In recent years, the Canadian approach to permanent establishment (“PE”) determinations has become a source of confusion and uncertainty for Canadian and non-resident taxpayers alike. Due to some debatable audit activity by the Canada Revenue Agency (“CRA”) and other tax administrations around the world, the potential existence of a PE has quickly become, along with transfer pricing, one of the biggest tax concerns among corporations today. The most alarming trend we have seen with CRA auditors is where they have made PE determinations on foreign companies that have structured a portion of their operations in Canada through a wholly-owned subsidiary acting as a captive service provider. These assessments are especially controversial in situations where the wholly-owned subsidiary does not habitually conclude contracts on behalf of its foreign parent. However, the tide appears to have finally turned in Canada, following recent guidance from the Tax Court of Canada (“Tax Court”) and CRA’s Income Tax Rulings Directorate (“CRA Rulings”).
General PE Principles
If a non-resident carries on business in Canada, the non-resident is liable for tax in Canada pursuant to subsection 2(3) of the Income Tax Act (“Act”), subject to the application of a relevant treaty. Generally, Article 7 of Canada’s treaties requires the business to be conducted through a PE in Canada in order for Canada to impose tax on Canadian source income. The threshold for a non-resident of Canada to be considered to carry on business in Canada is easily met. Therefore, for all intents and purposes, it is the treaty relief provided by the PE rule in Article 7 of Canada’s tax treaties that is of critical importance to foreign entities expanding their operations to Canada.
The official CRA views on PE determinations are generally consistent with the Organisation for Economic Co-operation and Development’s (“OECD”) Model Tax Convention on Income and on Capital (“OECD Model Treaty”) and its Commentary (“OECD Commentary”) on Article 5, which, as per the general rule in paragraph 1 of Article 5 (“General PE Rule”), establishes three key conditions for the finding of a PE:
- There must be a place of business (“1st PE Criterion”).
- The place of business must be fixed (“2nd PE Criterion”).
- The non-resident must be carrying on business wholly or partly through this fixed place of business (“3rd PE Criterion”).
In structures where the foreign entity has established a subsidiary in Canada, the General PE Rule analysis is generally the first prong in a two-pronged analysis to a PE determination. If the General PE Rule does not result in a finding of a PE, then the analysis generally turns to the dependent agent PE rule found in paragraph 5 of Article 5 of Canada’s tax treaties (“Dependent Agent PE Rule”). The key factors in the Dependent Agent PE Rule are:
- An agent's authority to conclude contracts in Canada; these contracts must relate to operations that constitute the business proper of the non-resident;
- The agent habitually exercises such authority to conclude contracts;
- Whether the agent is of dependent status, both legally and economically; and
- Whether the dependent agent is acting in the ordinary course of his or her business.
There is significant ambiguity within the text of Article 5 of the OECD Model Treaty as well as the OECD Commentary regarding scenarios where parent companies establish wholly-owned subsidiaries in Canada. In the absence of formal interpretive policies by CRA along with limited Canadian jurisprudence on the issue, CRA auditors have had some flexibility in assessing parent companies as having a PE in Canada where they have felt the circumstances supported it. The technical basis for PE determinations by CRA involving wholly-owned subsidiaries is not always clear, but commonly involves some variation of the “space at the disposal” argument, the “agency” argument, the “place of management” argument, or a combination thereof.
Space at the Disposal
It is generally accepted that a non-resident of Canada does not need to own or lease its own facilities in Canada to meet the 1st PE Criterion. This criterion can also be met in circumstances where space is made available to the non-resident at, for example, its client’s or wholly-owned subsidiary’s facilities in Canada. CRA generally abides by the OECD Commentary to Article 5 with respect to what constitutes “space at the disposal” in determining whether a non-resident entity has a PE in Canada. Generally, this means that CRA will not pierce the corporate veil and automatically consider the premises of the wholly-owned subsidiary to be at the disposal of the parent corporation. However, if the employees of the parent corporation have a significant presence in Canada (e.g., they spend significant time in Canada) at the premises of its Canadian subsidiary, a PE may be determined to exist under the General PE Rule.
