Nick Shaxson ■ Which countries have the right to tax Google’s income?
Recently, amid the furore over Google’s surprisingly low tax payments in the UK and in other countries, it has been suggested, as one observer put it to us:
“The claim is that international tax law accrues profits to where products are created, and not where sales are made. For example, a UK company that makes lots of sales in the US still pays most of its corporation tax in UK.”
This takes us back to the age-old fundamental question in international tax: when a multinational company operates in many countries, which country gets to tax which bits of its income?
The most general answer is that the international tax system, as it has grown up over the last century or so, has tended to see multinationals paying most tax in the countries where they are resident. So US companies pay most of their tax in the U.S., and other countries find it much harder to tax them, even when they are generating sizeable economic value in their countries, as Google is.
A second related problem, of course, is that the rules are so easy to game that it’s possible to build devious pathways through the thickets of international tax law so that the multinational uses transfer pricing and other tricks to pay little or no tax anywhere. The OECD, a club of rich countries, was tasked with setting up systems to beef up the international tax system, and the resulting BEPS (Base Erosion and Profit Shifting) project has made some progress, but also been subject to heavy criticism – not least by us.
Now TJN Senior Adviser Prof. Sol Picciotto has written us this guest blog, to put the Google case in this bigger context.
A guest blog by Sol Picciotto:
Current international tax rules require tax authorities to analyse each firm’s business to identify the “functions performed, assets employed and risks assumed” by each of its component entities. The key functions for Google, as for most firms in the “knowledge economy” relate to intangibles.
The BEPS project resulted in some significant revisions of the chapter in the OECD Transfer Pricing Guidelines on Intangibles, which is now much longer. Cutting through the technical complexity, they stress as key functions for intangibles “design and control”, “direction of and establishing priority”, and “management and control”. Clearly this favours the US and other developed states, which can argue that innovation and R&D are “controlled” there, even if much of it takes place elsewhere.
In the Google case, Google UK performs both “marketing” functions, as well as some R&D. The R&D is not insignificant, as the Google tax story came out at about the same time as the story that
“the AlphaGo machine, developed by Demis Hassabis and colleagues at Google’s DeepMind subsidiary in London, triumphed over Fan Hui, the European Go champion, by five games to zero”.
DeepMind was a UK start-up which was bought up by Google last year. Google can easily make such acquisitions from its cash mountain held offshore, which reports suggest has grown to over $43b (£30b). So Google’s enormous revenues, substantially swollen by tax avoidance, are not just used to finance internal research initiatives, but to buy up smaller and often more innovative research groups.
Some have speculated that DeepMind’s success might mean that Google could pay more UK tax in the future. This would depend, first on whether DeepMind’s success can be tied to an identifiable revenue stream, and secondly on whether the “control”, “direction” and “management” functions can be considered to take place in the UK. Both of these seem doubtful. Google is likely to argue that its main revenue flows, which come from advertising sales, derive from the algorithm behind its search engine, rights to which were transferred to its affiliate formed in Ireland but resident in Bermuda. Any enhancements to the software resulting from the R&D, such as that done by DeepMind, could be attributable mainly to the “control” functions, which now presumably take place in Mountain View, California.
This could of course be arguable. HMRC could say that significant control takes place in the UK. It could also argue that important value is added in the UK. This applies also to the “marketing” function. The customer lists and other data which help drive advertising sales are also important intangibles, and must result to a significant extent from the work of its UK employees. The “settlement” (or deal) it reached suggest that HMRC shrank from taking an aggressive position towards Google, which could have led to an open legal conflict with Google, and also a more secret political conflict with the US government.
The more important point is one we have often stressed: the false assumptions behind the “independent entity” principle. No doubt much “cool stuff” takes place in Mountain View, but Google is now a global organisation. Enormous value is contributed around the world, not only by Google’s own employees, but also its users. It is more like an organism than a top-down organisation. That is why, as we have often argued, the only viable solution is to treat firms such as Google in accordance with the economic reality that they are unitary firms. The allocation of its tax base should depend on a combination of payroll costs (production factors) and sales (consumption factors).
TJN would add a couple of points. Google doesn’t automatically have to pay little or no tax in the UK. There is nothing forcing Google to set up abusive structures via Bermuda and Ireland. This is a choice by Google management, to be aggressive (and we’ll leave aside the economic illiteracy of Google’s management, for now). This is not a value-free choice. It has moral and ethical dimensions. Google’s earlier ‘don’t be evil’ adage would clearly be violated here.
But this is in the context of a system that does indeed make it easy for the likes of Google to game the system. And, as Picciotto notes, the system is broken, and needs radical overhaul.