Comments on OECD Public Discussion Draft on BEPS Action 7: Additional Guidance on the Attribution of Profits to Permanent Establishments

Brussels, 5 September 2016

Submitted by email to: TransferPricing@oecd.org

Dear Mr VanderWolk,

EBIT is grateful for this opportunity to comment on the OECD Public Discussion Draft on BEPS Action 7: Additional Guidance on the Attribution of Profits to Permanent Establishments (the “Discussion Draft”) dated 5 July 2016.

EBIT has a number of general concerns with the Discussion Draft which are set out below.

  • EBIT welcomes the recognition in the Discussion Draft that there may be PEs, especially under the lower thresholds recommended in Action 7, to which little or no profit should be attributed.
  • There has been to date little consistency of approach by tax authorities in relation to attribution of profit. There are different interpretations of the authorised OECD approach (AOA) and many authorities adopt methods which could not be described as falling within the AOA. Additional clarity in the AOA and greater and more consistent adoption by tax authorities would ease the compliance burden considerably. Where tax authorities do not follow the AOA, fees for intangibles like know-how and software might not be deductible in the PE and that no margin would typically be allowed on fees for services to the PE.
  • The processes for assessing transfer prices under Article 9 and PE profit attribution under Article 7 would be more greatly coordinated in the case of a dependent agent PE (DAPE) if the compliance activities of the agent and the PE were either handled by the same tax authority compliance officer/ team or there were active collaboration between respective officers/ teams resulting, where necessary, in a single tax audit. EBIT has experienced separate enquiries and audits being carried out by different officers/ teams putting forward conflicting arguments. In one case a transfer pricing audit of a subsidiary providing local marketing and support services to its head office concluded that the service fee was not arm’s length. Additional income was subsequently attributed to the subsidiary by the Tax Officer. Independent of the transfer pricing audit the Tax Authority deemed the subsidiary to constitute a DAPE of the head office and apportioned some of the profits on local sales derived by head office to the DAPE. As the two audits were carried out in isolation, profit apportionment to the DAPE gave no consideration to the additional service fee income apportioned to the subsidiary following the transfer pricing audit. This resulted in the same profits being taxed twice in the subsidiary country with no permitted credit in the head office’s country of residence.
  • We do generally agree that the Article 9 analysis should be carried out first and then the Article 7 analysis should follow, but in practical terms this has not always taken place. That is also vital to the proper operation of intra-country transfer pricing arrangements and transactional taxes like VAT as well as cross-border withholding taxes.
  • Having a PE with a zero profit attribution is not the same as not having a PE. It can, for example, mean we that companies would need to sort out payroll withholding from day one for business visitors, and in some territories there would be a presumption that a fixed establishment for VAT or other taxes had been created. Therefore, we would urge the OECD to discourage territories from speculatively claiming the existence of a PE where the ultimate analysis is likely to lead to a low or zero profit attribution to the PE. As the OECD recognises elsewhere, the recording and reporting of the activities of a PE creates administrative cost for both MNEs and tax administrations: this is particularly burdensome where the profit attributable to the PE is zero or very small. Consideration should therefore be given to establishing a de minimis threshold below which the activity of a PE would not need to be reported, and the related administrative burdens and exposure to other taxes and reporting obligations avoided.
  • This is particularly important where a PE would not have arisen but for a broad interpretation of the new anti-fragmentation rules. For example, new paragraph 22.3 of the commentary on Article 5 (as set out in the 2015 Final Report on Action 7) gives the example of an independent logistics company in State S, operating a warehouse in which it maintains a stock of goods belonging to an enterprise of State R. The commentary states that in some circumstances the enterprise of State R may have a sufficient level of access to the building for the purposes of inspection that this constitutes a fixed place of business of that enterprise. That right of inspection may be required for regulatory reasons. In most commercial situations – particularly where facts differ from the online retailer example in new paragraph 22 - the enterprise of State R could analyse those activities as preparatory or auxiliary, such that subparagraph 4(b) of Article 5 exempts it from PE status. However, Example B of new paragraph 30.4 of the Commentary appears to state that if a subsidiary of RCo is distributing products in State S, and taking delivery from the warehouse, then the exceptions of paragraph 4 cannot apply – and a PE exists regardless of how minimal the activities are. This seems to apply even in the present case where (unlike in paragraph 30.4) RCo does not own or operate the warehouse. The Discussion Draft suggests – rightly in our view – that zero profit would be allocated to that PE. This is by analogy to Scenario C, where a return commensurate with economic ownership of the warehouse is contemplated, noting that in the present case the warehouse is not owned by RCo. Regardless of this, however, PE obligations would have been created.
  • The practical implications of setting up a branch outside the territory of residence of the head office, which may be mandatory in the overseas territory when a PE is created, should not be underestimated. Registration and filing/ reporting requirements are often quite extensive. In the case of a dependent agent PE (DAPE), the position could be greatly simplified if the dependent agent were able to fulfil these administrative requirements on behalf of the PE and in the course of its own tax compliance. The information could then be shared with the head office’s tax authority under the relevant treaty or tax information exchange agreement, etc.

