Why There Is No Offshoring Tax

In Satyam Computer Services Limited v Commissioner of Taxation [2018] FCAFC 172, decided on 11 October 2018, the Full Federal Court considered payments made by Australian clients to an Indian IT services company for work performed in India by India-based staff. Ordinary Australian source rules would not have taxed that income.

The Court held it was taxable here anyway. The source rule in the India tax treaty deemed the payments to have an Australian source, and the treaty's definition of royalties is broader than the domestic one, reaching services that make technical knowledge available. The taxpayer argued that a tax treaty can only limit a country's taxing rights and never extend them. The Court disagreed.

The practical effect was an Australian tax on offshore IT services income. Not designed as an offshoring tax, but working like one.

On 2 April 2022, alongside the Australia-India Economic Cooperation and Trade Agreement, the two governments exchanged side letters on taxation. Australia agreed to amend its domestic law to stop taxing the offshore income of Indian firms providing technical services to Australian customers.

That became the Treasury Laws Amendment (Australia-India Economic Cooperation and Trade Agreement Implementation) Act 2022, Act No. 60 of 2022, which received royal assent on 23 November 2022 and applies to income years starting on or after 29 December 2022, the day the trade agreement entered into force. The services it covers are the ordinary subject matter of offshoring contracts: engineering, architectural and computer software development services delivered remotely.

So the sequence is: Australian law reached this income, the Commonwealth defended that position successfully in the Full Federal Court, and four years later Parliament gave it up in exchange for tariff concessions. Whatever one thinks of that trade, it is the reason the most obvious lever no longer exists.

Australia does tax imported services, through the GST reverse charge in Division 84. The Australian business buying the service accounts for the GST itself. For most businesses this is neutral, because they claim it straight back as an input tax credit.

It is not neutral for a business making input-taxed supplies. Lending money is input taxed, so a bank buying offshore services generally has to recognise reverse-charge GST on those costs, including general overheads, and cannot fully reclaim it.

The result is that the closest thing Australia has to a tax on offshoring is an unintended side effect of the GST treatment of financial services, and it falls hardest on the banks, which are the largest offshorers in this site's data. It was not designed to discourage anything, and at a few per cent of contract value it plainly does not.

The best-known attempt is the United States' Bring Jobs Home Act: a tax credit of up to 20 per cent of the cost of moving operations back onshore, paired with denying the deduction for the expenses of moving them offshore. It was introduced in the 113th, 114th, 115th and 117th Congresses, from 2012 onwards, and stalled every time. A version was still being introduced a decade later. None has been enacted.

Taxes pointed the other way, at inbound digital services, have fared better, but they are the opposite measure: they tax foreign suppliers selling in, not domestic buyers sending work out.

A credit or discount for employing Australians is the obvious alternative, and legally it is a far cleaner instrument than a penalty. All three obstacles above largely fall away. Tax treaties become irrelevant, because rewarding a domestic wage involves taxing no foreign supplier at all. The trade law problem mostly goes too: the WTO's binding subsidy rules sit in the agreements covering goods, while for services, GATS Article XV did no more than commit members to negotiate disciplines in future. Those negotiations began in 1995 and have never concluded. And the definitional problem inverts in your favour, because you no longer need to define offshoring, only an Australian employee, which PAYG withholding and superannuation already define exactly. You would be rewarding something easy to verify instead of penalising something easy to disguise.

Australia already does precisely this where it chooses to. The R&D Tax Incentive is worth billions a year and expenditure on R&D conducted outside Australia is simply not eligible unless the company obtains an Advance Overseas Finding. Getting one requires showing a significant scientific link to core activities performed in Australia, and that the work genuinely cannot be done here, and the overseas spend must stay below the spend on the related Australian activities. That is a major tax concession with a majority-onshore test built into it. The Future Made in Australia production tax credits, legislated in February 2025, work the same way: a 10% refundable offset for processing critical minerals in Australia and $2 per kilogram for renewable hydrogen produced here. The design the question implies is proven, in force, and uncontroversial. It has simply never been pointed at services labour.

Two things stop it working as a wage subsidy, though. The first is arithmetic. The gap being closed is roughly $100,000 against $30,000, so you need to find about $70,000, more than two thirds of the wage. A credit of a politically plausible size, 10 or 20%, moves the Australian cost to $90,000 against $30,000 and changes nothing about the decision.

The second is that most of the money goes to hiring that would have happened anyway, and Australia has a recent and unusually stark lesson in this. The JobMaker Hiring Credit was announced in the 2020-21 Budget at an estimated $4 billion, intended to support around 450,000 positions for young workers. Treasury documents obtained the following year showed that 5,278 people had been hired using it, just over 1% of the forecast, and the next Budget wrote $3.9 billion of the $4 billion back out of the estimates. Designing the additionality test is what makes these schemes hard: JobMaker required both headcount and payroll to rise against a baseline, and capped the credit at the payroll increase, and still almost nobody used it.

There is also something perverse in the existing settings. Australia does not merely fail to discount local hiring, it taxes it. State payroll tax charges employers roughly 5% on Australian wages and exactly nothing on an offshore services contract. Reducing payroll tax is the most direct discount for hiring locally available, needs no new scheme and no agreement with anybody, and would cost the states their second largest source of revenue, which is why it does not happen.

Nothing in a tax treaty or a trade agreement obliges a government to hand public money to a company that offshores, or to buy from one. Procurement rules, local content requirements and conditions attached to subsidies need no renegotiation with anybody. The same goes for the concessions already on the books: the R&D Tax Incentive shows that a condition requiring the work to be performed in Australia is both lawful and administrable, and nothing stops that test being written into other concessions.

Qantas is the clearest illustration of that gap. It received $2.7 billion in taxpayer support during COVID, outsourced 1,800 ground crew around the same time in a move later ruled illegal, and in November 2026 announced 850 corporate roles moving to Accenture, with the work delivered mostly from India. No condition attached to the public money prevented any of it.

The Commonwealth is also one of Accenture's larger Australian customers in its own right. This site records Accenture engagements at 18 government bodies and 12 listed companies, which means the same government that could impose onshore delivery conditions is a major client of the vendor those conditions would bind. See the vendors behind offshoring and government outsourcing.

It is also why the main union demand has been disclosure rather than taxation. The ASU and FSU have pushed for companies to be required to notify and consult before moving work offshore, on the basis that the information is not currently published anywhere. Compiling it by hand is what this site does.