RDTI Five-Year Evaluation

Angela Hodges • 26 November 2025

RDTI delivers billions in value – but many firms still miss out.

Here’s how to fix that.


The Research and Development Tax Incentive (RDTI) recently reached a major milestone with the release of its first statutory five-year independent evaluation. Introduced in 2019 to replace the Callaghan Innovation Growth Grants, the RDTI provides a 15% tax credit for eligible R&D expenditure and was designed to broaden access and stimulate innovation across the economy.


The evaluation confirms that the RDTI is working, that it is generating material additional R&D, and that it provides a better economic return than the former Growth Grants regime. It also highlights friction points, particularly for smaller innovators, where the compliance burden can be disproportionately high.


1.0 What the Evaluation Found: A strong case for retaining (and refining) the RDTI


Over its first five years:

  • 1,752 firms received support.
  • $1.074 billion in tax credits was provided (nominal value).
  • By 2023, RDTI-supported firms accounted for 65% of all measured R&D expenditure, compared with a 44% peak under Growth Grants.


Just over half of entrants had never received any Callaghan Innovation support, potentially demonstrating a broadening of access.


2.0 Additionality and macro-economic return: compelling evidence


The evaluation found that:

  • The RDTI generated additional R&D spend of approximately $274k per firm per year.
  • Total additional R&D created was $1.833 billion (present value).
  • The “bang for buck” is 1.4, meaning each dollar of government support generated ~$1.40 in additional R&D, comparable to strong OECD performers.
  • Economy-wide impact is estimated at 4.2x the government investment, or approximately $6.8 billion in GDP return.


These findings strongly support continuing the RDTI as one of the anchors of New Zealand’s R&D support system.


3.0 Innovation and productivity impacts


The evaluation found:

  • Innovation rates increased after two+ years of RDTI support (average uplift 6.1 percentage points).
  • No clear productivity effect yet, likely because productivity gains from R&D take much longer to materialise.


4.0 Where the System Struggles: High complexity


The evaluation makes clear that understanding RDTI eligibility, documentation rules, and expenditure tests remains a significant challenge, even for moderate-spend or technically sophisticated firms.


One of the most important passages in the evaluation states:

“Many firms found the in-depth description of eligible activities (IR1240) difficult to navigate, making tax consultants a valuable resource in interpreting scheme requirements. It was common for businesses to struggle to understand what was required without external advice.”


A firm interviewed summarised the challenge:

“It’s hard enough to understand [expenditure eligibility] when someone’s drip-feeding it to you… If we had to go and work out what we have to record, what is the difference between supporting and something else, and what percentage of total salaries can be claimed, we’d really struggle.”


From our experience advising clients across multiple sectors, these difficulties are familiar. Even highly capable taxpayers often:

  • Misinterpret the scientific/technological uncertainty test.
  • Misclassify supporting vs core activities.
  • Misallocate expenditure categories.
  • Under-claim eligible salaries and overheads.
  • Overlook small R&D projects entirely due to documentation effort.
  • Struggle to reconcile GA approvals with SR expenditure requirements.


5.0 Administrative inefficiencies


Although approval processes have improved, the evaluation continues to identify issues with Supplementary Return processing delays.


6.0 Software development: a systematic grey area


The evaluation identifies software R&D eligibility as one of the most contested and inconsistently assessed components of the scheme. Firms frequently encounter:

  • Difficulty documenting uncertainty within agile methodologies.
  • Confusion around what constitutes scientific/technological uncertainty.
  • Different interpretations and conflicting guidance between agencies and reviewers.


This is expected to be a major reform focus.


7.0 Compliance costs: the biggest pain point – especially for low spenders


A recurring theme in the evaluation is that the RDTI is disproportionately expensive for low-spend firms. Many businesses reported needing $300k–$500k in annual R&D spending before the incentive became financially worthwhile.


This is at odds with the scheme’s intent, which was to support early-stage innovators and firms with low or irregular R&D expenditure. Instead, smaller firms often:

  • Choose not to claim, even when eligible.
  • Face documentation requirements that outweigh the 15% credit.
  • Encounter heightened friction in software-related claims.


For precisely the firms the scheme seeks to support, the RDTI can be economically unattractive in practice, despite being available in theory.