More specifically in relation to the five examples included in the Discussion Draft:

  • It is important to note that the overly simplistic ‘binary’ examples set out in the Discussion Draft do not reflect the reality of what EBIT members have experienced. Many situations are far more complex, with an entity potentially have a large number of PEs and a particular dealing constituting a PE of more than one entity.
  • A clear distinction should be made in the case of different kinds of warehousing. Many businesses will need to use warehouses for a range of purposes relating to a wide variety of inventory risk issues. Where a warehouse does in these circumstances constitute a PE (or contribute to the recognition of an operation as a PE) under the revised thresholds, it is a very different profit attribution scenario from that applicable to the type of online retailer often referred to in relation to the ‘preparatory and auxiliary’ versus core activity dispute. As in Example 1 were there to be a warehouse in Country B, the value attributable to warehousing for businesses other than online retailers would tend to be considerably lower (the position is further considered in Example 5).
  • Further, Example 1 illustrates the recognition of a balancing figure for costs where sales revenue is allocated to the DAPE. This is the only way to arrive at the appropriate attribution of profit for the DAPE, but is a problem that we have experienced with some tax authorities which don’t have a concept of using such a balancing figure (preferring an allocation of local source income minus deductible expenses).
  • Example 2 illustrates the situation in which an associated person acts on behalf of the non-resident enterprise and performs control functions related to risks contractually assumed by the non-resident enterprise. We would not generally expect an associated sales agent to control many functions without oversight on a regional or global basis.
  • In Example 3, the non-resident enterprise acting as a principal sends an employee to the host country to perform activities that give rise to a DAPE. We find it a reasonable illustration of the situation in which the functional and factual analysis might attribute to the DAPE: inventory risk and receivables risk (significant people functions or SPFs performed by the employee), economic ownership of the company vehicle used (place of use and paragraph 75 of the 2010 Attribution of Profits Report) and capital to fund the risks and assets attributed to the DAPE.
  • In Example 4, both the dependent agent and the head office have capacity to decide credit terms with customers and both actually perform the decision-making functions. The agent actively manages the recovery of customer receivables resulting from the decisions to extend credit, but significant decisions on account management are taken by head office (which contractually bears the bad debt risk, exercises control over the risk and has the financial capacity to assume the risk). On a functional analysis the attribution of risks and economic ownership of assets to the DAPE based on the SPFs undertaken by the agent on behalf of the head office (but also the head office itself) includes a proportion of the decisions that lead to the assumption of credit risk together with economic ownership of receivables. That is, in the context of the Article 7 analysis, there are considered to be split SPFs and so there is then a split of return on the receivables and then also on the fees paid. We believe that splitting SPFs can lead to significant complexity.
  • Example 5 distinguishes the functioning of a warehouse in which scenarios B and C are most relevant to EBIT (Scenario A relating to ownership of inventory by customers). In Scenario B, in which the agent’s employees operate the warehouse, a small return on inventory may be appropriate as noted above but additional clarity would be welcome to confirm that it is not assumed that there would be any allocation of customer revenues to the DAPE (a service fee would be more likely). In Scenario C, where the operation of the warehouse is carried out by a third party, if there is a DAPE it has no people and the profit attributable is likely to be zero on the basis of a risk free return and payment for funding costs and investment management services.

EBIT trusts that the above comments are helpful and will be taken into account by the OECD in finalising its work in this important area. EBIT is committed to constructive dialogue with the OECD and is always happy to discuss.

Yours sincerely,

European Business Initiative on Taxation – September 2016