Garron and Antle — Federal Court of Appeal decisions
Canadian Tax Alert, December 13, 2010
Please download the PDF version below.
The Federal Court of Appeal (FCA) recently released its decisions in two cases dealing with the residency and validity of Barbados trusts, Garron and Antle, in which it dismissed the taxpayers’ appeals and affirmed the decisions of the Tax Court of Canada (Tax Court). These two decisions are significant as they may challenge the validity of existing tax planning involving trusts.
In Garron, the more significant of the two decisions, the FCA ruled that the residence of a trust is not a legal test decided solely on the basis of the residency of the majority of the trustees as suggested by Thibodeau and that, where a question arises with respect to the residence of a trust for tax purposes, it is appropriate to undertake a fact-driven analysis with a view to determining the central management and control of the trust. In Antle, the FCA agreed with the Tax Court that no valid trust was created at law, but went on to state that the trust itself was also a sham, contrary to the Tax Court’s decision.
Mr. Garron and his business partner, Mr. Dunin, both Canadian residents, owned and operated a profitable Canadian company (OPCo). As part of a 1998 reorganization, Mr. Garron and Mr. Dunin each incorporated holding companies and established irrevocable, discretionary family trusts for the benefit of themselves and their families (the Garron Trust and the Dunin Trust, collectively, the Trusts).
Each of the Trusts was settled by an individual residing on the island of St. Vincent, and had a corporate trustee, St. Michael Trust Corp., resident in Barbados. Each of the Garron Trust and the Dunin Trust subscribed for common shares of the respective holding company of the appellants. The holding companies, in turn, subscribed for new common shares of OPCo.
In 2000, the business was sold to a private equity fund, which purchased the shares of the holding companies owned by the Barbados Trusts, resulting in a capital gain of approximately $450,000,000.
The applicable non-resident withholding tax, under section 116 of the Income Tax Act (Canada) (the Act) was remitted to the Canada Revenue Agency (CRA). The Trusts then filed Canadian income tax returns requesting a refund of the tax so withheld, on the basis that they were residents of Barbados and therefore exempt from Canadian income tax under the Canada–Barbados Tax Treaty (the Treaty).
The following issues were raised in the appeals:
- Were the Trusts resident in Canada under general legal principles?
- If the Trusts were not residents of Canada under general legal principles, were the Trusts deemed to be residents of Canada under subsection 94(1) of the Act?
- Even if the Trusts were deemed to be resident in Canada by virtue of section 94 of the Act, were they nevertheless residents of Barbados and not residents of Canada for purposes of the Treaty and thus entitled to the benefit of the exemption in paragraph 4 of Article XIV of the Treaty?
- If the Trusts were not residents of Canada and were residents of Barbados for purposes of the Treaty, would the general anti-avoidance rule (GAAR) apply to tax the capital gains in Canada?
Residence of the Trusts under general legal principles
The Tax Court held that the Trusts were residents of Canada on the basis that the management and control of the Trusts were exercised in Canada. The taxpayers appealed on the basis that the Tax Court “made a palpable and overriding error in concluding that the Trusts were controlled and managed by anyone other than St. Michael Trust Corp.” and that Justice Woods erred in law in applying a management and control test to determine the residency of a trust and, in the alternative, that if the management and control test was correct, the management and control of the Trusts was exercised by the Barbados trustees.
The FCA reviewed the case law and made a number of observations. Generally, the question of residence of a person is fundamentally a question of fact requiring consideration of a number of factors that point to or away from an economic or social link between the person and a particular country. As to the residence of a corporation, case law has established that the residence of a corporation is primarily determined based on the location of the corporation’s central management and control, which is a question of fact. With respect to trusts, however, there is little jurisprudence. The case law does not establish a single test for determining the residence of a trust and, in particular, the case law does not support the conclusion that the residence of the trustee is an invariable legal test for the residence of the trust. Nor has the case law rejected conclusively the central management and control test as an appropriate legal test for determining the residence of a trust. The FCA agreed with the Tax Court’s conclusion that Thibodeau was not a blanket rejection of the management and control test as the appropriate test of residence for a trust.