Consider the situation where a non-resident corporation does not directly own or lease a place of business in Canada or have space at its disposal at another entity’s Canadian facilities in which the non-resident’s employees are spending significant time. Based on these facts, once it is determined that there is no dependent agent in Canada, or that the dependent agent does not habitually conclude contracts, one would assume that there is generally no need for a tax administration to pursue a PE analysis further. However, recent audit activity shows that CRA’s PE assessments on non-resident parent corporations that have expanded their business operations in Canada through a subsidiary are not limited to those dependent agent scenarios where the agent (i.e., subsidiary) habitually exercises its authority to conclude contracts on behalf of the parent company. Instead, where the purported agent (i.e., subsidiary) does not have the authority to conclude contracts, CRA auditors have occasionally relied on what appears to be a combination of Canadian “agency law” and a contentious interpretation of paragraph 10 of the OECD Commentary to Article 5 to support a PE determination. Accordingly, it is not uncommon for parent corporations with subsidiaries in Canada to be challenged by CRA on PE issues using some variation or combination of the space-at-the-disposal argument and the agency argument, even when its Canadian subsidiary is not empowered to conclude contracts on behalf of the parent company.
Place of Management
CRA has raised PE assessments on non-resident corporations on the basis that the Canadian subsidiary constitutes a “place of management” of the parent corporation. This argument is often used in conjunction with the agency argument. The basis for these assessments is unclear. Nevertheless, the OECD Commentary states that those places listed in paragraph 2 of Article 5 of the OECD Model Treaty “constitute a PE only if they meet the requirements of paragraph 1.” In our experience, CRA has made such reassessments with questionable support for a PE determination under the General PE Rule.
In the case of the Canada-United States Income Tax Convention (“Canada-US Treaty”), once CRA determines that a U.S. parent corporation does not have a PE in Canada under the General PE Rule or Dependent Agent PE Rule, it will generally then review the service PE rule in Article 5(9) of the Canada-US Treaty to determine if a deemed PE exists. A discussion of the deemed service PE rule is outside the scope of this article.
On May 16, 2008, the Tax Court simultaneously released the decisions of Knights of Columbus and American Income Life Insurance Co. Both cases dealt with a fixed place of business PE determination and the Tax Court’s decision in both cases was highly dependent on whether the place of business was “at the disposal” of the enterprise. In each case, the Tax Court held that the non-resident company did not have a PE in Canada.
In Knights of Columbus, the issue of the General PE Rule was to be determined by considering the 3rd PE Criterion: whether the business of the parent was being carried on through a fixed place of business, i.e., in the context of this case, the subsidiary’s office.
The Knights of Columbus’ insurance business was handled through agents. In Canada, there were general agents who oversaw field agents. The general agents were not actively involved in selling insurance products to members of the Knights of Columbus; rather, they earned income from a commission override from the field agents and were responsible for recruiting, training and managing the field agents. The field agents did the front-line work of soliciting applications for sales of Knights of Columbus’ insurance products from members. The field agents typically conducted business from their home offices but usually met with clients at the clients’ homes. The Knights of Columbus had no access to the home office nor was there anything there that would indicate a connection with the Knights of Columbus.
Justice C.J. Miller concluded the following in response to expert testimony:
For the Field Agents' residences to be considered fixed places of business of the Knights of Columbus, the Knights of Columbus must have a right of disposition over these premises. A right of disposition is not a right of the Knights of Columbus to sell an agents' house out from under him. The Knights of Columbus might be viewed as having the agents' premises at its disposal, for example, if the Knights of Columbus paid for all expenses in connection with the premises, required that the agents have that home office and stipulate what it must contain, and further required that clients were to be met at the home office and in fact the Knights of Columbus' members were met there.
And later, at paragraph 80:
The Knights of Columbus make no operational decisions at the Field Agent's premises. The Knights of Columbus had no officers, directors or employees even visit the agents' home offices, let alone have any regular access. All risks connected with carrying on business at the home offices are borne by the agents themselves. The agents are not carrying on the Knights of Columbus' core business from these premises. Their premises cannot therefore be found to be a fixed place of business permanent establishment.
