December 22, 2011
The Canadian Tax System Gets a Christmas Present from the Supreme Court
Yes Virginia, there is a GAAR 1 … it isn’t just a figment of your imagination.
On Friday of last week, the Supreme Court finally released its decision in the Copthorne Holdings case (2011 SCC 63). Many expected a split decision because of the time that it took the Court to render its decision. However, the Court decided unanimously that the GAAR applied. Let me say at the outset that I think the result is correct. Any other result would have seriously weakened or neutered the GAAR, and opened the possibility of a flood of abusive transactions to create cross-border paid-up capital and the avoidance of Canadian tax on dividends paid to nonresidents.
The facts are well known so I have provided only a brief summary here. Copthorne, a Canadian corporation that was part of the multinational group ultimately owned by Li Ka-Shing and his son, owned all of the shares of VHHC Holdings, another Canadian corporation. The paid-up capital of the shares of VHHC Holdings was approximately $67 million. In 1993 the shares of VHHC Holdings were transferred to Copthorne’s parent company, thereby making Copthorne Holdings and VHHC Holdings sister companies.
In early 1994 Copthorne and VHHC Holdings were amalgamated to form Copthorne II. The paid-up capital of the amalgamated company was equal to the aggregate of the paid-up capital of the each of the predecessor companies. Some other transactions subsequently took place, including the amalgamation of Copthorne II with three other corporations to form the taxpayer, Copthorne Holdings. In early 1995 Copthorne Holdings redeemed some of its shares held by its nonresident parents.
According to the taxpayer, the amount for which the shares were redeemed did not exceed their paid-up capital; therefore, there was no deemed dividend subject to Canadian withholding tax. The Canada Revenue Agency applied the GAAR to reduce the paid-up capital of the shares of Copthorne Holdings by $67 million, the paid-up capital of the shares of VHHC Holdings. This resulted in a deemed dividend of $67 million and Canadian withholding tax of 15 percent on that amount.
Both the Tax Court and the Federal Court of Appeal had no difficulty finding that the GAAR applied. The Tax Court thought that the transactions had artificially increased the paid-up capital on the amalgamation and the amount distributed to shareholders as a tax-free return of capital, and that this was an abuse of sections 84(3), 87(3) and 89(1), the definition of paid-up capital. The Federal Court of Appeal took the view that section 89(1) was the only provision abused as a result of the duplication of paid-up capital on the amalgamation.
The Supreme Court granted leave to appeal and the case was argued in January 2011. Given the unsatisfactory result of the split decision in the Lipson case ( SCC 1), the decision in Copthorne was eagerly anticipated by the tax community. The decision was expected as early as September and as the months passed the speculation increased. Many expressed the view that the delay indicated a serious disagreement among the judges of the Court concerning the GAAR.
In my judgment, the decision was worth waiting for (although I don’t really understand why it took so long because the correct answer was pretty obvious). The reasons for judgment, written by Justice Rothstein for a unanimous court, are refreshingly concise, straightforward, and comprehensive. They provide a solid platform for lower courts to continue developing the GAAR into an effective tool to protect the tax system against abusive tax avoidance; they will provide welcome encouragement to the CRA officials charged with responsibility for controlling tax avoidance; and they represent a clear signal to taxpayers and their professional advisers that the GAAR is neither a paper nor a toothless tiger.
Now let’s get into the details of Justice Rothstein’s reasons.
First, was there a tax benefit? The only issue for the Supreme Court was whether the Tax Court made a palpable and overriding error in finding that there was a tax benefit. For this purpose, the Court thought it appropriate to engage in a comparison of the transactions actually carried out and a hypothetical arrangement that might have been carried out but for the tax benefit. The Court adopted the approach suggested by Professor David Duff and his co-authors in Canadian Income Tax Law, 3rd edition 2009. 2
To me this comparative analysis, while it may be necessary in some cases, was not necessary in this case. As the Supreme Court said in Canada Trustco, any deduction, credit, or allowance is a tax benefit without the need for any further analysis. I would add the avoidance or reduction of Canadian withholding tax to this list. I cannot imagine a situation in which it would be appropriate to conclude that a reduction of withholding tax is not a tax benefit.
Second, was there an avoidance transaction? This question raises the issue of a series of transactions, which I, and many others, thought (mistakenly as it turns out) would be the central issue in the Copthorne case. The taxpayer conceded that the sale of the shares of VHHC Holdings to Copthorne’s parent, making it a sister company, and the amalgamation of Copthorne and VHHC Holdings were both part of a series.
The crucial question, however, was whether the subsequent redemption of the shares of the taxpayer was a related transaction carried out in contemplation of the series. If so, the redemption is deemed to be part of the series pursuant to subsection 248(10). The proper test for the application of subsection 248(10) has been controversial since the introduction of the GAAR.