St. Michael Trust Corp. argued that the test of central management and control cannot be applied to a trust because a trust is a legal relationship without separate legal personality. The FCA concluded that although it is true as a matter of law that a trust is not a person, given that a trust is treated as a person for income tax purposes, it is not inconsistent to recognize that the residence of a trust may not always be determined by the residence of its trustee. The FCA also opined that section 104 of the Act, which provides that the trust is embodied as a taxpayer in the person of the trustee, was enacted to solve the practical problems of tax administration and not to reflect an intention of Parliament that the residence of a trust must in all cases be the residence of the trustee.
The Tax Court had concluded that St. Michael Trust Corp. was not exercising the main powers and discretions as the trustee under the trust indentures. Rather, its true role was “to execute documents as required, and to provide incidental administrative services”. The Tax Court found that substantive decisions respecting the Trusts were made by the Canadian-resident beneficiaries. The FCA noted that some of the factors considered by the Tax Court were neutral and common to many ordinary trusts, such as the fact that the beneficiaries had the right to appoint a protector who had the power to replace the trustee, or the fact that the beneficiaries and the trustee employed common advisors. However, the cumulative body of evidence led to the conclusion that management and control was not exercised in Barbados. There could be no other explanation for the lack of documentary or other evidence showing that the Barbados trustee took an active role in assessing any of the important decisions relating to the disposition of the trust property.
Regarding the factors considered by the Tax Court in finding that the Trusts were managed and controlled in Canada, the FCA stated as follows:
However, there is a line to be drawn. On one side of the line are recommendations, even strong ones, by the beneficiaries to the trustee, leaving the trustee free to decide how to exercise the powers and discretions under the trust. In that case, the trustee is still managing and controlling the trust. On the other side of the line the beneficiaries are really exercising the powers and discretions under the trusts, managing and controlling the trusts, and displacing the appointed trustee. As mentioned above, on which side of the line a case falls is a factual question, requiring consideration of the evidence in its totality.
While the FCA’s conclusion that the Trusts were resident in Canada on the basis of general legal principles was sufficient to dispose of the appeal, the FCA went on to comment on the other issues argued in the appeal.
Deemed residence under Subsection 94(1) of the Act
Section 94 of the Act deems non-resident trusts to be resident of Canada for purposes of Part I tax. To be deemed a resident, paragraph, 94(1)(b) of the Act sets out the contribution test, which will be met if the non-resident trust acquires property “directly or indirectly in any manner whatever” from a Canadian-resident beneficiary or person related to that beneficiary. The Tax Court concluded that there had not been a transfer of property directly or indirectly to the Trusts because this would require “looking through” the holding companies of Mr. Garron and Mr. Dunin to find that the Trusts were the recipients of the transferred property. The Tax Court was uncomfortable in adopting such a broad interpretation of the term “directly or indirectly” due to the potential number of persons that could be caught.
However, the FCA disagreed with the Tax Court’s opinion on this point. A crucial fact in reaching this determination was the change in value of the common shares of OPCo as a consequence of the 1998 reorganization. At the Tax Court, the Minister argued that the common shares were worth substantially more than the preferred shares issued on the reorganization. The Tax Court agreed with the Minister but did not find it necessary to actually establish the value of the common shares. Consequently, the FCA accepted as fact that the common shares were worth substantially more than the pre-reorganization value attributed to them.
Kieboom has long established that the value of a corporation is represented by its shares, which are property, and that the allocation of value between shares of different classes is determined by their terms and conditions. Therefore, it is possible to shift value from one class of shares to another by changing their terms and conditions. The FCA held that the reorganization shifted value from one shareholder to another since the pre-reorganization value of the common shares was substantially higher than the redemption value of the preferred shares for which they were exchanged. As a consequence, value accrued to the new common shares issued to the new holding companies, and this value was in turn transferred to the Trusts.
