Bien que la règle générale anti-évitement du Canada (la « RGAÉ ») célébre son trentième anniversaire l’an prochain, il est surprenant que certains aspects de cette règle n’aient pas encore été réglés. La décision de 2005 de la Cour suprême du Canada dans l’affaire Trustco Canada a clarifié le cadre général d’application de la RGAÉ. Il demeure toutefois encore une grande incertitude à l’égard des éléments individuels de l’analyse à trois volets que la Cour a établie dans cette affaire, ce qui fait qu’il est possible qu’il soit toujours difficile de prédire l’issue des dossiers relatifs à la RGAÉ.

Une version française de ce billet sera disponible prochainement.

With Canada’s general anti-avoidance rule (GAAR) celebrating its 30th birthday next year, it is surprising that certain basic aspects of the rule remain unsettled. The Supreme Court of Canada’s 2005 decision in Canada Trustco clarified the general framework for determining when the GAAR applies. However, there is still considerable uncertainty about the individual components of the three-part analysis that the court developed in that case, with the result that it can still be quite difficult to predict the outcome of GAAR cases.

Part of the problem – the manner in which taxpayers have chosen to argue GAAR cases – is well illustrated in the recent decision of the Tax Court in the case of Perry Wild v. Her Majesty the Queen.[1] In part because the taxpayer did not contest certain points, the court ended up making a ruling that arguably further “muddies the waters” on two key elements of the Canada Trustco analysis: whether the transaction conferred a “tax benefit” and (if so) whether it was an “avoidance transaction”.

The Facts: A Butterfly Transaction With A Twist

The facts were fairly straightforward, involving a clever twist on a standard “butterfly” transaction. Perry Wild (PW), an individual resident in Canada, owned all of the shares of P.W. Rentals Ltd. (PWR), a Canadian corporation. In an effort to protect the assets of the corporation, PW decided to have PWR transfer assets to two new corporations (the “transferee corporations”) pursuant to a double-wing butterfly. In this reorganization, PW would transfer a portion of his shares of PWR to one of the transferee corporations for shares of the transferee corporation. PWR would then transfer a portion of its assets to that transferee corporation for shares of the transferee corporation. As a result of these transfers, PWR would own shares of the transferee corporation and the transferee corporation would own shares of PWR. The cross shareholdings would have equal values and would be redeemed in exchange for notes which would then be set-off one against the other. The same steps would be taken to transfer assets from PWR to the other transferee corporation.

In a typical butterfly reorganization, on the first transfer, in this case the transfer of the shares of PWR by PW to the first transferee corporation, the transferor receives common shares of the transferee corporation. On the second transfer, in this case the transfer of the assets by PWR to the transferee corporation, the transferee corporation issues fixed value preferred shares having a value equal to the net value of the assets transferred.

The clever twist employed by PW was to have the transferee corporations issue the same class of shares on the first and the second transfers. On the first transfer, the paid-up capital (PUC) of the shares issued by the transferee corporation would be limited, pursuant to subsection 84.1(1), to the PUC of the shares of PWR that were transferred. This was a nominal amount. However, on the second transfer, subsection 85(2.1) would apply to limit the PUC of the shares of the transferee corporation issued on this transfer to the cost of the assets transferred which was a much higher number. Because the same class of shares was issued on each transfer, the higher PUC of the second tranche of shares issued on the second transfer was averaged with the PUC of the first tranche of shares issued on the first transfer. The result was that the PUC of the shares of the transferee corporations held by PW was higher than the PUC of the shares of PWR that were transferred to the transferee corporations – a result that section 84.1 is designed to avoid. In addition, because PW claimed the lifetime capital gains exemption on the transfer of the shares of PWR to the transferee corporations, the adjusted cost base of the shares of the transferee corporations held by PW was at least equal to the PUC of the shares. As a result, PW would be able to extract cash or other property from the transferee corporations having a value equal to the PUC of the shares without realizing either a deemed dividend or a capital gain. There was, however, no evidence that PW had extracted in this manner any amounts from either of the transferee corporations.