It is important to note that there was no debate as to whether the Canadian agents were, in fact, agents in accordance with common law principles. As stated earlier, CRA has made PE determinations using a combination of the space-at-the-disposal argument and the agency argument in circumstances where the status of the Canadian wholly-owned subsidiary as an agent was highly debatable.
In our experience, CRA auditors have used the following logic to determine that the premises of the subsidiary was at the disposal of the non-resident parent, even in cases where the non-resident parent was never present in Canada:
Since the Canadian subsidiary is the non-resident parent’s agent, the subsidiary’s employees who carry out the strategies and activities are carrying on the parent’s business. This suffices to constitute the premises of the subsidiary in Canada as a fixed place of business for the non-resident parent as the personnel of the subsidiary can be viewed as being those of the non-resident parent.
The Knights of Columbus decision appears to confirm that the above analysis is flawed and that the presence of an agent in a jurisdiction is not sufficient to render the agent’s premises “at the disposal” of the parent under the 1st PE Criterion or a fixed place of business through which the parent company’s business is wholly or partially carried on. Additional analysis is required.
In American Income Life Insurance, Justice Miller provided a list of factors to be considered in determining whether the parent’s business was being carried on from the fixed place of business. These factors include:
- Use of premises by parent;
- Control by parent over premises;
- Legal right to exercise control over premises;
- Degree to which premises identified with parent business;
- Who paid for expenses of premises;
- Who paid for equipment used at premises;
- Who made management decisions;
- What contracts were concluded from premises;
- What parent products were kept on premises;
- Did parent have any Canadian employees;
- Who bore the risk of the operation from premises;
- How many principals were represented by the agent; and
- Were agents subject to detailed instructions or comprehensive control.
Cause for Concern
Despite the guidance of the Tax Court in Knights of Columbus and American Income Life Insurance, CRA has occasionally maintained its aggressive approach on the PE issue. One factor that has seemingly caused CRA auditors to not follow the judgments of the Tax Court is a faulty interpretation of paragraph 10 of the OECD Commentary to Article 5 (“Paragraph 10”).
Paragraph 10 of OECD Commentary to Article 5
On occasion, CRA has relied on Paragraph 10 when making PE determinations in parent-subsidiary scenarios, particularly when that subsidiary is acting as a captive service provider to the parent.
Paragraph 10 states in part as follows:
The business of an enterprise is carried out mainly by the entrepreneur or persons who are in a paid-employment relationship with the enterprise (personnel). This personnel includes employees and other persons receiving instructions from the enterprise (e.g. dependent agents). The powers of such personnel in its relationship with third parties are irrelevant. It makes no difference whether or not the dependent agent has the authority to conclude contracts if he works at the fixed place of business… [emphasis added]
We believe the CRA is interpreting Paragraph 10 incorrectly, in a manner that allows it to consider the foreign parent to be carrying on its business in Canada through a PE at facilities owned by the purported agent.
Paragraphs 2 to 11 of the OECD Commentary on Article 5 focus on the interpretation of the General PE Rule. Paragraph 4 of the OECD Commentary on Article 5 deals with the 1st PE Criterion, while paragraphs 5 and 6 discuss the 2nd PE Criterion. Paragraphs 7 to 10 focus on the 3rd PE Criterion (the “carrying on business” condition), and it is in this context that Paragraph 10 must be interpreted. The underlined excerpt in Paragraph 10 above could lead to much confusion if not interpreted in the context of the 3rd PE Criterion of the General PE Rule. The correct interpretation of this sentence is that the authority of the dependent agent to conclude contracts is not relevant so long as the 1st PE Criterion and 2nd PE Criterion of paragraph 1 of Article 5 are present. Hence, a non-resident will be found to have a PE in a foreign country if the non-resident has a fixed place of business through which the non-resident’s employees or dependent agents are carrying on its business, and in the case of the latter, whether or not the dependent agents have the authority to conclude contracts on behalf of the non-resident.