Further, the FCA did not share the same concerns about looking through the holding companies as the Tax Court did:
Once it is accepted that an indirect transfer of the shares of a corporation from Shareholder A to Shareholder B can be achieved by a corporate reorganization that shifts part of the value to the corporation from the class of shares owned by Shareholder A to the class of shares owned by Shareholder B, I see no principled basis for concluding that the same transaction cannot also be an indirect transfer of property “in any manner whatever” to the person that owns Shareholder B. This does not imply any change to the legal principle that the property of a corporation is not the property of the shareholders. It merely recognizes the fact that any increase in the wealth of Shareholder B will necessarily increase the wealth of whoever owns Shareholder B.
It was the FCA’s opinion that the words “directly or indirectly in any manner whatever” were deliberately chosen by Parliament to capture every possible means by which wealth and income-earning potential represented by the shares of a Canadian corporation could be transferred to a non-resident trust from a Canadian-resident beneficiary.
The next issue considered was whether a trust deemed resident in Canada pursuant to subsection 94(1) of the Act is precluded from claiming an exemption from Canadian tax by virtue of paragraph 4 of Article XIV of the Treaty. In Crown Forest, the common element in the treaty definition of residence is that it allows a state to impose full tax liability on worldwide income. Pursuant to Crown Forest, a taxpayer who is liable to tax in Canada only on income from a source as opposed to its worldwide income would not be considered a resident of that state for treaty purposes. Applying Crown Forest, the Tax Court concluded that since a deemed resident trust is not taxable on its worldwide income, but is only taxable in Canada for the purposes of Part I of the Act — which is relevant to the determination of its Canadian-source income and its foreign accrual property income — the Trusts would not be considered residents of Canada for purposes of the Treaty. The FCA agreed with the Tax Court on this issue.
The last issue considered was the application of the GAAR in section 245 of the Act. Assuming that the residence of the Trusts was to be determined only on the basis of the residence of the trustees, such that the Trusts would be resident of Barbados for Treaty purposes, would the GAAR apply to deny the Treaty benefits?
The Tax Court concluded that the series of transactions that resulted in the Trusts becoming entitled to the Treaty exemption in the face of section 94 of the Act was not a misuse or abuse of the Treaty and therefore the GAAR did not apply. The FCA agreed.
The FCA noted that the Trusts fall squarely within the non-resident trust rules and are subject to taxation under the provisions of section 94 of the Act. The Trusts did not avoid being taxed under section 94 of the Act but rather were entitled to an exemption under the Treaty, in which Canada has agreed not to tax certain capital gains realized by residents of Barbados.
As noted above, the FCA also recently released its decision in Antle. As was previously discussed in the September 2009 Tax Alert, this case concerned the validity of a trust also resident in Barbados. Rather than dealing with residency, Antle was concerned with whether a valid trust was in fact created in the circumstances.
On the advice of his tax advisors, Mr. Antle decided to implement a tax planning strategy known as a “capital step-up strategy”, to avoid paying Canadian income tax on the accumulated gain arising from the sale of certain common shares in a private Canadian corporation (the “subject shares”). He settled a qualifying spousal trust for the benefit of his wife; selecting a trustee resident in Barbados (the “Trust”). He then gifted the subject shares to the Trust, a transaction that occurred on a tax-deferred “rollover” basis. The Trust then sold the subject shares to Mrs. Antle at fair market value and she then sold the subject shares to an arm’s-length purchaser. Shortly thereafter, the Trust was wound up. As a result, the cost base of the subject shares was stepped up to fair market value prior to the ultimate sale so no capital gains would be taxable in Canada. Further no tax was payable by the Trust in Barbados. This series of transactions was to take place on the same day so that the proceeds of disposition would find their way back into Mr. Antle’s newly incorporated business the next day by way of a loan from his wife.
The Minister reassessed Mr. Antle to include in his income the taxable capital gain arising from the sale of the subject shares, arguing that the Trust was not validly constituted or, if found to be validly constituted, was a sham.