Our fixed-fee model directly solves this problem:

NZ Tax Desk offers fixed-fee RDTI and RDTLC engagements, ensuring:

  • Claims are viable even for firms spending far less than $300k.
  • Early-stage and loss-making innovators can claim without fear of runaway costs.
  • R&D documentation systems are set up correctly from day one.
  • Clients maintain compliance confidence through complex SR reviews.


The evaluation makes clear that external advice is not just beneficial – it is often essential. Our structured, predictable pricing ensures smaller innovators are not locked out of the incentive. We don’t take a percentage of your claim, because R&D funding is for your innovation, not our commission.


8.0 What R&D-performing businesses should be considering now


The five-year evaluation of the RDTI makes one point very clear: navigating New Zealand’s R&D tax regime is complex, especially for smaller firms, software developers, and businesses without specialist tax capability. Many businesses interviewed for the evaluation reported that they struggled to interpret the rules, document uncertainty, and often needed external advice simply to understand what to record and how to present their R&D activities.


NZ Tax Desk can help. We offer fixed-fee RDTI and RDTLC advisory services, making claims viable even for low-spend businesses. We can draft all documentation (including General Approval submissions, activity descriptions, and Supplementary Returns), and ensure your claims are accurate, defensible, and optimised.


Contact us now to turn your innovation into tangible financial benefit.





Disclaimer:
The information provided in this article is general in nature and does not constitute personalised tax advice. You should consult with a qualified tax adviser familiar with both New Zealand tax rules and any relevant overseas tax systems before making decisions based on this content.


