The following article will appear in an upcoming edition of the Hamilton Law Association Law Journal.
The Tax Court of Canada recently stuck a stick into the tax law ant hill with two decisions on trusts and tax avoidance in Garron Family Trust v. The Queen, 2009 TCC 450, and Antle v. The Queen, 2009 TCC 465. Although the decisions involve trusts established in Barbados and complex questions relating to the general anti-avoidance rule and treaty-shopping, the actual bases of the decisions are relatively narrow and are of interest to all professionals who deal with trusts and tax planning in their practices.
Garron
Background
In Garron, the two Canadian-resident shareholders of a successful Canadian operating company (“PMPL”) effected an ‘estate freeze’ in 1998. They transferred their Common Shares in the capital of PMPL to Canadian holding companies. The holding companies exchanged their PMPL Common Shares for freeze shares with a total redemption amount equal to $50 million (the supposed fair market value of the shares at that time). Two other Canadian corporations subscribed for Common Shares in the capital of PMPL for a nominal amount; trusts resident in Barbados held all of the Common Shares of the latter corporations. The trustee of both trusts was a trust corporation resident in Barbados that was controlled by Price Waterhouse Barbados (later PWC Barbados). In August, 2000, an arm’s length purchase acquired all of the issued shares of the holding companies for proceeds equal to $532 million, the bulk of which was paid to the trusts. The trusts did not pay tax in Barbados on the large capital gains they realized. Naturally, they took the position that they were not subject to Canadian tax either because they were non-resident and the gains were treaty-protected under the Canada-Barbados Income Tax Agreement.
The CRA assessed the trusts on the basis that they were resident in Canada and therefore liable to tax in Canada on the capital gains they had reazlied. The CRA, among the many other arguments it advanced in support of the assessments, took the position that the central management and control of the trusts was in Canada and that, therefore, the trusts were resident in this country.
Decision
The trusts argued that they were not resident in Canada because the trustee of the trusts was a non-resident. The trusts relied on Dill and Pearman, Trustees of the Thibodeau Family Trust v. The Queen, [1978] DTC 6376 (FCTD) for the proposition that a trust is resident where the majority of its trustees reside regardless of the location of the trust’s “central management and control”.
Madame Justice Judith Woods rejected this proposition. She pointed out that Thibodeau did not “purport to state a general test of trust residence. The decision was intended to be limited to the particular facts of the case” (¶137). In any event, the judge in Thibodeau made a finding of fact to the effect that the Canadian trustee of the trust did not control it. Arguably, then, Thibodeau’s rejection of a central management and control test was obiter.
The judge in Thibodeau did reject the test however:
The judicial formula for this respecting a corporation, in my view, cannot apply to trustees because trustees cannot delegate any of their authority to co-trustees. A trustee cannot adopt a “policy of masterly inactivity” as commented upon in Underhill on the Law of Trusts and Trustees, 12th Edition, page 284; and on the evidence, none of the trustees did adopt such a policy.
Justice Woods respectfully disagreed:
[143] It may be correct that trustees would generally be in breach of fiduciary obligations if they adopted a policy of “masterly inactivity.” The difficulty that I have with applying the obiter comment in all circumstances, though, is that it presumes that trustees always comply with their fiduciary obligations.[1]
What is to be done, then, to determine the residence of a trust? Justice Woods found that there was no common law precedent that would provide useful guidance on the question, and so she held that she should apply the “central management and control test”, which is the common law test used for determining the residence of a corporation. Why?
[158] First, the basis for applying this test to corporations is equally applicable in a trust context. In one of the most quoted passages in Canadian tax jurisprudence, the reasons for adopting this test were stated by Lord Loreburn in De Beers Consolidated Mines Ltd. v. Howe, [1906] AC 455, at 458:
In applying the conception of residence to a company, we ought, I think, to proceed as nearly as we can upon the analogy of an individual. A company cannot eat or sleep, but it can keep house and do business. We ought, therefore, to see where it really keeps house and does business. […]
It remains to be considered whether the present case falls within that rule. This is a pure question of fact to be determined, not according to the construction of this or that regulation or by-law, but upon a scrutiny of the course of business and trading. [Emphasis added.]