The Assessment: There was a Misuse or Abuse

The Canada Revenue Agency (CRA) assessed the transactions under GAAR alleging that the transactions resulted in a misuse or abuse of section 84.1. Because no actual amounts had been extracted from the transferee corporations, the CRA used determinations to reduce the PUC of the shares of the transferee corporations to a nominal amount.[2]

The taxpayer conceded that the series of transactions included avoidance transactions and resulted in a tax benefit. The only issue for the Tax Court was therefore whether the transactions resulted in a misuse or abuse. The Court cited the decision of Mr. Justice Hogan in Descarries, in which it was stated that “the object, spirit or purpose of section 84.1 of the Act is to prevent taxpayers from performing transactions whose goal is to strip a corporation of its surpluses tax-free through the use of a tax-exempt margin or a capital gain exemption”. The Court held that the transactions undertaken by PW frustrated the object, spirit or purpose of section 84.1. Its reasons are summarized in the following passage:

The avoidance transactions achieved an outcome, and in the context of the series of which those were part, that section 84.1 was intended to prevent (individual shareholders in a non-arm’s length share for share exchange convert corporate distributions, that would otherwise be taxed as dividends, into exempt returns of capital by utilizing the capital gains exemption) and defeated its underlying rationale – and the underlying rationale of subsection 89(1) – by misusing the PUC calculation that triggered the share averaging effect.

Our Analysis: Was Canada Trustco Correctly Applied?

Based on the framework established in Canada Trustco, the first step in applying the GAAR is to determine whether there is a tax benefit arising from a transaction or series of transactions of which the transaction is part. The second requirement for application of the GAAR is that the transaction giving rise to the tax benefit be an avoidance transaction. While there was little analysis with respect to these first two aspects of the GAAR test given the taxpayer’s concessions, it is interesting to consider these aspects of the case in light of statements made in prior cases.

The “tax benefit” issue

The Court defined the tax benefit to be the increase in the PUC of the shares of the transferee corporations. “Tax benefit” is defined in subsection 245(1) as a reduction, avoidance or deferral of tax or other amount payable under the Act or an increase in a refund of tax or other amount under the Act. It is well settled that the reference to “tax or other amount under this Act” is the object of the preposition “of” that precedes it so the benefit includes an increase in a refund of tax or an increase in the refund of some other amount under the Act, such as interest. A tax benefit does not include an increase in any amount under the Act so it would not include an increase in the PUC of shares.

This interpretation of “tax benefit” was confirmed by the Supreme Court of Canada in Copthorne in the following passage:

In the agreed statement of facts, the parties agreed that the sale of VHHC Holdings to Big City, and the subsequent amalgamation of VHHC Holdings with Copthorne I to form Copthorne II were part of a series of transactions. However, these transactions themselves did not result in a tax benefit. The tax benefit was only realized when Copthorne III redeemed its shares without its shareholder incurring an immediate tax liability.

In Copthorne, the sale of VHHC Holdings to Big City and subsequent amalgamation of VHHC Holdings with Copthorne I increased the PUC of the amalgamated entity but that increase was not the tax benefit. As the Court noted, the benefit did not arise until shares were redeemed thus utilizing the PUC so created to avoid withholding tax on the deemed dividend. As noted, in the Wild case, there is no evidence that the PUC created on the butterfly was ever used. Therefore, the transactions in issue did not result in a tax benefit. If and when the PUC is returned, it would be a question of fact whether the return of capital was part of the series that included the butterfly reorganization. The entire discussion of the meaning of series in Copthorne would have been irrelevant if the increase in PUC on the amalgamation was itself a tax benefit.[3]

The “avoidance transaction” issue

A similar issue arises with respect to the test for avoidance transaction. The agreed statement of facts identified the avoidance transactions as the first and second transfers to each of the transferee corporations, on which transfers the transferee corporation issued the same class of shares to the transferor.

Based on prior case law, it is difficult to see how these transactions could be considered avoidance transactions. Both parties agreed that the purpose of the series of transactions was to protect the assets of the corporation. Each one of the transactions undertaken was required to effect that purpose on a tax-deferred basis. There was no suggestion that being able to split the assets on a tax-deferred basis resulted in an abusive tax benefit. The misuse or abuse related only to the increase in the PUC of the shares of the transferee corporations. While it is clear from MacKay that an overall non-tax purpose for a series of transactions does not save every transaction in the series, it is also clear from Canadian Pacific that it is necessary to identify at least one transaction that leads to the impugned tax benefit and that was not undertaken for bona fide purposes other than to obtain that tax benefit. In identifying such a transaction, it is not open to the Crown to artificially split off various aspects of the transaction in order to create an avoidance transaction.

