The world is full of ‘Taxation of the Digital Economy’ material, but I saw a 1 hour ‘webcast’ by KPMG, on this subject, which I thought was very useful in that, it covered the fundamentals of the subject well, and gave good insight into the direction of future taxes, both direct and indirect.

This is an overview of their webcast: ‘Taxation of the Digital Economy – an Asia Pacific Perspective‘.

Why is ‘digital’ the lighting rod for changes to our tax system?

Taxation of the ‘digital economy’ was thought of, as a discrete action plan in itself, when the OECD’s ‘Base Erosion and Profit Shifting’ (BEPS) initiative began, in 2013 (it was ‘Action 1’). But all involved, soon realised that the whole economy is ‘digitalised’ to some extent, albeit to very different extents. The spread of ‘digitisation’ seems to be unstoppable, at present, so it carries real clout.

The OECD countries’ plans turned to a ‘whole of economy’ focus, at least in the longer term, rather than try to create a ‘ring fenced’ system of special rules, just for those businesses, more conspicuously ‘digital’. The whole of the tax system, and all the rest of the BEPS changes, would have to do the work, though some changes would have to be made, to the whole system, to meet the ‘digitised’ challenge.

Why did digital businesses ‘start the hare running’ in the tax world?

The features of digital businesses made it increasingly difficult for a country to tax the profits made by digital businesses, in their country.

These features are:

  1. It is easy to have ‘scale without local mass’. Think of the turnover Google, Facebook, Uber, AirBnB, and Netflix get from Australians, without (as far as I am aware) and presence here, or any significant presence here. This is largely due to the features of the internet and the development of new ways for creating value electronically.
  2. New ways of creating value – that can be delivered electronically and are highly dependant on intellectual property, which can be placed anywhere, and the use of data, and algorithms, to create value.
  3. Highly automated – in many cases, there are few individuals involved in delivering the product and it is highly automated, and the individuals could be based anywhere and occupy less obvious places in the value creation function.
  4. A high use of ‘user data’ – particularly in some sectors, such as Google and Facebook, where the usefulness of their service, to the public, leaves them with a digital footprint of a huge section of the public, which they can monetarise in various ways.

A survey of the types of digital businesses we have (so far)

Before asking why these features are such a challenge to direct taxes (in particular), it would be helpful to survey the types of digitised businesses we are familiar with, today.

  1. Use of websites as a ‘shop front’ – nearly every business now has its own website, for this purpose (if not more) – even this humble tax barrister:
  2. Online ordering of retail goods and services – now widely available, lead, perhaps by Amazon, for goods. And, for services, online banking might be a good example, or an airline booking site.
  3. Supply of digital information or content – the prime example of this is ‘Netflix’ (which has no physical presence here), but other examples might be the above KPMG webcast, and this article, on this site.  A ‘weather app’ might be another example.
  4. Digital platforms to find, sift, compare and rank data usefully, probably to facilitate a purchase. Examples of this might be: Webjet; Travago; Trip Adviser or iSelect.
  5. Digital platforms to match buyers and sellers – examples of these might be Uber, AirBnB; and the emerging ‘FinTech’ industry.
  6. Social media – like ‘Facebook’ where, it’s hugely popular information sharing platform, attracts enormous amounts of traffic, which it can monetarise, by charging advertisers, a fee, or by using the rich source of user information other ways, such as selling it to others (subject to ‘privacy issues’, which Mark Zukerberg has so publicly been apologising for breaching).
  7. Search engines – like Google, monetarises its services in a very similar fashion to social media. Its search engine function, is even more popular, and widely used, than Facebook and leaves a richer vein of personal data and preferences, which it monetarises in a similar way: charging people for advertising and sale of forms of that data, to persons who find it useful.

How do the ‘digital’ features create difficulty – particularly for ‘direct’ taxes (eg. ‘income tax’)?