If Paragraph 10 were to be interpreted otherwise (i.e., beyond the context of the 3rd PE Criterion), it would render paragraphs 5 and 7 of the OECD Model Treaty irrelevant. Simply stated, the fixed place of business condition must still be met and the fixed place of business must be that of the non-resident and not that of the resident entity. Once it is determined that the non-resident has a fixed place of business, then the General PE Rule will be met if employees or dependent agents perform duties at this fixed place of business.
Recent Technical Interpretation
As discussed above, CRA has occasionally relied on its characterization of a wholly-owned subsidiary as an agent of the non-resident parent to support a PE determination. What is most concerning is the apparent ease with which certain CRA auditors conclude that a parent-subsidiary or contractor-subcontractor relationship is in fact an agent-principal relationship for the purposes of Canada’s bilateral tax treaties. CRA Rulings recently released a technical interpretation which may invite CRA auditors to continue this highly alarming and increasingly common trend.
The recent technical interpretation arose when a taxpayer requested CRA’s assistance in determining whether a non-resident company had a PE in Canada as per a hypothetical fact scenario. While we concur with CRA’s conclusion that the company had a PE in Canada, we disagree with its determination that there was an agency relationship between the contractor and subcontractor, as the facts were, in our opinion, insufficient to support such a determination. Moreover, CRA Rulings could have made a PE determination on the basis of the available facts by using a purposive approach to treaty interpretation, regardless of whether an agency relationship existed.
The use of the “agency” argument in this technical interpretation—when it was arguably not necessary to achieve the desired result—adds fuel to the fire of a growing trend we have encountered in our practice. However, as discussed below, CRA Rulings recently made up for this questionable technical interpretation by issuing an historic Advance Income Tax Ruling (“ATR”) that should provide tremendous guidance to taxpayers moving forward.
Recent CRA Ruling
The first Canadian ATR (“Ruling”) on whether a non-resident corporation (“NRco”) has a PE in Canada where part of its business activities have been subcontracted to its wholly-owned subsidiary in Canada (“Canco”) was recently issued by CRA Rulings. Gowlings assisted NRco in obtaining the favourable ruling that the relocation of part of its manufacturing operations to Canada, to be operated by Canco, would not in and by itself create a PE in Canada for NRco.
The facts of the Ruling are briefly summarized as follows:
- NRco carries on a multitude of separate business divisions worldwide, which include manufacturing, processing and distribution activities.
- NRco entered into a supply agreement (“Supply Agreement”) with a Canadian client (“Client”) whereby NRco would provide parts to Client. Dealings between NRco and Client usually consist of commercial meetings, the frequency and duration of which never exceed an aggregate of 90 days over any 12-month period.
- NRco incorporated Canco as a 100% wholly-owned Canadian subsidiary. NRco and Canco entered into a service agreement (“Service Agreement”) pursuant to which Canco performed a portion of the manufacturing duties regarding the parts sold by NRco to Client. As per the Service Agreement, Canco would not in any case carry out activities such as negotiating commercial issues with customers or concluding sales contracts on behalf of NRco.
- In connection with the Service Agreement, NRco’s personnel visited Canco from time to time for business meetings and to audit and monitor Canco’s performance.
- Under the proposed transaction, NRco would modify the Service Agreement to subcontract to Canco additional activities, in connection with NRco's obligations to Client under the Supply Agreement that were initially being carried out by NRco in NRco’s country of residence. This would include the relocation of a further portion of the manufacturing operations to Canada to be performed by Canco.
The fact structure briefly described above is a fairly routine structure that, on the surface, does not appear to create a Canadian PE issue for the parent company. However, as alluded to above, transactions similar to those described in the Ruling have come under attack by CRA auditors from the perspective of PEs. In fact, it is now not uncommon for CRA to raise both a transfer pricing adjustment and a PE assessment, giving rise, effectively, to triple tax on the same economic income.
Presuming that the subsidiary in Canada is set up as an independent contractor, and that the subsidiary has no authority to conclude contracts or otherwise bind the parent as a purported agent of the parent, the CRA should only be able to rightfully establish an agency relationship in exceptional cases. Captive service providers of this nature are generally not the kind of wholly-owned subsidiary in which a tax administration would pierce the corporate veil and ignore its existence (i.e., a puppet). Justification to pierce the corporate veil is rare in Canada and is generally only accepted by the Canadian courts in cases of fraud or improper conduct. It is our opinion that CRA should only be challenging those captive service subsidiaries that are mere shell companies with little activity or minimal employees and where every aspect of their Canadian operations is controlled and approved by the parent. Unfortunately though, that is not always the case.
