The Bank of Montreal v. The Queen

The Bank of Montreal v. The Queen
Court (s) Database
Tax Court of Canada Judgments
Date
2018-09-12
Neutral citation
2018 TCC 187
File numbers
2016-445(IT)G
Judges and Taxing Officers
David E. Graham
Subjects
Income Tax Act
Decision Content
Docket: 2016-445(IT)G
BETWEEN:
THE BANK OF MONTREAL,
Appellant,
and
HER MAJESTY THE QUEEN,
Respondent.
Appeal heard on June 4, 5, 6, 7 and 8, 2018 at Toronto, Ontario
Before: The Honourable Justice David E. Graham
Appearances:
Counsel for the Appellant:
Martha MacDonald
Jerald Wortsman
Patrick Reynaud
Counsel for the Respondent:
Natalie Goulard
Sara Jahanbakhsh
Marie-France Camiré
JUDGMENT
The appeal of the reassessment of the Appellant’s taxation year ending October 31, 2010 is allowed and the matter is referred back to the Minister of National Revenue for reassessment on the basis that section 245 of the Income Tax Act did not apply to the transactions in question.
Costs are awarded to the Appellant. The parties shall have 30 days from the date hereof to reach an agreement on costs, failing which the Appellant shall have a further 30 days to file written submissions on costs and the Respondent shall have yet a further 30 days to file a written response. Any such submissions shall not exceed 10 pages in length. If the parties do not advise the Court that they have reached an agreement and no submissions are received within the foregoing time limits, costs shall be awarded to the Appellant as set out in the Tariff.
Signed at Ottawa, Canada, this 12th day of September 2018.
“David E. Graham”
Graham J.
Citation: 2018 TCC 187
Date: 20180912
Docket: 2016-445(IT)G
BETWEEN:
THE BANK OF MONTREAL,
Appellant,
and
HER MAJESTY THE QUEEN,
Respondent.
REASONS FOR JUDGMENT
Graham J.
[1] This appeal involves the application of the general anti-avoidance rule (the “GARR”) to a complex series of cross-border financing transactions entered into by the Bank of Montreal between 2005 and 2010 that avoided the application of the dividend stop-loss rules in subsection 112(3.1) of the Income Tax Act (the “Act”). The Bank of Montreal takes the position that it did not receive a tax benefit from the transactions and, even if it did, the receipt of that benefit was not an abuse of subsection 112(3.1).
[2] The parties filed a Partial Agreed Statement of Facts which describes the structure and the financing in detail. A copy of the Partial Agreed Statement of Facts is attached as Appendix “A”. In these Reasons for Judgment, I have simplified the facts somewhat for ease of presentation. [1] Simplified transaction diagrams prepared by counsel for the Bank of Montreal are attached as Appendix “B”. [2]
A. Overview
[3] In 2005, the Bank of Montreal (“BMO”) wanted to lend a total of $1.4 billion USD to a number of its US subsidiaries referred to as the Harris Group. BMO chose to borrow those funds from third parties.
i. Tower Structure
[4] It would not have been tax efficient for BMO to simply borrow the funds and lend them to the Harris Group. Such a structure would have resulted in BMO having to pay US withholding tax on the interest payments it received from the Harris Group. As a result, BMO implemented what is commonly referred to as a “tower structure”. A tower structure is a complicated structure often used by Canadian companies to finance US subsidiaries in a tax efficient manner. It allows the deduction of interest costs by the Canadian company for Canadian tax purposes and the deduction of the corresponding interest costs by the US subsidiary for US tax purposes without having to pay withholding tax to the US on the repatriation of the funds.
[5] The tower structure implemented by BMO consisted of the following entities:
(a) a Nevada limited partnership named BMO Funding L.P. (“Funding LP”) in which BMO had a 99.9% interest and a wholly owned subsidiary of BMO named BMO G.P. Inc. (“BMO GP”) had a 0.1% interest;
(b) a Nova Scotia unlimited liability company named BMO (NS) Investment Company (“NSULC”) that was wholly owned by Funding LP; and
(c) a Delaware limited liability company named BMO (US) Funding LLC (“LLC”) that was wholly owned by NSULC.