The series aspect of the GAAR is critical to its effectiveness. It must apply both to a series of transactions that constitutes abusive tax avoidance and to abusive transactions that are inserted into a series that has an overall legitimate commercial purpose. For this purpose, the concept of a series needs to have a broad meaning. Early on, the courts adopted the narrow meaning of a series from UK tax law (each transaction must be pre-ordained). This narrow meaning of series would be relatively easy for taxpayers to avoid, and therefore the importance of subsection 248(10) is obvious.
The Tax Court suggested that the appropriate test for subsection 248(10) was a “strong nexus” between the related transaction and the series. The Federal Court of Appeal disagreed and held that all subsection 248(10) required was that the series be “a motivating factor” for the related transaction. The Supreme Court endorsed the test adopted in Canada Trustco that the related transaction must be carried out “because of” or “in relation to” the series; this test does not require a “strong nexus” but does require more than a mere possibility. Each case must be decided on the basis of its own facts.
In my view, the “in relation to” test is meaningless because it merely restates the issue in almost identical terms. When is a transaction related to a series? When it is carried out in relation to the series? On the other hand, the “because of” test is a test of causation that can be applied objectively. Did the series cause or make possible the related transaction?
One difficulty with this interpretation of subsection 248(10) is that it effectively ignores the meaning of the words “in contemplation of.” The taxpayer in Copthorne argued that the words “in contemplation of” should be interpreted prospectively so that the provision is limited to transactions carried out because of a subsequent series and does not apply to transactions carried out after a prior series.3
Quite correctly, the Supreme Court rejected this narrow interpretation of subsection 248(10). It noted that the dictionary definition of “contemplation” is not limited to looking forward, and suggested that the context (and surely the purpose) of subsection 248(10) supports a broad meaning. Most importantly, Justice Rothstein relied on the clear statement in Canada Trustco that subsection 248(10) “can be applied to events either before or after the basic avoidance transaction.” This statement, he suggested, was adopted and expanded from the Federal Court of Appeal decision in OSFC Holdings ( FCA 260) (also written by Justice Rothstein).
The real problem that the wording of subsection 248(10) presents, and that courts and commentators have not confronted, is that contemplation is an attribute of a person. A particular person may carry out a series of transactions because that person anticipates or expects that a subsequent transaction will be carried out. If the subsequent transaction takes place, it would be in the contemplation of the particular person when carrying out the prior series. However, the person carrying out the subsequent transaction may not be aware of the prior series; if so, from that person’s perspective the prior series was not a cause of the subsequent transaction.
This should be irrelevant for purposes of applying subsection 248(10). The prior series sets the table for the feast; there is no other reason for setting the table. The same analysis applies where a person carries out a transaction because of a subsequent series of transactions that is anticipated or expected to occur, but here it is the contemplation of the person carrying out the related transaction, and not the series, that is relevant.
This shifting of the focus on who is doing the contemplating is not problematic to me. I can see nothing in the text, context, or purpose of subsection 248(10) that necessitates restricting it to the contemplation of the person carrying out the series or the related transaction, whichever comes first. If the Supreme Court had concluded that subsection 248(10) applied only prospectively, it would have been necessary for the government to amend it to ensure that the series aspect of the GAAR works properly.
Tax practitioners will undoubtedly be concerned that the Supreme Court did not provide any guidance concerning the “old and cold” issue. We know that 18 months – the time between the sale of the shares of VHHC Holdings to make it a sister company and the redemption of the shares of the taxpayer – is not sufficient. It will be a question of fact in each case whether prior or subsequent transactions are related to the series.
Having concluded that the redemption was part of the series, the Supreme Court easily confirmed the finding of the Tax Court that the sale of the shares of VHHC Holdings to make it a sister company was a transaction that was part of the series the primary purpose of which was to obtain a tax benefit. Therefore, the sale of the shares of VHHC was an avoidance transaction within the meaning of subsection 245(3).
Third, was the avoidance transaction abusive? The GAAR jurisprudence has resolved that the issue of abuse is the central focus of any GAAR case. The Canadian courts have done a good job in this regard. They have resisted the temptation to decide GAAR cases on the basis of peripheral issues like tax benefit, series, and the primary purpose of transactions.
In Copthorne, the Supreme Court confirmed the basic two-step approach adopted in Canada Trustco: first, determine the purpose of the provisions relied on for the tax benefit, and second, determine whether the avoidance transaction at issue frustrates or abuses the purpose of those provisions. Before applying this approach, Justice Rothstein set out several fundamental propositions that must be followed in conducting the abuse analysis:
- The GAAR imposes on the courts “the unusual duty of going beyond the words of the legislation to determine the object, spirit or purpose of the provision or provisions relied upon by the taxpayer;” [more about this later]
- The GAAR is a provision of last resort;
- Courts “must approach a GAAR decision cautiously” (Does this imply that other decisions do not require caution?) because taxpayers are entitled to “consistency, predictability and fairness in tax law.” (The repetition of a meaningless slogan from Canada Trustco that was not supported by any substantive analysis.) As a result, the GAAR can be applied only if the abuse is clear; in effect, any doubt must be resolved in favour of the taxpayer.
