mxx1's avatar

R&D Tax

The R&D Tax Incentive. First, you have to be a company, registered and eligible. Then you register for R&D tax. Then you register your annual R&D activities via their online form. Then you register again with the ATO when you file your tax return. So the process feels like register, then register again, then register once more. Each stage has the formality of “approval”, but none of them are.

The key is your self-assessment of your R&D: did you have a hypothesis, did you run experiments, do your records show it? If they review your self-assessment, AusIndustry decides whether your “R&D” was actually scientific research according to the Frascati Manual, or just ordinary product development dressed up in fancy words. The ATO checks whether the dollars you claimed really trace back to those activities.

I am not sure how your local SME is supposed to be an expert at the ridiculously obscure Frascati Manual bullshit approach to R&D. In truth, it’s anti-product development, so I am not sure these are skills we want to infect the broader business sector with. But wait, there’s R&D tax agents out there, so you can claim the R&D tax incentive without having to be contaminated with the scientific hypothesis.

R&D tax agents thrive on the gap between perception and reality. They will happily let you believe that all the thinking goes into the application, that once lodged it is somehow approved. They call it an audit if AusIndustry comes back, because “audit” is a familiar word, and less likely to make you question the premise. In truth, you have only registered your own self-assessment, and if the evil twins decide to “review” it, the burden of proof sits entirely on you.

The R&D tax agent takes your money for offering a few emails worth of advice and to cut and paste your self-assessment into the online form. What you are really paying for is the opportunity not to learn anything about the OECD Frascati Manual. Of course if you do get reviewed it doesn’t hurt the R&D tax agent at all. The risk is all yours and they get more money off you for managing the review process. If you lose, it doesn’t hurt their business at all; none of their other customers get to hear about it.

Tens of thousands of claims go through without being checked, while a few are “reviewed”. Of course the ATO doesn’t reveal the process for deciding which applications get reviewed, because that would let us design our self-assessments to avoid review. But my mate at the GPT reckons he knows.

He says; “The R&D Tax Incentive runs on self-assessment, but reviews are triggered through risk-based targeting rather than random audits. The ATO has said it uses “sophisticated systems” (read LLM) to detect non-compliance and specifically monitors industries of concern such as agriculture, construction, mining, and software development, as well as behaviours like claiming business-as-usual costs or poor record-keeping. AusIndustry has also shifted to a stricter compliance model, replacing informal education with formal examinations. In short, most claims go through untouched, but those that look unusual, aggressive, or fall into flagged sectors are far more likely to be examined.”

It is possible that the R&D tax agents pay for “drinks and dinner” (or near similar) with senior ATO folk in order to get the low-down on the review process. So maybe this also is what you are paying for when you pay them and then you do all the work.

But that’s Australia for you. Seems like 90% of the population earns a living by guiding the rest around the artificial barriers put up in their collective best interests. And the government carries on about our lack of “productivity”…

Simple maths: against the $4 billion headline cost of the R&D Tax Incentive you need to strip out the layers of leakage: government administration through the ATO and AusIndustry chews up perhaps 5–10% ($200–400 m); R&D tax agents and consultants take 10–20% of claims as fees ($400–800 m); companies themselves burn another 10–20% in staff time and compliance overhead ($400–800 m); the the R&D tax loans sector burns another $500m in profits extracted, and there’s the distortion cost of pseudo-research done only to satisfy the Frascati template, another $400–800 m with zero commercial yield. Add the opportunity cost of foregone product work and the fact that medium and large firms only get a non-refundable offset that is effectively repaid in future profitable years, and the $4 billion shrinks to zero gain, so we have negative productivity. That is, $4b of taxpayers money annually pissed up against the wall in the interest of a marketing campaign that no one cares about. Good job, Canberra!