ii. Borrowed Funds
[6] BMO borrowed $150 million USD from a third party. It invested those funds in Funding LP. Funding LP, in turn, used those funds to acquire shares of NSULC which, in turn, used those funds to acquire shares in LLC. LLC then took the funds that it had received and lent them to the Harris Group. [3]
[7] The balance of the required $1.4 billion USD came from a $1.25 billion USD loan obtained by Funding LP from a third party. Again, Funding LP used those funds to acquire shares of NSULC which, in turn, used those funds to acquire shares in LLC. LLC then took the funds that it had received and lent them to the Harris Group. [4]
iii. Quarterly Payments
[8] Interest payments and dividends flowed through the tower structure at the end of each fiscal quarter. The Harris Group would pay interest to LLC. LLC would then use the money to pay dividends to NSULC. NSULC would pay corresponding dividends to Funding LP. Funding LP would use the funds it received to pay interest on the $1.25 billion USD that it had borrowed and would distribute the balance to BMO and BMO GP. BMO would, in turn, use the funds it received from Funding LP to pay interest on the $150 million USD that it had borrowed.
[9] The dividends received by BMO from NSULC (indirectly through Funding LP) were taxable dividends. BMO benefited from a subsection 112(1) deduction in respect of those dividends.
iv. Hedging Foreign Exchange Risk
[10] From a business point of view, by borrowing US dollars to make an investment in a US asset, BMO effectively hedged its foreign exchange risk. If the Canadian dollar decreased in value against the US dollar between 2005 and 2010, then the increase in value (in Canadian dollars) of BMO’s indirect US dollar investment in the Harris Group would be matched by the increased cost (in Canadian dollars) of repaying the $1.4 billion USD in borrowed funds. Conversely, if the Canadian dollar increased in value against the US dollar between 2005 and 2010, then the decrease in value (in Canadian dollars) of BMO’s indirect US dollar investment in the Harris Group would be matched by the decreased cost (in Canadian dollars) of repaying the $1.4 billion USD in borrowed funds.
[11] However, from a tax point of view, BMO faced a potential problem with hedging its foreign exchange risk. There would not be any problem if the Canadian dollar decreased in value. Any increase in the value of the NSULC shares held by Funding LP that arose from a decrease in the value of the Canadian dollar would be taxable as a capital gain. That capital gain would be offset by the corresponding capital loss that would arise on the repayment of the $1.4 billion USD in borrowed funds. On the other hand, BMO would have a problem if the Canadian dollar increased in value. The resulting decrease in the value of the NSULC shares held by Funding LP would give rise to a capital loss. However, the stop-loss rule in subsection 112(3.1) would reduce that capital loss by an amount equal to the value of any non-taxable dividends that Funding LP had received from NSULC. As a result, the reduced capital loss would not be sufficient to fully offset the capital gain that would arise on the repayment of the $1.4 billion USD in borrowed funds.
v. Modified Structure
[12] To avoid this potential mismatch of the capital gain and capital loss, BMO implemented a modification to the tower structure. Subsection 112(3.1) applies separately to each class of shares. Therefore, BMO decided to create a structure whereby NSULC had two classes of shares. When the first set of quarterly dividends was being paid, instead of paying a cash dividend, NSULC paid a stock dividend consisting of preferred shares. This resulted in Funding LP holding two classes of shares of NSULC: common shares with a high cost base and preferred shares with a low cost base. From that point forward, all quarterly dividends were paid on the preferred shares. By isolating the dividends in this manner, BMO ensured that, if the Canadian dollar increased in value resulting in a capital loss on the common shares of NSULC, the loss would not be ground by the value of the non-taxable dividends that had been paid by NSULC and thus the loss would fully offset the capital gain on the borrowed funds.
[13] BMO’s planning proved fortuitous. The Canadian dollar increased in value significantly between 2005 and 2010. [5] The use of the two-class structure ensured that the capital loss on the disposition of the shares of NSULC matched the capital gain on the borrowed funds.
vi. Unwinding the Structure
[14] The tower structure was unwound in 2010. The Harris Group repaid the loans made by LLC. LLC was wound up and its assets were distributed to NSULC. NSULC, in turn, was wound up and its assets were distributed to Funding LP. Funding LP used the funds it received to repay the $1.25 billion USD that it had borrowed. It distributed the balance of the funds to BMO and BMO GP. BMO used the funds it received from Funding LP to repay the $150 million USD that it had borrowed.
vii. Reassessment
[15] The Minister of National Revenue was not pleased with the modifications that BMO had made to the tower structure to address its foreign exchange-related tax risk. The Minister concluded that, while the resulting structure technically complied with the Act, the GAAR should be used to reduce BMO’s share of the capital loss on the disposition of the common shares of NSULC by the amount by which the loss would have been ground under subsection 112(3.1) had BMO not modified the structure. Accordingly, the Minister reassessed BMO to reduce its capital loss by $287,766,503.
