26 November 2019
The anti-hybrids rules enacted in 2018 are detailed and complex rules which often depend upon the workings of the tax laws of other countries. Sometimes the way the foreign laws work will prove beneficial to a taxpayer under Australian law; sometimes it can be deleterious. Draft Taxation Determination TD 2019/D12 is the first ATO guidance on one of these elements in the anti-hybrids rules. It takes an approach that is detrimental to Australian taxpayers. It is arguably technically incorrect; it clearly contradicts the policy of the provisions.
Adjusting our law for what happens offshore
The anti-hybrids rules are focussed on income that falls through the gaps when two national tax systems interact, the most common instance being where one country allows a deduction for a payment and no other country observes assessable income. This hole can arise because of disagreements about the nature of the instrument, or about the characterisation of the entities involved, or some other disagreement. So it becomes critical to establish whether an amount was allowed as a deduction to someone, and whether an amount was included in income by some country. And if that state of affairs exists, it is then necessary to decide which country has the first option to remedy the outcome, and whether some other country must remedy the defect if that option was not pursued. To establish those facts, the anti-hybrid rules look at what is happening offshore.
Australia's legislation always alludes to what is happening offshore in very general terms; it legislates 'the vibe.' There is really no other option; national legislation could never keep track of the exact provisions in foreign tax legislation and refer to them with precision. Consequently, our legislation uses phrases like:
- is the amount used '... in working out its tax base for the foreign tax period ...'
- is the tax we are examining '... a law ... dealing with foreign income tax ...'
- does their law have '... a provision ... that corresponds to ...' a regime in our Act,
- does their law contain '... provisions ... [with] substantially the same effect as ...' a regime in our Act.
So just how close does the resemblance between the foreign rule and our rule have to be to satisfy those phrases? The Draft TD examines a situation where the taxpayer would benefit from a resemblance between the foreign law and our law; the Draft TD insists there must be a strong resemblance for the benefit to flow.
To see how this plays out in the Draft TD, consider a simple example: Aust Co pays a deductible amount to B Co located in Country B, which is owned by C Co located in Country C. Assume all the companies are member of the one corporate group. It should be sufficient to switch off the anti-hybrid rules (and allow Aust Co to retain its deduction) if the payment is included in someone's assessable income (whether that is B Co or C Co doesn't matter) by some country (whether that is Australia or Country B or Country C doesn't matter either). The country might be:
- Australia: if Australia viewed the amount as Australian-sourced income taxable by assessment in the hands of B Co, or
- Country B: if Country B imposes income tax on the net income B Co, or
- Country C: if Country C imposes income tax on the payment because --
- Country C sees B Co as transparent (and so includes B Co's income directly in the income of C Co) or
- Country C consolidates B Co into C Co (and so taxes C Co on an amount which includes the income B Co) or
- Country C sees B Co as opaque (but nevertheless imposes tax on the payment received by B Co in the hands of C Co under a CFC-type regime).
The anti-hybrid rules describe this last possibility in general terms: 'an amount of income ... is subject to foreign income tax [in Country C] if the amount is included in working out the tax base of [Company C] under a provision of a law of [Country C] that corresponds to section 456 or 457 of the Income Tax Assessment Act 1936 ...'
If Country C were the US, there would seem to be little doubt that its Subpart F regime (as it stood prior to 2017) is, 'a law of a foreign country that corresponds to section 456 ... of the Income Tax Assessment Act 1936 ...' because Subpart F was the model for CFC regimes around the world, ours included: s. 951(a) of the Internal Revenue Code 1986 (US) requires a shareholder with a 10% interest in a foreign company, that is 50% owned by US residents, to include in income its share of the CFC's 'Subpart F income.' Aust Co could keep its deduction if the US took the payment to B Co into account in working out the tax payable by C Co under Subpart F.
In 2017, the US expanded Subpart F to include the Global Intangible Low-Taxed Income ('GILTI') in Subpart F, and just like the prior rule, s. 951A of the Internal Revenue Code 1986 (US) includes an amount from a CFC in a US shareholder's income. The amount included is the return the CFC makes in excess of 10% of the value of the CFC's depreciable assets. The assumption behind this figure is that a company’s normal return on its tangible assets is 10%; so everything above 10% must be passive income earned on its intangible assets.
The Draft TD takes the view that s. 951A is not a regime that 'corresponds to' our CFC regime. There are only 2 paragraphs out of 7 pages to justify that conclusion; most of the Draft examines how GILTI came about and how it works. The Draft TD gives 3 reasons:
- the first justification given is, 'GILTI is widely referred to as a global minimum tax regime for which there is no equivalent in Australia' while our CFC regime is 'a general CFC anti-deferral regime.' However, the way the US has imposed this 'global minimum tax' is by expanding its existing 'general anti-deferral regime';
- secondly, the GILTI regime combines the amounts attributed from all CFCs while our regime operates CFC-by-CFC. This a matter of mechanics;
- GILTI is calculated to reflect 'the deemed high ... return on certain depreciable property whether or not the deemed return in fact comprises passive or tainted income ...' while our regime includes only 'actual passive or tainted income.' The US rules do presume that the supra-normal return is passive income without bothering to verify that. Just why this is fatal is not obvious; actual passive income will often be repatriated under Subpart F while this rule goes even further.
On the other hand, there are strong similarities between this portion of Subpart F and our own CFC rules:
- the regime attributes income only from a foreign company
- the foreign company must be controlled by a US shareholder or group
- the amount is included in the income of the US shareholders
- the amount is included whether or not the CFC pays an amount as a dividend
- the amount is calculated to reflect the CFC's presumed passive income.
The reality is, where GILTI is triggered, the US has included some amount in the income of the US parent under one element of their Subpart F regime that most definitely 'corresponds to' our CFC regime; the amount is 'subject to foreign income tax' which should immunise the Australian payer from adverse consequences under our anti-hybrid rules.
One can't help suspecting the real concern is that the maximum effective tax rate on GILTI ends up being only 10.5% because a shareholder which is a US company is given an automatic tax deduction equal to 50% of the amount attributed. In fact, the marginal rate on the GILTI could be much lower if tax was actually paid by B Co in Country B because a US corporate shareholder can also enjoy a tax credit for any underlying tax paid by the CFC. But since the anti-hybrid rules don't care (for the most part) about tax deductions or tax credits or the tax rate, only whether the amount was included in income, another line of argument is to deny the deduction to the Australian payer by making the case that this part of the US CFC attribution regime doesn't bear a close enough resemblance to our CFC regime.