by Angela Hodges • 28 September 2026
Artificial intelligence is now part of a growing number of product-development projects. For New Zealand businesses, this raises a practical tax question: could any of the development expenditure qualify for the Research and Development Tax Incentive (RDTI) or the R&D Tax Loss Credit (RDTLC)? The starting point is to identify the particular development activity, the scientific or technological uncertainty being addressed and the systematic process used to investigate it. Two forms of R&D tax support The RDTI generally provides a 15% tax credit for eligible R&D expenditure. It may assist established businesses undertaking R&D as well as earlier-stage companies, subject to the applicable eligibility, approval, expenditure and filing requirements. The RDTLC is directed at eligible loss-making companies. Broadly, it may allow a company to obtain the cashflow benefit of tax losses attributable to qualifying R&D expenditure rather than carrying all those losses forward. It has its own eligibility rules, calculations and future tax consequences. When might AI development qualify? The strongest starting point may be where a business is developing its own AI technology. This could include experimental work on a new model, algorithm, training method or technical architecture where it is not known at the outset whether the required capability or performance can be achieved. However, potential R&D is not limited to businesses creating the underlying AI model. It may also arise when an existing AI model is integrated into a new product and that integration creates technological problems that cannot readily be resolved using existing knowledge and established techniques. For example, experimental development may be required to determine whether an integrated system can: achieve the required level of accuracy and consistency; identify unreliable or fabricated outputs; validate results against trusted data; meet defined processing-speed, scale, security or privacy requirements; operate reliably when information is incomplete or unusual; or continue to perform when a third-party AI provider changes its model. The potentially eligible R&D is the systematic work undertaken to resolve those technical uncertainties. It is not simply the decision to include AI in a product. Using AI to undertake ordinary development Many businesses now use AI assistants to write code, design workflows, analyse data, prepare content or automate routine tasks. These tools may reduce the time and cost of development, but their use does not make the underlying work R&D. If a business uses an AI coding assistant to build a conventional customer portal, for example, the project does not qualify merely because AI generated some of the code. The development activity must independently satisfy the R&D requirements. This distinction is important when preparing a tax credit claim. The relevant question is not how the development team used AI. It is what scientific or technological uncertainty the team was seeking to resolve. Example: an AI-enabled equipment-monitoring system Consider a business developing a platform that monitors specialist industrial equipment. The system receives data from sensors operating under variable environmental conditions and is intended to identify emerging faults before the equipment fails. The business incorporates a third-party AI model but also develops its own analytical layer to interpret incomplete sensor data, distinguish genuine warning signs from background variation and decide when an alert should be escalated for human review. It is not known whether the AI model and the analytical layer can be combined to achieve the required accuracy without producing an unacceptable number of false alerts. Established methods do not provide the answer, so the development team designs and tests different technical approaches against defined performance measures. That experimental work may qualify as R&D. However, the whole platform would not necessarily qualify. Standard dashboard development, customer account functions, routine reporting, deployment and ongoing monitoring may fall outside the eligible core activity. Review eligibility before claiming the credit AI development can involve genuine R&D, including where a business builds its own technology or undertakes experimental work to integrate an external model into a new product. However, neither the use of AI nor the novelty of the finished product determines tax-credit eligibility. Before including expenditure in an RDTI or RDTLC claim, the business should establish: which activities may satisfy the relevant R&D definition; where those activities begin and end; which expenditure is eligible under the applicable regime; and what approvals, returns and supporting evidence are required. Need help with an R&D tax claim? If your business is developing an AI-enabled product and you would like to understand whether any of the work may qualify, contact us to discuss your R&D activities and the requirements of the RDTI or RDTLC regime. Disclaimer: This article provides general information only and does not constitute tax advice. The RDTI and RDTLC rules are detailed, and their application depends on the particular claimant, activities, expenditure, contractual arrangements and supporting evidence. The examples are illustrative and do not confirm that any specific project or cost is eligible. Businesses should obtain advice based on their own circumstances before making a claim.
by Angela Hodges • 26 August 2026
Inland Revenue’s revised view of section CB 3 could widen the tax net for property sales, developments and subdivisions.
World map showing continents and oceans with blue-toned land and water.
by Angela Hodges • 21 July 2026
New Zealand’s foreign investment fund (FIF) rules have long caused problems for migrants holding overseas shares, particularly US citizens and Green Card holders living in New Zealand. The United States generally continues to tax these individuals on their worldwide income even when they are tax resident in New Zealand. At the same time, New Zealand’s FIF rules can impose tax on deemed income that does not match the income or gains recognised in the United States. This mismatch can create cash flow problems and, in some cases, double taxation because the tax paid in one country may not be available as a credit in the other. A new revenue account method (RAM) was introduced to address these concerns. However, the original rules were narrowly targeted. Following lobbying from tax advisers, migrants, investors and others affected by the rules, the Government announced a significant expansion as part of Budget 2026. These changes have already been widely discussed. In this article, we explain what they mean, particularly for US citizens living in New Zealand, and provide an update on where they have reached in the legislative process. Why