The Uncertain Demise of the Duke of Westminster: A Comment on Lipson v Canada

My colleague Julian Ho has already provided a succinct summary of last Thursday’s Supreme Court decision in Lipson v Canada, 2009 SCC 1 [Lipson], so today I thought I’d comment briefly on the general anti-avoidance rule (GAAR) and the Court’s adoption of a purposive approach in determining whether a transaction or series of transaction is abusive. The majority’s reasons in Lipson are somewhat disappointing as they fall short of resolving several pressing issues that arose out of Canada Trustco Mortgage Co v Canada, 2005 SCC 54 [Canada Trustco] (which, prior to Lipson, was the leading case on interpreting the GAAR). Specifically, the Court has failed to clarify what the parameters of the GAAR are and how the GAAR may coexist with tax law’s longstanding approach to statutory interpretation, which favours form over substance.

The purpose of the GAAR is to undo any transaction that, in effecting a tax benefit for the taxpayer, has unacceptably manipulated or exploited the rules of the ITA. Prior to the enactment of the GAAR in 1988, the government’s only available response to tax plans that were legal but overly aggressive was to enact specific anti-avoidance rules prohibiting a particular transaction from reoccurring. However, specific anti-avoidance rules have proven only partially effective; they are reactive measures that may create further loopholes which in turn spawn more aggressive tax planning. The usefulness of a general anti-avoidance rule for the Canadian Revenue Agency is obvious: the GAAR is an umbrella rule that prohibits any abusive transaction or series of transactions that have not yet been contemplated by the ITA. Uncertainty arises, however, as to what constitutes an abusive transaction.

In Canada Trustco, the Court created a three-prong test with which to determine whether a transaction triggers the GAAR: (1) the transaction must produce a tax benefit to the taxpayer; (2) the transaction must be an avoidance transaction as defined in the ITA; and (3) the transaction must result in an abuse or misuse of the ITA provisions relied upon by the taxpayer. The third prong of the test is problematic because neither abuse nor misuse are defined. In Canada Trustco, the Court conflates the two terms. In establishing whether there has been an abuse or misuse of an ITA provision, a court must first determine what the purpose of that provision is. It must then ask whether the taxpayer’s transaction (or series of transactions) frustrates that purpose.

The SCC has reasoned that, in order to find an abuse or misuse of an ITA provision, one first must define what the use of that provision is. Sounds reasonable, but this purposive approach subverts the “textual, contextual and purposive” approach to statutory interpretation that has long been relied upon by tax planners. The GAAR ignores the plain text meaning of the provision and moving straight to a larger purposive analysis. This is inconsistent with tax law’s fundamental principle – as stated in Commissioners of Inland Revenue v Duke of Westminster, [1936] AC 1 (HL) – that a taxpayer has the right to order her affairs as she sees fit to minimize her tax payable. Duke of Westminster embodies the proposition that a textual reading of a tax provision should precede and override any contextual or purposive analysis. Lipson leaves unclear how much the GAAR should be permitted to erode that rule.

While divining a legislative purpose is nothing new for the Court, a predictable interpretation is essential for taxpayers to plan their financial affairs. Deference to statutory text, rather than statutory purpose, allows a taxpayer certainty in tax planning. That certainty is removed when a provision’s underlying objectives – unwritten and unbeknownst to the average taxpayer – are allowed to override the provision’s clear wording on which the taxpayer relies.

The majority’s reasons in Lipson also provide no guidelines on how to determine an ITA provision’s purpose or how to choose between competing interpretations. The majority held that s. 74.1(1) of the ITA, the property attribution rule for spouses, was an anti-avoidance rule designed “to prevent spouses from reducing tax by taking advantage of their non-arm’s length relationship when transferring property between themselves”; the appellant’s series of transactions subverted this purpose by allowing income-splitting.

In his dissent, however, Justice Binnie noted that that s. 74.1(1) only applied if spouses did not opt out of s. 73(1), the spousal rollover rule. Since Parliament gave spouses a choice regarding property attribution, s. 74.1(1) was in effect not an anti-avoidance rule at all, but one that encouraged tax minimization. “Far from offending the ‘object, spirit or purpose’ of the spousal attribution rules, the taxpayer’s tax plan fulfilled them, or at a minimum did not abuse them.” Using a purposive approach to determine whether there’s been an abuse of an ITA provision produces uncertainty, since it is rarely clear what the true purpose behind a tax rule is.

Writing for the majority, Justice LeBel responds to Justice Binnie’s concerns by stating that “such uncertainty is inherent in all in all situations in which the law must be applied to unique facts.” Yet in situations of genuine ambiguity over what an ITA provision means, there is supposed to be a residual presumption in favour the taxpayer (Placer Dome Canada Ltd v Ontario, [2006] 1 SCR 715). In Lipson, the majority did not consider whether Justice Binnie’s competing interpretation of s. 74.1(1) could bring genuine ambiguity to a provision’s purpose. Neither do they explain what steps must be followed to find a provision’s underlying purpose.

The majority’s decision in Lipson diminishes the importance of the Duke of Windsor method of interpreting the ITA, by allowing a purposive analysis of a provision to supersede any textual or contextual interpretation. I agree with Justice Binnie’s view that “the application of the GAAR in these circumstances […] means paying lip service to the Duke of Westminster principle without taking seriously its role in promoting consistency, predictability and fairness in the tax system.” The majority’s reasons speak of the need to limit the GAAR’s scope but fail to explain how this may be done. It remains unclear what kind of aggressive tax plan could exploit the word and letter of the ITA and still withstand a challenge under the GAAR.