Justice Woods also opined at ¶¶160-1 that
[A]dopting a similar test of residence for trusts and corporations promotes the important principles of consistency, predictability and fairness in the application of tax law. […]The development of a test of trust residence in Canada has been left by Parliament to the courts. If courts decide to develop a totally different test of residence for trusts than they have for corporations, there should be good reasons for doing so. I am not satisfied that there are good reasons.
After adopting this test, Justice Woods concluded that the trusts in question were really resident in Canada because that was the location of their real central management and control (see ¶¶187ff). She found that the trustee was selected for its willingness to comply with direction from the two original shareholders on key issues, which willingness could be enforced through a protector who had the ability to remove the trustee at any time. The original shareholders could replace the protector at any time.
Justice Woods made note of several other facts that did not assist the appellants. An internal memorandum prepared by the trustee seemed to contemplate a limited role for it, and Justice Woods noted that little evidence was presented that the trustee knew much about the transactions it was supposed to be implementing. She pointed out that the trustee did not appear to have any institutional expertise in managing property. Indeed, the investment advisers for the trusts were located in Canada and the investment policy appears to have been determined by the pre-freeze shareholders. The appellants provided “very little documentary evidence that St. Michael [the trustee] had any involvement in the affairs of the Trusts other than in the execution of agreements, and in administrative, accounting and tax matters” (¶206).
The oral evidence of the pre-freeze shareholders did nothing to counter the foregoing impressions because the evidence struck the judge as “disingenuous” at times. For example, one of the pre-freeze shareholders who was responsible for negotiating the sale of the shares to the arm’s-length purchaser claimed that he was indifferent about whether the non-resident trusts sold shares. The judge found this unbelievable given the difference in the tax result. She noted the following:
[229] In terms of Mr. Dunin’s other testimony, the general impression that I had was that, in key areas, Mr. Dunin’s words appeared to be so carefully chosen that his answers were often non-responsive. I had no confidence that the answers he gave provided a complete picture.
Based on the foregoing, Justice Woods concluded that the trusts that had sold shares were resident in Canada and so were subject to tax under the Income Tax Act (Canada) (the “Act”) in respect of the gains realized on the sale. She disposed of the appeals on this basis.[2]
It is important in reading Justice Woods’ decision to recognize that she made her findings about the residence of the trusts in part because the appellants failed to convince her of the opposite proposition. That is, the onus of proof on the appellants to rebut the Minister’s assumptions about the central management and control of the trusts played an important role in the case. Justice Woods drew an adverse inference from the fact that numerous individuals who might have cast light on the operations of the trusts were not called as witnesses, and it seems that the evident purpose of the whole structure—the avoidance of tax—placed an additional burden on the appellants that they did not meet.
Antle
Miller J.’s judgment in Antle begins as follows:
[2] If the capital property step-up strategy [used by the taxpayers in this case] is considered acceptable tax planning, then there would be two tiers of taxation of capital gains in Canada: one tier for those whose capital gain can justify professionals’ fees to implement the strategy, in which case there is no tax on a capital gain in Canada; the second tier for everyone else, in which case capital gains are subject to tax in accordance with Part I of the Act. This is an unacceptable result to the Respondent. The real question before me is whether it is for the legislators to introduce legislation to defeat such a result, or can existing legislation and jurisprudence be relied upon by the Courts to do so?
It was almost all downhill from there for the appellants.
How was the “capital property step-up strategy” supposed to work? Mr. Antle wished to sell his shares in the capital of a corporation to a third party. Rather than sell the shares directly and pay tax in Canada, he purported to transfer his shares to a spousal trust resident in Barbados on a rollover basis. The trust, in turn, sold the shares to Mr. Antle’s wife for a note. As a result, the non-resident trust realized a capital gain on which it did not pay tax. Mrs. Antle sold her newly-acquired shares to the third party for cash. Mrs. Antle realized no gain because “her” shares had a tax cost equal to the price payable by the third party. She used the cash proceeds to repay the note to the trust, and the trust then made a distribution of capital to Mrs. Antle. As a result, no taxpayer was required to include in income for the purposes of the Act any gain on the sale of the shares.
The Minister argued that the scheme failed to achieve its tax objectives for a number of reasons, including that the scheme constituted a sham, that it was legally ineffective and that the GAAR applied to it.