- There is no distinction between misuse and abuse.
With respect to first step in the GAAR analysis – determining the purpose of the relevant provisions – Justice Rothstein indicated that, although the same approach that is used for statutory interpretation generally (a unified textual, contextual, and purposive approach) is used for this purpose, the objective is different. In an exercise of statutory interpretation the objective is to determine the meaning of the words, whereas in a GAAR analysis the objective is to determine the purpose of the provision: “The search is for the rationale that underlies the words that may not be captured by the bare meaning of the words themselves.” This is dangerously close to a search for the overriding purpose which the Supreme Court rejected in Canada Trustco.
With respect to the second step – whether the avoidance transaction abuses or frustrates the relevant provisions – the Court confirmed that, where the transaction is part of a series, the series as a whole must be considered. (This principle was first enunciated in the Lipson case.)
Justice Rothstein identified subsection 87(3) as the only provision that was potentially abused. This is an improvement on the Federal Court of Appeal’s finding that the definition of paid-up capital had been abused. I guess it is possible to frustrate the purpose of a definition but I think it makes more sense to consider the question of abuse in the context of the operative provisions in which the definition applies.
I don’t see why in this case subsection 84(3) hasn’t also been abused. After all, the reason for the preservation of the paid-up capital was to eliminate the deemed dividend that would otherwise have arisen. Justice Rothstein did not address this point; he simply stated that subsection 84(3) doesn’t alter the computation of paid-up capital. This is true, but it doesn’t necessarily follow that there cannot be an abuse of subsection 84(3).
Justice Rothstein determined the object, spirit, and purpose of subsection 87(3) by conducting a textual, contextual, and purposive analysis of the provision. (Forget about the fact that it doesn’t make sense to engage in a purposive analysis to determine the purpose of a provision. Instead, think about it as considering all of the available information to determine the purpose of a provision.) Based on the words, Justice Rothstein found that subsection 87(3) ensures that the paid-up capital of a corporation cannot be inappropriately increased as a result of an amalgamation. In particular, on a vertical amalgamation, any intercorporate paid-up capital must be eliminated.
His contextual analysis of the Act led to the same conclusion about the purpose of subsection 87(3). Provisions such as sections 84.1 and 212.1 grind the paid-up capital of a corporation with respect to certain non-arm’s length transactions to prevent surplus stripping. According to the Supreme Court, because subsection 87(3) is similar to the other grinds to paid-up capital:
It is reasonable to conclude that it shares the same purpose of precluding the preservation of PUC where such preservation would allow for a withdrawal, without liability for tax, of an amount in excess of the investment made with after-tax funds.
The Court rejected several specious, desperate (?) arguments made by the taxpayer – for example, because tax will eventually be paid by shareholders in the form of capital gains or dividends, there can be no abuse. As the Court correctly pointed out, capital gains may be taxable at a lower rate than dividends, and in the case of nonresidents in treaty countries, may be realized without any Canadian tax. The taxpayer also argued that because there were no rules in the Act with respect to paid-up capital similar to the stop-loss rules for losses, there was no policy against the preservation of paid-up capital. The Court did not fall for this negative implication argument, which boils down to: because there is no specific rule preventing taxpayers from doing something, there can be no abuse if they do it. Justice Rothstein was not fooled:
If such an approach were accepted, it would be a full response in all GAAR cases, because the actions of a taxpayer will always be permitted by the text of the Act. . . . and the GAAR would be rendered meaningless.
Unfortunately, this will not be the last time that such a negative implication argument will be made.
As to the purpose of subsection 87(3), the Court concluded that “[T]he parenthetical portion [of subsection 87(3)] seeks to preclude corporations from preserving PUC of the shares of a subsidiary corporation on amalgamation with the parent corporation,” and rejected all of the taxpayer’s arguments. In particular, the Court rejected the argument that, if validly created paid-up capital cannot be preserved, taxpayers would be placed “in a state of impermissible uncertainty.” There should be some kind of penalty (a 10 minute misconduct for counsel?) for this type of outrageous hyperbole. The taxpayer’s better argument was that Canada Trustco prevents courts from searching “for an overriding policy of the Act that is not based on a unified, textual, contextual and purposive interpretation of the provisions at issue.”