are the FIF rules difficult? The FIF rules generally apply when a New Zealand tax resident holds shares in a foreign company. The most commonly used calculation method is the fair dividend rate (FDR) method. Broadly, FDR taxes 5% of the opening market value of a foreign share portfolio, together with an adjustment for certain shares bought and sold during the year. This can produce taxable income even where: no shares have been sold; no dividends have been received; the investment has produced no cash; or the investment has ultimately fallen in value or failed. This can be particularly harsh for people holding shares in an overseas private company or start-up. Tax may be payable each year on deemed income, even though the company has not paid a dividend and the shares cannot readily be sold to fund the tax. It is also problematic for US citizens. New Zealand may impose tax each year under FDR, while the United States may not tax the gain until the shares are sold. Because the two countries are taxing different amounts at different times, effective foreign tax credits may not be available. RAM: a “pay when you sell” method The new RAM taxes qualifying foreign shares on a realisation basis. Instead of being taxed each year on deemed income, an eligible taxpayer is generally taxed on: dividends received; and 70% of the gain realised when the shares are sold. A corresponding 70% of a realised loss may be recognised, although the use of RAM losses is restricted. For example, if an eligible taxpayer receives a $1,000 dividend and realises a $100,000 gain, their taxable RAM income would generally be: $1,000 dividend + 70% of the $100,000 gain = $71,000. RAM does not impose tax at a flat rate of 70%. Rather, 70% of the gain is included in the person’s taxable income and taxed at their applicable marginal tax rate. This approach is much closer to the capital gains regimes familiar to migrants from countries such as the US. It also means the taxpayer generally has cash from the sale available to fund the resulting tax. RAM was introduced but the initial rules were narrow The original RAM legislation applies from 1 April 2025. However, it is broadly limited to recent migrants and returning New Zealanders who: became fully subject to New Zealand tax on or after 1 April 2024; and had been non-resident for at least five years before becoming resident. Ordinary RAM generally applies only to qualifying unlisted foreign shares acquired before the person became a New Zealand tax resident. An extended version of RAM can cover listed and unlisted shares held by people who remain liable to tax in another treaty country because of their citizenship or right to live and work there. This is principally aimed at US citizens and Green Card holders. However, under the enacted rules, these individuals must still satisfy the recent-migrant requirements. This excludes US citizens who have already lived in New Zealand for several years, even though they face the same double-tax problem. Budget 2026 proposes a major expansion Following lobbying by the tax industry and people directly affected by the rules, the Government announced a broader RAM regime in Budget 2026. The changes would: allow all New Zealand residents to use RAM for qualifying unlisted foreign shares; allow qualifying US citizens and others subject to citizenship-based taxation to use extended RAM for listed and unlisted foreign shares, regardless of when they migrated to New Zealand; and increase the FIF de minimis threshold from $50,000 to $100,000. The removal of the recent-migrant restriction from extended RAM will be particularly helpful for US citizens and Green Card holders who have lived in New Zealand for many years. Once enacted, these taxpayers should be able to align their New Zealand tax more closely with the time at which the United States recognises the gain. This should reduce cash flow problems and the risk of tax being paid in both countries without an effective foreign tax credit. The expansion of ordinary RAM will also help New Zealand residents holding shares in overseas private companies. Instead of potentially paying tax each year on an investment that produces no cash, tax would generally arise when a dividend is received or the shares are sold. The de minimis threshold is also increasing Individuals generally do not have to apply the FIF rules where the total cost of their relevant foreign investments does not exceed $50,000. The Government proposes increasing this threshold to $100,000 from 1 April 2026. This should remove many smaller investors from the FIF regime, significantly reducing their compliance costs. The Budget changes are not yet law The expanded RAM rules and the $100,000 de minimis threshold were not included in the Budget night legislation and have not yet been enacted. We expect these changes to be included in the 2026–27 Annual Rates Bill later in 2026. If enacted as proposed, they should apply from 1 April 2026. The detailed legislation will be important. In particular, it should confirm the final eligibility requirements and whether the expanded RAM will be available to companies as well as individuals and qualifying trusts. Until then, taxpayers should distinguish between: the original, narrowly targeted RAM rules, which are already law; and the broader Budget 2026 changes, which remain proposals. Is RAM necessarily the best method? RAM will not always produce the lowest amount of New Zealand tax. FDR may be more favourable where an investment increases significantly in value because annual FIF income is generally based on 5% of opening value. Under RAM, 70% of a realised gain becomes taxable when the investment is sold. The method chosen can also have longer-term consequences. It may not be possible to move freely between RAM and the existing FIF methods, and leaving RAM may trigger a deemed disposal. The choice should therefore be made after considering: the nature and expected performance of the investments; whether the shares are listed or unlisted; when they were acquired; the taxpayer’s New Zealand residency history; the taxpayer’s US tax obligations; the availability and timing of foreign tax credits; and the treatment of losses. A welcome and practical change The proposed expansion of RAM is a welcome response to concerns raised by the tax industry and the people directly affected by the FIF rules. It should reduce the risk of taxpayers having to fund tax on income they have not received, provide much-needed relief for long-term US citizens living in New Zealand and make New Zealand a more attractive place for migrants and returning New Zealanders to live and invest. However, the rules are complex, and the choice of method matters. US citizens and other affected investors should review their portfolios before preparing their FY26 tax returns and obtain coordinated New Zealand and overseas tax advice. Contact us if you need assistance. Disclaimer The information provided in this article is general in nature and does not constitute personalised tax advice. The proposed FBT reforms are subject to legislation and may change before implementation. You should seek professional advice tailored to your specific circumstances before making any business or tax decisions based on this content.
