees time-sheets are not required) and overhead allocations based on counting emails and envelopes, rejecting invoices from suppliers that only contribute to R&D activities for not detailing the R&D activity and accurate calculations of percentages of office space excluding walkways and amenities.
e. The ATO will increasingly rely on its own determination of whether the R&D entity is conducting R&D unless an assessment has already been by a delegate of Innovation and Science Australia.
Attachment C – The implications of the Moreton Decision (pg.40)
The Tax Laws Amendment (Research and Development) Act 2011 sets out the object of the legislation and the definition of core activities. The legislation requires a test that relies on the uncertainty in outcomes that are only able to be determined by the conducting experimental activities. It also accepts that R&D includes the experimental development of new and improved materials, products, devices, processes or services. That is, it recognises that business enterprise expenditure on R&D (BERD) is about the development of new knowledge as defined and grouped in the Frascati definition as Pure and Basic Research, Applied Research and Experimental Development:
the systematic “Experimental development” work that draws on the experiments that developed the new knowledge and/or practical knowledge to produce new or improved materials, products, devices, processes, systems and services 38.
ISA errors in R&D Tests (pg.41)
It was found that there were two errors in the application of the tests to determine if the activities are R&D. The first is that it is incorrect to consider that the first line in s 355-25(1) of the legislation creates a distinction between experimental activities and other R&D development activities that meet the requirements in s 355-25(1)(a) and (b). This line is:
Core R&D activities
1) Core R&D activities are experimental activities:
The Federal Court’s decision was to reject the ISA position that these two highlighted words create an additional test or requirement over:
a. whose outcome cannot be known or determined in advance on the basis of current knowledge, information or experience, but can only be determined by applying a systematic progression of work that:
i. is based on principles of established science; and
ii. proceeds from hypothesis to experiment, observation and evaluation, and leads to logical conclusions; and
b. that 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).
That is, the term experimental activities are those that meet the tests in a. and b., not that eligible activities are a subset of activities that meet the tests in a. and b.
The second place that ISA was found to be incorrect is its application of whether the definition restricts R&D to excluding experimental development activities because they are seeking to apply current technologies. The R&D entity was seeking to generate energy from mine gases by using an existing gas turbine generator. Both mine gas power generation and generation using gas turbines are known technologies. However, the R&D entity identified clear and significant uncertainties requiring experiment development activities to use these technologies under the planned circumstances. The Federal Court agreed with the R&D entity that there was uncertainty in determining the outcome in advance of any experimentation on the basis of current knowledge, information or experience.
ISA errors in the purpose: (pg.41,42)
Purpose test by ISA since January 2017 has been to sharply reduce this to more of scientific research support program. This has often seen the rejection of R&D activities on the basis that they are project outcome focused, not knowledge research focused.
This reflects the progress of the development of the Bill in that its intent is to encourage R&D and not just research.
The errors identified in the case are very applicable to R&D in software development.
This is often the only purpose of different phases of development of software development. This would apply to the development of a completely new item of software up to the MVP (minimum viable product) for testing by alpha and beta testers. It would also apply to a project to create new functionality to an existing product or improve the operating processes of that product.