The Tax Court found that the Trust was not validly constituted because it lacked certainty of intention and certainty of subject matter. In any event, the Tax Court found that no transfer of the subject shares to the Trust had taken place; there were many inconsistencies and inaccuracies in the documentation and mechanics of the transaction with the result that the subject shares were never effectively transferred. The Tax Court also found that the surrounding circumstances could not support a conclusion that signing the trust deed reflected any true intention to settle a discretionary trust, regardless of the clarity of the language in the trust deed itself. Finally, the Tax Court also found, in obiter, that the series of transactions were abusive of the Act and the Treaty, although Justice Miller did not find that the Trust was a sham.
Issues raised in the appeal:
- Did the Tax Court err in law in determining that the Trust had been invalidly constituted?
- Were the subject transactions subject to the GAAR?
Validity of the Trust
On appeal, Mr. Antle argued that the Tax Court applied the wrong legal test to determine if the Trust had been validly constituted and maintained that the trust agreement was clear and unambiguous in its intent. In addition, the appellant argued that the Tax Court made an error in law in making its determination on the validity of the Trust by looking at circumstances external to the trust agreement itself. The FCA rejected that argument and confirmed that a finding of whether or not a trust has been created is to be made on the facts of the case, as evidenced by both the documents and the actions of the parties. The FCA upheld the Tax Court’s finding that the Trust had not been validly constituted.
In Antle, the Tax Court expressed the view, in obiter, that notwithstanding the deficiencies in settling the Trust, it was not a sham because there was no intentional deceit on the part of Mr. Antle and the trustee. Even though the FCA agreed with the Tax Court’s finding that the Trust was not validly constituted, and it was therefore not necessary to consider the issue of whether or not the Trust itself was a sham, the FCA made a number of important statements in obiter on this issue. Justice Noel concluded that the Tax Court judge was wrong in finding that there was no intentional deception and therefore the Trust was not a sham. In fact, the FCA stated that Justice Miller of the Tax Court misconstrued the notion of intentional deception in the context of a sham. In explaining the test for sham, Justice Noel wrote that when the courts speak of an intentional deception as a necessary condition for a sham, the required intent or state of mind is not equivalent to criminal mens rea or even the tort of deceit. It suffices that parties to a transaction present it as being different from what they know it to be.
The FCA noted that the Tax Court had found as a fact that both Mr. Antle and the trustee knew with absolute certainty that the trustee had no discretion or control over the subject shares and yet they both signed a document stating the opposite. The Tax Court judge nevertheless held that they did not have the requisite intention to deceive.
Giving a false impression of the rights and obligations created between Mr. Antle and the trustee was sufficient to hold that the Trust was a sham. The FCA stated as follows:
When regard is had to the reasons as a whole, it is apparent that the only reason why the Tax Court judge reached the conclusion that he did is his finding that the appellant and the trustee – as well as the participants in the plan – could say “with some legitimacy” that they believed that the trustee had discretion over the shares (Reasons, para. 71). While the claim to “some legitimacy may show that there was no criminal intent to deceive (as would be required in a prosecution pursuant to subsection 239(1) of the Act) and perhaps no tortuous deceit, it does not detract from the rights and obligations created between them. Nothing more was required in order to hold that the Trust was a sham.
It follows that it was the parties’ intention at the time the document was created that is relevant to the analysis of whether there was a sham.
The FCA did not comment on the GAAR.
The decisions in Garron and Antle have very important implications for taxpayers and their advisors, notably in relation to tax planning involving the use of trusts and the interaction of tax treaties and the Act.
Implications of Garron
This decision has important implications for all taxpayers who have discretionary trusts. The FCA and the Tax Court agreed on every issue except whether an indirect transfer of property had occurred for the purposes of subsection 94(1).
Residence of trusts
The FCA’s approval of the central management and control test used by the Tax Court clearly establishes that the residency of the trustee does not determine the residence of a trust. While the decision does not override Thibodeau or state that the case is incorrect, one must consider where management and control is exercised.