by Angela Hodges • 17 June 2026
A More Practical Road Ahead?
by Angela Hodges • 27 May 2026
How Land Owning Structures Become “Accidentally Tainted” Over Time
by Angela Hodges • 23 April 2026
Some Common Sense for Taxpayers
Laptop displaying bar graph on a desk with financial documents, calculator, books, and an alarm clock.
by Angela Hodges • 8 February 2026
How To Prepare For Your Year-End Accounting.
by Angela Hodges • 27 January 2026
The Government has recently released a proposal that would fundamentally change how shareholder loans are taxed in New Zealand (Officials’ Issues Paper Improving taxation of loans made by companies to shareholders). At its core, the proposal could turn loans from a company to shareholders into a deemed dividend. Broadly, where a company advances funds to a shareholder and that loan is not repaid within a specified period, the outstanding balance would be treated as taxable income to the shareholder, most likely as a deemed dividend. This would apply to new loans made on or after 4 December 2025, with a proposed $50,000 de minimis per company (not per loan). This would mean that, for example, money taken out of a company by shareholders and left in an overdrawn current account could be treated as taxable income for the shareholders. Alongside this, Inland Revenue proposes a separate rule for companies that are removed from the Companies Register. Any shareholder loan still outstanding at the time of removal would be taxed at that point, on the basis that these loans are frequently never repaid and Inland Revenue has no practical way to recover tax once the company no longer exists. The stated problem: large loans that are never repaid Inland Revenue’s explanation for these changes relies on the concern that shareholders are taking funds out of the company, not declaring dividends, and not paying the funds back. The concern is not ordinary short-term lending. It is large shareholder loan balances that: build up over many years, fund private consumption, are never realistically repaid, and are often abandoned when a company is liquidated or removed from the register. From Inland Revenue’s perspective, these arrangements allow shareholders to enjoy company profits without ever paying shareholder-level tax, while IRD has (apparently) no effective recovery mechanism once the company disappears. That concern is understandable. However, the difficulty lies in how far the proposed solution strays from that original framing and the practical reality of how to implement the proposal. The $50,000 de minimis tells a different story Despite repeated references to very large balances and long-term non-repayment, the proposed rules would apply once shareholder loans exceed a $50,000 de minimis. This threshold applies to the company, so it will include all shareholder loans, not on a loan-by-loan basis. That threshold is not particularly high in the context of owner-managed businesses and does little to confine the rules to the behaviour Inland Revenue says it is targeting. In practice, the proposals could capture many ordinary commercial arrangements that bear little resemblance to the “never repaid” loans highlighted in IRD’s communications. New Zealand’s deliberate departure from Australia This tension becomes clearer when compared with Australia. Australia is cited as a model for taxing shareholder loans, but the Australian regime includes a critical safeguard: a commercial loan exemption. Where a shareholder loan is structured and documented on commercial terms, it is not treated as a disguised (or deemed) dividend. Inland Revenue has rejected adopting a similar exemption for New Zealand. The Issues Paper states that a commercial loan carve-out would be too easy to manipulate and would undermine the integrity of the regime. That decision has far-reaching consequences. It means that even a genuinely commercial loan, indistinguishable from third-party debt, remains exposed to the proposed deemed dividend rules purely because the borrower is also a shareholder. When a “loan” is taxed like income but still behaves like a loan Rejecting a commercial loan exemption also creates a series of unresolved technical and practical issues. If a shareholder loan is deemed to be income for tax purposes, but continues to exist legally, several questions follow: What happens to interest? Is this still taxable income for the company? Remember that, for tax purposes, the loan has been repaid via a deemed dividend. How are repayments treated? If the shareholder later repays the principal, should there be a deduction available to the shareholder for that repayment? i.e., to reverse the tax impact of the deemed dividend? What about future dividends? At this stage, the deemed dividend appears to be a tax fiction. The retained earnings remain in the company for accounting purposes. Unless the deemed dividend is matched by a reduction in retained earnings or tracked some other way, the same underlying profits could be distributed again later as an actual dividend — and taxed again in the ordinary way. What this would mean in practice From a practical perspective, the proposals would mean that overdrawn shareholder current accounts could be treated as taxable income for the shareholder, rather than simply being viewed as loans that remain outstanding. Inland Revenue has framed the changes around situations where large shareholder loans are not repaid, and shareholder-level tax is not ultimately collected. The proposed rules would apply more broadly than those scenarios, including to loans that are documented, interest-bearing, and intended to be repaid. As the proposals currently stand, further guidance will be needed on how deemed income amounts interact with ongoing loan balances, interest payments, repayments of principal, and future dividends funded from the same company profits. These interactions will be important in determining the overall tax outcome. If you would like to understand how these proposed changes could affect your business or existing shareholder loan arrangements, please get in touch with the team at NZ Tax Desk.  Disclaimer: The information provided in this article is general in nature and does not constitute personalised tax advice. You should consult with a qualified tax adviser familiar with both New Zealand tax rules and any relevant overseas tax systems before making decisions based on this content.
by Angela Hodges • 23 October 2025
Why You Can Now Trigger Tax Without Exercising
by Angela Hodges • 24 September 2025
What you need to know for new residents