Justice Miller at ¶74 rather grudgingly agreed with counsel for the appellants that the scheme did not constitute a sham because the latter concept requires an element of deceit. Justice Miller, however, also found that the scheme failed because it was legally ineffective: no trust was created because there was no certainty of intention, no certainty of subject matter and no real discretion exercised by the “trustee” in carrying out the terms of the “trust” agreement.
Justice Miller, in arriving at these conclusions, was able to point to a number of deficiencies in the execution of the scheme. In a passage that the CRA will no doubt quote again and again to taxpayers, the Justice wrote:
This conclusion emphasizes how important it is, in implementing strategies with no purpose other than avoidance of tax, that meticulous and scrupulous regard be had to timing and execution. Backdating of documents, fuzzy intentions, lack of transfer documents, lack of discretion, lack of commercial purpose, delivery of signed documents distributing capital from the trust prior to its purported settlement, all frankly miss the mark – by a long shot. They leave an impression of elaborate window dressing. In short, if you are going to play the avoidance game, it is not enough to have brilliant strategy, you must have brilliant execution.
What were some of the deficiencies? The Justice found that Mr. Antle did not really intend to settle a discretionary trust other than as an instrument for tax avoidance and that he never properly transferred his shares to the trust. The Justice also found that Mr. Antle appeared to reserve some of his rights in the shares that he purported to transfer—Mr. Antle later successfully sued a third party as if he continued to have rights in the shares—which created doubt about whether there was certainty of subject matter. Some of the documents were backdated, and it appeared that the trustee did little if anything in connection with the trust other than sign documents as required by the scheme. In light of these findings, Justice Miller held that no trust was created and Mr. Antle was subject to tax under the Act on the sale of his shares as if he had sold them directly.
Conclusion
A casual acquaintance with Garron and Antle might induce the reader to take the former decision as the more radical of the two cases. In Garron, Justice Woods created a new common law rule for determining the residence of a trust. In Antle, Justice Miller found that the scheme before him failed merely because the pressure of events meant that its implementation was, in the Court’s opinion, sloppy. Antle admonishes tax advisers to be careful in executing their plans. Tell them something they don’t already know.[3]
The foregoing impressions might be revised by subsequent readings of the two cases. It is true that Justice Woods may have departed from existing precedent in applying the central management and control test to a trust, but the correctness of that precedent had been called into question for some time before Garron and for good reason. Cautious tax advisers (is there any other kind?) have always treated Thibodeau with circumspection, and no good tax adviser rendered an opinion on the residence of trust by paying attention only to the residence of its titular trustees, especially given the CRA’s comments on the subject in Interpretation Bulletin IT-447.[4] Seen in this light, Justice Woods’ decision in Garron does not seem so radical.
On the other hand, some of the comments made by the Court in Antle leave one with the impression that the court felt able to ignore the legal consequences of the transactions entered into by the parties merely because they were tax motivated. Such an approach would be radical. The Supreme Court of Canada, in Shell Canada Ltd. v. The Queen, [1999] 3 S.C.R. 622, 1999 CanLII 647, held that a tax court cannot disregard the legal effect of transactions just because they are tax-motivated. The court in Antle seemed to come dangerously close to saying (for example) that, just because Mr. Antle’s intention was to avoid tax, he could not have intended to settle a valid trust.
What is to be done? Tax advisers and allied professionals must proceed with renewed caution in implementing tax plans. In particular, it will be necessary to ensure that, the more aggressive the plan, the better the documentation and implementation. Moreover, it will not be enough to create all the right paper. It will be necessary as well to ensure that the substance of the transactions confirms the paper to satisfy the Tax Court that the latter is more than “window dressing”.
[1] Justice Woods also noted that a similar argument had been rejected by the Federal Court of Appeal in Robson Leather Company Ltd. v. MNR, [1977] DTC 5106.
[2] Justice Woods addressed numerous other arguments raised by both the Crown and the appellants—including arguments relating to section 94 of the Act and the GAAR—but she explicitly stated that her decision on central management and control disposed of the appeals. That is, her comments on the other issues were, strictly speaking, obiter dicta.
[3] For that matter, Garron makes the same point.
[4] See especially ¶5 of IT-447.