Thus, the test for the determination of the residence of a trust has now become a multi-step approach based on the particular facts and is no longer a single legal test. Due diligence should be taken when examining and determining the residence of a trust since the application of the test is fact specific and no two trusts will have the same fact patterns. While larger trusts will typically have active trustees, for smaller trusts decisions may continue to be made by the contributor because the cost of having an active trustee may be too high.
This decision is not limited to non-resident trusts, but will have an impact on current interprovincial trusts and all future tax planning involving trusts. The CRA has long held that central management and control is the proper test for determining the residence of a trust and will be quick to examine any interprovincial trusts to ensure their legitimacy.
The most important implication of this decision is that for the purpose of determining the residence of a trust, unlike individuals and corporations, little guidance has been provided by either the CRA or the FCA. With respect to individuals, a guide has been published that indicates the factors that the CRA will consider when making their determination. Garron suggests that, depending upon the facts, a trust may be resident in the jurisdiction in which the contributor or beneficiary resides on the basis that the central management and control does not in fact rest with the trustee. It remains to be seen how the case law applying the management and control test in connection with the residence of a corporation will be applied to trusts.
Justice Sharlow, in obiter, greatly increased the scope of the term, “directly or indirectly in any manner whatever”, to a point that most transfers into a trust could become caught; with the result that many trusts can now be deemed residents of Canada. In the present case, had the Trusts in fact been resident in Barbados they would have been entitled to the benefits of the Treaty. The Trusts would have been able to use to the Treaty to avoid any capital gains taxable under the deemed residency provision.
However, this exemption from Canadian taxation is not available under the proposed legislation in connection with non-resident trusts that will be retroactive to 2007. Under the proposed legislation, a deemed resident under section 94 of the Act would be considered a resident of Canada for Treaty purposes, as the CRA has proposed for it to be listed in the Income Tax Conventions Interpretation Act. Furthermore, should the trust be resident of another country as well, it would not be eligible for the Treaty tie-breaker rules as the CRA views only natural persons as eligible for such tie-breaker rules. In the event that a trust is entitled to the tie-breaker rules, the rules outlined would be of no use as they were designed to determine the residence of natural persons, mentioning factors such as permanent home and habitual abode amongst others, which clearly have no applicability to trusts.
The FCA agreed with the decision of the Tax Court that the series of transactions that resulted in the Trusts becoming entitled to the Treaty exemption in the face of section 94 was not a misuse or abuse of the Treaty.
Implications of Antle
The FCA upheld the finding of the Tax Court with respect to the factors that can be considered when determining whether a trust has been validly created. More significantly, the FCA made new and far-reaching comments about the doctrine of sham. Prior to this decision, the concept of a sham has been interpreted to mean that the parties’ true intent is that others shall be misled by the terms appearing in the transactional instrument. In the context of a trust, the settlor’s true intent is to retain control of the asset even though the trust instrument is intended to convey the impression that there has been a transfer of assets. The phrase “intentional deception” as used in the caselaw connoted something akin to fraud and, therefore, most tax practitioners likely thought that a sham was meant for more egregious situations than merely giving a false impression. While the FCA would say it is merely clarifying the law, one result of the Antle decision is that the threshold for a finding that a sham exists is now quite low; certainly lower than previously thought.
Where a corporation exercises the management and control of a trust, the residence of that corporation is determined based on the normal factual tests for determining residence of a corporation. An exception to the general rule may be encountered where the management and control of a trust is exercised by a branch office, for example, a branch of a trust company. In these circumstances, the trust may be determined resident in the jurisdiction where the branch office is located even though the corporation itself is resident outside that jurisdiction.
This publication is produced by Deloitte & Touche LLP as an information service to clients and friends of the firm, and is not intended to substitute for competent professional advice. No action should be initiated without consulting your professional advisors. Your use of this document is at your own risk.