Despite the favourable Ruling obtained by the taxpayer in this case, structures involving non-resident parents and wholly-owned subsidiaries are not without risk. In order to obtain tax certainty, we strongly recommend that taxpayers consider filing an ATR request with respect to those proposed structures where they feel there is potential risk to a PE issue. When submitting an ATR request, the taxpayer is requesting CRA’s confirmation that a proposed transaction will result in the tax treatment anticipated by the taxpayer in its ATR request. A proposed transaction is required in order to qualify for an ATR and the ATR must be requested before the transaction begins. The required proposed transaction can come as a result of a new or amended service agreement between parent and subsidiary.
An ATR is binding provided that the statement of facts and the proposed transaction are accurate and constitute complete disclosure, and that the proposed transaction is carried out as described in the ATR request. ATRs may be issued on questions of fact scenarios, but only if it is possible to determine all the material facts and only if those facts can reasonably be expected to prevail. The ATR will remain binding so long as the facts and transactions remain unchanged. With respect to ATR requests regarding PEs, it is often possible to establish the pertinent facts, which are generally that:
- There is no agency relationship between the non-resident entity and any Canadian entity;
- The non-resident does not own or lease a place of business in Canada; and
- The non-resident’s personnel will not be present in Canada, such as at the premises of its Canadian subsidiary, for a significant period of time that would create a fixed place of business for the non-resident.
By establishing a lack of physical presence in Canada by personnel of the non-resident, the CRA should have the pertinent facts it needs to ensure that a PE will not arise due to the space-at-the-disposal element of the 1st PE Criterion.
As mentioned above, the recently issued Ruling is the first of its kind in the context of a PE determination in Canada involving a parent-subsidiary structure. For the purpose of an ATR request for a PE determination, the ruling can be limited to the impact that a particular subsidiary may have on the non-resident’s PE status in Canada. Given that a PE analysis is entirely fact dependent, the taxpayer and CRA should be in constant communication to ensure that the pertinent facts relating to the issue of the existence of a PE can be ascertained to CRA’s satisfaction in order to give a favourable ruling.
Until this Ruling, the CRA had not published any formal guidelines as to the circumstances in which it may consider whether the premises of a subsidiary can constitute a PE of its parent company. With so much uncertainty in the tax community as to whether CRA auditors will challenge these structures in the context of the parent company having a PE in Canada, this Ruling is an extremely positive step. One must remember that in the absence of jurisprudence and formal guidelines from CRA on these issues, CRA auditors may propose adjustments based on their interpretation of the law. These interpretations are not necessarily the formal views of the CRA. The recent Canadian jurisprudence and the Ruling from CRA described above are huge factors that will either prevent auditors from raising unsupportable PE assessments or strengthen a non-resident taxpayer’s chances of an expedient and successful resolution of the matter in CRA Appeals, the courts or the competent authority process.
 In most of Canada’s tax treaties, there is also a special provision that treats a building site or construction or installation project as a PE only if it lasts more than twelve months. In addition, in the Canada-United States Income Tax Convention, there is a deemed PE rule for certain service activities that extend beyond 183 days in any twelve-month period.
 American Income Life Insurance Co. v. Canada, 2008 TCC 306 at para. 35.
 Knights of Columbus v. The Queen, 2008 TCC 307
 American Income Life Insurance Co. v. Canada, 2008 TCC 306
 2010-0391541E5 E - Article V (9) of the Canada-US Tax Convention (Released April 13, 2011).
 For a detailed analysis of this CRA technical interpretation, see Jim Wilson and Pierre Alary, Deemed Services PEs - Subcontractors vs Agents - CRA Adds Fuel to the Fire, Canadian Tax @ Gowlings, May 2011.
 CRA document # 2011-0396421R3 (E).