According to the taxpayer, the only possible justification for a finding of abuse in this case would be a general policy against surplus stripping. Justice Rothstein agreed that it would not be proper to find abuse based “on a broad statement of purpose such as anti-surplus stripping.” However, in this case the purpose is derived from the paid-up capital provisions of the Act and not from a broadly stated policy. This distinction is likely to surface in future GAAR cases. The bottom line seems to be that the GAAR will be applied only in cases where the Crown can establish a clear statutory policy based on the provisions at issue.
Having determined the purpose of subsection 87(3), Justice Rothstein turned to the second step in the GAAR analysis: whether the transactions in the case frustrate or abuse that purpose. He concluded that the preservation or double counting of $67 million of paid-up capital on the amalgamation frustrated or defeated the purpose of subsection 87(3). Unfortunately, Justice Rothstein stumbled in his analysis at this point. He stated that the double counting of paid-up capital is not sufficient evidence of abuse by itself; it is only abusive where it is “artificially” preserved.
Nowhere in the wording of the GAAR is the term “artificial” used4 or the manner in which a transaction or series of transactions is carried out referred to. Artificiality was the test under former subsection 245(1). However, that test was intentionally rejected for purposes of the GAAR, quite rightly in my opinion. Many transactions are artificial – just think about income trusts or streaming borrowed funds to eligible uses to maximize interest deductions – but are clearly acceptable (i.e., not abusive) for income tax purposes.
The test of abuse in subsection 245(4) focuses on the result of an avoidance transaction, not the manner in which the transaction was carried out.5 Despite this misstep, Justice Rothstein got to the right answer:
The sale of VHHC Holdings shares to Big City [its parent] circumvented the parenthetical words of s. 87(3) and in the context of the series of which it was a part, achieved a result the section was intended to prevent and thus defeated its underlying rationale.
A few concluding comments. The case is important for what it doesn’t say. Although the Court in Copthorne often refers to Canada Trustco as authority, it is important to recognize that elements of Canada Trustco are ignored. First, Neither Lipson nor Copthorne mention or apply the test of abuse adopted in Canada Trustco, which was:
Abusive tax avoidance may be found where the relationships and transactions as expressed in the relevant documentation lack a proper basis relative to the object, spirit or purpose of the provisions that are purported to confer the tax benefit, or where they are wholly dissimilar to the relationships or transactions that are contemplated by the provisions.
The fact that the Court has ignored its own test of abuse is a good thing because the Canada Trustco test is tautological and cannot be justified on the basis of the wording of subsection 245(4) (where, for example, does the reference to relationships come from?).
Second, Justice Rothstein never mentions that, according to Canada Trustco, step one of the abuse analysis is a question of law and step two a question of fact. These characterizations are inappropriate and should be abandoned. In my view, the determination of whether a transaction is abusive is a quintessential legal question or, at least, a mixed question of fact and law.
My final comment relates to the relationship between statutory interpretation and the GAAR. In Copthorne Justice Rothstein wrote that:
The GAAR is a legal mechanism whereby Parliament has conferred on the court the unusual duty of going behind the words of the legislation to determine the object, spirit or purpose of the provision or provisions relied upon by the taxpayer. While the taxpayer’s transactions will be in strict compliance with the text of the relevant provisions relied upon, they may not necessarily be in accord with their object, spirit or purpose.
The search is for the rationale that underlies the words that may not be captured by the bare meaning of the words themselves.
These statements reveal a fundamental inconsistency in the Supreme Court’s approach to statutory interpretation and its application of the GAAR. Many times the Court has indicated that all statutes, including tax statutes, must be interpreted in accordance with the text or wording of the relevant provisions, the context, and the purpose of the provisions. Therefore, courts must always go beyond “the bare meaning of the words” to determine their meaning, and the GAAR does not involve anything “unusual” for them. Indeed, in Canada Trustco, the Supreme Court said that “the inquiry under s. 245(4) requires the court to look beyond the mere text of the provisions” but in the next sentence went on to indicate “[T]here is nothing novel in this.”
The fundamental problem that the courts have never confronted is that, if the provisions of the Act other than the GAAR are interpreted in a textual, contextual, and purposive manner, what is the point of going through the same exercise again under subsection 245(4)? If the transactions in Copthorne frustrated the purpose of subsection 87(3), why wasn’t the provision applied to limit the paid-up capital in accordance with the textual, contextual, and purposive approach to statutory interpretation? Or, if the transactions did not frustrate the purpose of subsection 87(3), how can the result change when the same exercise is performed under subsection 245(4)? These are mysteries that I would be much happier to ignore.
Finally, I want to wish season’s greetings and happiness and prosperity for 2012 to all of the readers of my reports. I realize that there is an ocean of material for you to read. If you choose to read my reports, I take that as a real compliment even if – or especially if – you don’t agree with what I write. Thank you and please keep reading.