The Court referred to the following paragraph in the agreed statement of facts as an avoidance transaction:

18. Because Perry Wild caused 1245 [the first transferee corporation] to issue Class E preferred shares to himself and to PWR, the PUC of Perry Wild’s 1,989 Class E preferred shares of 1245 was increased to $467,115.62 and the PUC of the 1,826.242 Class E preferred shares of 1245 received by PWR was reduced to a like amount as a result of subsection 89(1).

This statement appears to be simply an explanation of a particular tax result arising from the transactions and not a transaction in and of itself. If the suggestion is that the choice to have the transferee corporations issue the same class of shares on the first and second transfers is an avoidance transaction, surely that contradicts the Federal Court of Appeal’s findings in Canadian Pacific. In that case, the taxpayer obtained a tax benefit by borrowing in Australian dollars rather than Canadian dollars. As there was a bona fide non-tax purpose for the borrowing, the Crown attempted to classify the designation of the currency of the borrowing as a transaction. The Federal Court of Appeal noted that:

In the present case, the Australian dollar borrowing was one complete transaction and cannot be separated into two transactions by labelling the designation in Australian dollars as a separate transaction.

Similarly, the first and second transfers were each complete transactions that cannot be separated into two transactions by treating the choice to issue the same class of shares as a separate transaction from the actual issuance of the shares.

As noted above, the misuse analysis focused primarily on section 84.1 and the choice to have the transferee corporation issue the same class of shares on the first and second transfers. As the Court notes: “The strategy directed by Mr. Wild, facilitated by indifferent non-arm’s length parties (two of which were newly-created companies), resulted in a five thousand-fold increase in his PUC even though no new capital had been contributed by him to PWR nor 1245 [a transferee corporation].” The judgement contains excerpts from PW’s examination for discovery in which he was questioned about the choice to use the same class of shares. While his answer was simply that this is what he was instructed to do by his advisors, could he not just as well have responded by asking why he would be required to take a separate class?

From the comments made by the Court to the above question, the answer to this would appear to be that arm’s length parties would have insisted on separate classes of shares. That answer appears too simplistic. Arm’s length parties do not always negotiate for tax attributes that are not required or that cannot be used. In certain circumstances, an arm’s length purchaser may take steps to acquire a separate class of shares to avoid PUC averaging but a Canadian resident corporate purchaser acquiring in excess of 10 per cent of the shares of another corporation may very well be indifferent. Arm’s length non-residents that cannot benefit from a capital dividend account often accept structures that allow such dividends to be paid disproportionately to Canadian resident shareholders. The CRA has sanctioned such structures. By effectively requiring PW to structure the transactions using separate classes of shares to avoid PUC averaging, it seems the judgement offends the basic principle that the GAAR does not require a taxpayer to structure a transaction in the least tax efficient manner. While the transactions did create the attributes that would allow PW to use his capital gains exemption to strip surplus from PWR in the future, a result that section 84.1 is designed to avoid, this result by itself does not automatically lead to a finding of misuse or abuse. A similar tax benefit was obtained in Evans, a post Canada Trustco decision, in which the Tax Court found that the transactions that created this tax benefit did not result in an abuse or misuse of the provisions of the Act. As noted by Mr. Justice Bowman:

I do not think that it can be said that there is an abuse of the provisions of the Act where each section operates exactly the way it is supposed to. The Crown’s position seems to be predicated on the view that since everything worked like clockwork there must have been an abuse. The answer to this position is, of course, that if everything had not worked like clockwork we would not be here.


We are left with the GAAR applying to a series of transactions that arguably, did not result in a tax benefit, did not include an avoidance transaction and did not result in misuse or abuse. Nearly 30 years after the enactment of the GAAR and 12 years after the Supreme Court’s guidance in Canada Trustco, taxpayers, their advisors and the courts continue to struggle with the basic principles of this perplexing rule.