The core activity test is the same in every sector: a systematic approach, a material purpose of creating new knowledge or new or improved processes, services or goods, and a material purpose of resolving scientific or technological uncertainty. New Zealand's economy puts particular weight on agritech, food, software and specialised manufacturing, and Inland Revenue's approval decisions have built up a picture of where the eligible work sits in each. This page walks through the main sectors; the industry guides go deeper.
Where qualifying work sits by sector
| Sector | Usually core or supporting | Usually excluded | Watch for |
|---|---|---|---|
| Software | Resolving technological uncertainty in algorithms, architecture, performance or data where the answer was not deducible | Routine de-bugging, internal administration software, use of known techniques | Describing the uncertainty, not the product, in the approval application |
| Agritech and primary industries | Trials of genetics, inputs, husbandry, sensing and processing with uncertain outcomes | Normal farming and growing, routine data collection | Seasonal trials spanning income years; feedstock on produce sold |
| Food and beverage | Formulation and process experiments with measured technical targets | Cosmetic or stylistic changes, taste-led recipe changes, market testing | The cosmetic exclusion |
| Manufacturing and engineering | New materials, processes and products developed through trials | Preproduction: tooling up, trial runs, demonstration of commercial viability | The preproduction exclusion; where the core activity ends |
| Life sciences | Preclinical and clinical experimental work, device development | Routine testing, regulatory compliance | Overseas trials and the 10% cap |
| Cleantech and energy | Materials, storage and generation experiments | Installing proven systems | Government co-funding to review |
Software
Software is the largest sector in RDTI applications and the one where Inland Revenue declines most activities. Its guidance is direct: the uncertainty must be technological, meaning that a competent software professional could not deduce whether or how a result could be achieved from publicly available knowledge, and the activity must be a systematic attempt to resolve it. Building a new product with known techniques, integrating systems through documented interfaces and configuring platforms do not qualify. Schedule 21 adds two specific exclusions: routine de-bugging and routine software maintenance, and software developed for the internal administration of the business.
Applications succeed when they name the technical unknown, explain why the literature and standard practice do not answer it, and describe the experiments: alternative architectures benchmarked, models trained and evaluated against a target, protocols tested under failure conditions. The supporting activities, such as building test datasets and harnesses, follow once the core activity is clear. The core activity ends when the uncertainty is resolved; the rest of the release is excluded preproduction.
Agritech and food
Primary industry R&D fits the definition well because experiments are visible: trial plots, treatment groups, controlled feeding studies and processing trials with measured outcomes. The traps are structural. Trials often span two income years, and each year's activities must be covered by an approval for that year. Produce or animals from trials that are sold trigger the feedstock adjustment. And normal farming and growing operations around the trial are not supporting activities unless required for and integral to the core activity. In food and beverage, the cosmetic and stylistic exclusion removes taste-led recipe changes and packaging design, while formulation to achieve a technical result such as shelf life, texture or stability can qualify.
Manufacturing and engineering
Process trials on the factory floor qualify where a technological uncertainty was resolved systematically: a new alloy through an existing line, a parameter study for a new material, a fixture that had to be developed and tested. The preproduction exclusion is the one to watch. Schedule 21 excludes from core activities preproduction activities such as demonstration of commercial viability, tooling up and trial runs, which means the core activity must end when the uncertainty is resolved and the move to production is outside the claim. Trial runs that are themselves the experiment, because the uncertainty can only be resolved at production scale, need to be described that way in the application. Prototypes sold or used in the business trigger the feedstock adjustment.
Life sciences
Drug, biotech and device development is close to the paradigm of a core activity. New Zealand's issues are location and funding. Clinical trials and specialist testing run overseas count only within the 10% overseas allowance and only where integral to a New Zealand core activity, which pushes companies to run what they can at home. Many life sciences companies are loss-making, so the refundability rules and their labour-related tax cap decide how much of the credit arrives as cash; the refundability page explains them.
Other sectors
Cleantech and energy. Development qualifies in the ordinary way; installation of proven systems does not. Co-funding is common and should be reviewed for its effect on eligible expenditure.
Construction and engineering services. Novel solutions can qualify, but the claimant must control the R&D, bear the risk and own the results; work for a client who owns the outcome is the client's R&D.
Screen, games and creative technology. Technology development, such as rendering or simulation methods, can qualify; content creation and research in the arts cannot.
Illustrative example
Illustrative example. Three hypothetical companies. Waikato Logistics does not know whether a constraint-solving approach can produce delivery schedules for 3,000 stops within 20 seconds; it benchmarks three formulations and records the results. Core activity; the driver app is excluded. Southland Alloys develops a casting process for a new aluminium alloy, running a designed matrix of trial pours before porosity was controlled. Core activity; the tooling up for production afterwards is excluded preproduction. Nelson Juices reformulates a drink to taste sweeter. Excluded, cosmetic. Its parallel trial to keep a cold-pressed juice stable for 60 days without preservatives, with microbial testing at intervals, is a core activity.
Related pages
Official sources
Educational content only, not tax advice. See the disclaimer.