It is useful to survey the building blocks for a direct tax, such as income tax, and how the digitisation of the economy is creating challenges for countries, like Australia, to maintain its tax base (as the scope of digitised business keeps growing).

  1. Income must have an Australian ‘source’, to be taxed in Australia (as the extent of digitisation grows). The income Netflix earns, from streaming movies, made outside Australia, into Australian homes, does not have an Australian source. The income KPMG earns, from consulting in Australia, has an Australian source. The money this site earns, from advertising, has an Australian source (because the content is created here and site is here). The the money Facebook earns, even from Australian advertisers, does not have an Australian source, as the place where the platform was created and operates, is outside Australia. In fact, there will be businesses, where the value created, is in no particular place, making the determination of source very difficult.
  2. A foreign resident must have a ‘permanent establishment’ in Australia, before Australia can, typically, tax its income, under most of our treaties with foreign countries (like a foreign bank with a branch here). However, Uber and AirBnB are examples of income that has an Australian source (supply of a ride in Australia or use of a house in Australia) but their income is not taxable here, because they have no physical presence here.
  3. Attribution of profits to the Australian ‘permanent establishment’, is required, under these double tax agreements, even for income with an Australian source, earned through a ‘permanent establishment’. This is typically done on an analysis of where the ‘value was created’, typically using indicia of where people are, and the assets are, and what risks have been taken, it supplying the thing that created the income. The conceptual task, at least, would be simple enough, in the case of, say, a foreign bank, with an Australian branch. But what amount, of the value, has been created, in Australia, by a highly automated process, invented and operated outside Australia, via servers that could be anywhere? This is why the Australian Multi-national Anti-Avoidance Legislation (MAAL) resulted in a big increase in income booked here, but not income tax (as there was not much value attributable to what happened in Australia). Interestingly, though, the threat of deeming the affected companies, to have an Australian ‘permanent establishment’, was effective to get the companies to restructure, and book income here (about A$6 b), and this did generate significant amounts of extra GST.
  4. Transfer pricing rules apply to ensure that taxable income is not ‘leached’ out of Australia by related party transactions understating income and overstating expenses. Understating and overstating, is determined by reference to, chiefly, arm’s length equivalent prices, which in turn depends on factors very similar to the ‘allocation of profits’ to a permanent establishment, namely where and to what extent the value was created. The same problems arise.

High intensity user participation (user data)

Examples of digital businesses, that have a high use of data, collected from users of their platform, or service, raises interesting questions for income tax purposes. Chief among these, is the degree to which the value, those business create, is from that data.

If data from Australian users, were the chief source of the value created, which Google then sells, for its advertising income or income from Australian preference profiles, then, all of a sudden:

  • That income might have an Australian source.
  • A high (not low) proportion of its profits might be allocated, to any Australian ‘permanent establishment’ it had; and
  • The transfer pricing issues, could fall the same way.

On the other hand, Google would say:

  • The data is something they just purchase, in a kind of no-cash, bata transaction, in return for a very useful search engine service, which was developed, and is owned, and operated, outside Australia.
  • So, it would say, even if you were trying to tax it, on a money equivalent, of the (search engine) services, it supplied, it would not have an Australian source, and would not be taxable here.

The likely ‘long-term’ solution for direct taxes

We have some idea what the long term solution might look like, but no agreement yet, between the countries that currently collect income tax, from these sorts of digitised activities, and that that don’t (yet).

Income tax might be levied on such income as follows.

  1. It would be levied on persons incorporated in, or resident in that particular jurisdiction if they also had a ‘significant digital presence’ in that jurisdiction, also.
  2. It would be dependant, also, on the person having a ‘permanent establishment’, under with a ‘digital’ extension to the definition of ‘permanent establishment’.
  3. It would be levied on the supply of services over the internet or some other electronic network, that were essentially automated.
  4. And profits would be allocated, to a non-resident ‘permanent establishment’ on an extension of the existing (arm’s length) framework, supplanted by some suitably tailored rules.

Significant digital presence

The definition of ‘significant digital presence’, currently being used in India (for its 6%  ‘equalisation levy’) is as follows:

  • It’s supplies (including electronic) exceed a particular monetary threshold (which could be set, from time to time, or by the relevant country, concerned).
  • It solicits business in (India) through digital means, on a sustained basis.

The Indian Parliament has imposed, this 6% ‘equalisation levy’, on business to business (B2B) advertising by a foreign company.

Digital Permanent Establishments

An example of the kind of extension, to the definition of ‘permanent establishments’ (to meet the ‘digitised’ challenge) might be, any of the following.

  1. Gross revenue, from digital services (into that jurisdiction), over A$7 m.
  2. Over 100,000 users within the jurisdiction.
  3. Over 3,000 on-line contracts.

Interim measures

Whilst waiting for consensus to emerge, about a long term solution, a number of countries have (and more might) go it alone and adopt interim measures, which are generally a type of indirect tax (a bit like the Indian 6% Equalisation Levy).

These have their problems though, as they are not ‘creditable’ through supply chains (in the way GST or VAT is), they would be payable, when the supplier was not profitable, and they would probably not be creditable, in a foreign country, against their income tax.

A comparison with indirect taxes (especially GST or VAT)

Many have observed, that direct taxes are set to head towards the conceptual model, used for indirect taxes, namely to tax in the country, where the consumption arises, or the payment is made.

GST or VAT taxes are conceptually appropriate for the digitised era, but they usually require amendment to address the issue of collecting the tax from foreign suppliers. The ways this might be done, include the following.

  1. Requiring certain foreign suppliers, platform owners, or shippers to register and pay GST/VAT on the things supplied (which is what the Australian Parliament has legislated for – starting 1 July 2018). This has produced modest furore, in Australia, when Amazon announced that it would only supply Australians, from its new Australian ‘hub’, which would pay the GST. The Australian hub has about 4 million items it supplies, which sounds alright until you understand that the number of line items, supplied in the US, is measured in the hundreds of millions.
  2. Requiring recipients of the supply to withhold some of the purchase price, and remit it to the revenue authority, on account of the foreign supplier’s GST. We have just had a purchaser withholding regime enacted, but not for recipients of foreign supplies, but rather for the supply of ‘new residential premises’ and ‘potential residential land’.
  3. Imposing the tax (and collecting from) the persons who provide the means for paying for these digitised supplies (typically credit and debit card services).

United States’ law regarding the nexus required, for the individual states, to impose their sales taxes, provides an interesting model of how competing indirect tax systems can attract the necessary nexus, to impose their tax, whilst not ‘over-reaching’ a neighbour’s taxing rights.

The KPMG presenters conclude that the future direct and indirect taxing systems are ‘morphing’ into shadows of each other.


John Morgan


Study questions

  1. Was the webcast, that prompted this article, by KPMG?
  2. Is there a ‘digital’ economy, for which a ‘ring fenced’ set of customised rules can be devised?
  3. Is a feature of digital businesses, that they can have ‘scale without local mass’?
  4. Does Google exhibit the other characteristics of ‘digital’ businesses, namely: new ways of creating value; highly automated and high use of user data?
  5. Can a person with an Australian ‘permanent establishment, pay tax on foreign source income?
  6. Did the A$6 b in extra revenue, that foreign entities booked, in Australia, since the MAAL was legislated, result in a similar increase in income tax?
  7. Could high levels of Australian user input, potentially give the entity’s advertising income an Australian source?
  8. Might the long-term reform to direct taxes involve digital extensions to nexus: ‘significant digital presence’ in lieu of source and extensions to ‘permanent establishment’ based on levels of digital revenue, number of digital users and numbers of on-line contracts?
  9. Might the short-term answer be a kind of ‘indirect tax’ on certain digital revenues?
  10. Is it true that direct and indirect taxes are moving closer together, in the way they operate, in this new digital age?



About the author