[16] It is important to note that the GAAR reassessment did not challenge the use of the tower structure itself, only the modification to the structure that BMO made by introducing the second class of shares.
viii. Issues, Positions and Conclusion
[17] There are three conditions that must be present for the GAAR to apply. First, there must have been a tax benefit arising from a transaction. Second, that transaction must have been an avoidance transaction. Finally, the avoidance transaction must have been abusive. [6]
[18] The Respondent submits BMO received a tax benefit when its share of Funding LP’s capital loss on the NSULC common shares was not reduced by the $287,766,503 in non-taxable dividends received by Funding LP on the NSULC preferred shares. BMO submits that there was no tax benefit. It asserts that, even if a second class of shares had not been introduced into the tower structure, subsection 112(3.1) would still not have applied to grind the capital loss. BMO argues that, as the Act read in 2010, subsection 112(3.1) did not apply to capital losses that arose from foreign exchange fluctuations.
[19] BMO accepts that, if I find there was a tax benefit, the benefit resulted from a series of transactions that were primarily undertaken to obtain the benefit. In particular, if I find there was a tax benefit, BMO accepts that the stock dividend by which the preferred shares of NSULC were created and the subsequent payment of all dividends on those preferred shares to the exclusion of the common shares was an avoidance transaction.
[20] However, BMO argues that the Respondent has failed to prove that the avoidance transaction was an abuse of subsection 112(3.1). The Respondent submits that the object, spirit and purpose of subsection 112(3.1) is to prevent the overstatement of capital losses incurred on the disposition of shares by reducing the losses by any tax-free dividends received on those shares. The Respondent says that BMO abused that object, spirit and purpose.
[21] For the reasons that follow, I find that there was no tax benefit. Having reached that conclusion, there is no need for me to consider whether subsection 112(3.1) was abused.
B. Tax Benefit
[22] Subsection 245(1) defines a “tax benefit” as a reduction, avoidance or deferral of tax. The Respondent says that the tax benefit BMO received was the reduction in its tax payable as a result of subsection 112(3.1) not applying to reduce its share of the capital loss on the disposition of the common shares of NSULC.
[23] The relevant portions of subsection 112(3.1) read as follows:
. . . [W]here a taxpayer . . . is a member of a partnership, the taxpayer’s share of any loss of the partnership from the disposition of a share that is held by a particular partnership as capital property is deemed to be that share of the loss determined without reference to this subsection minus,
. . .
(b) where the taxpayer is a corporation, the total of all amounts received by the taxpayer on the share each of which is
(i) a taxable dividend, to the extent of the amount of the dividend that was deductible under this section . . . in computing the taxpayer’s taxable income or taxable income earned in Canada for any taxation year,
. . .
[Emphasis added.]
[24] Subsection 112(3.1) reduces a taxpayer’s share of any loss from the “disposition of a share”. BMO says that its loss from the disposition of the NSULC common shares was not a loss from the “disposition of a share” because subsection 39(2) deemed it to be a loss from the “disposition of currency”.
[25] Subsection 39(2) has since been amended but, in 2010, the relevant portions of subsection 39(2) read as follows:
Notwithstanding subsection (1), where, by virtue of any fluctuation after 1971 in the value of the currency or currencies of one or more countries other than Canada relative to Canadian currency, a taxpayer has made a gain or sustained a loss in a taxation year, the following rules apply:
(a) the amount, if any, by which
(i) the total of all such gains made by the taxpayer in the year . . .
exceeds
(ii) the total of all such losses sustained by the taxpayer in the year. . . . , and
(iii) if the taxpayer is an individual, $200,
shall be deemed to be a capital gain of the taxpayer for the year from the disposition of currency of a country other than Canada, the amount of which capital gain is the amount determined under this paragraph; and
(b) the amount, if any, by which
(i) the total determined under subparagraph (a)(ii),
exceeds
(ii) the total determined under subparagraph (a)(i), and
(iii) if the taxpayer is an individual, $200,
shall be deemed to be a capital loss of the taxpayer for the year from the disposition of currency of a country other than Canada, the amount of which capital loss is the amount determined under this paragraph.
[Emphasis added.]
[26] The Respondent accepts that subsection 112(3.1) applies after subsection 39(2) applies and thus that a loss deemed by subsection 39(2) to be a loss from the disposition of currency would not be a loss from the disposition of shares for the purpose of subsection 112(3.1). Capital gains and losses determined under section 39 are used to determine a taxpayer’s income under section 3. A taxpayer’s income under section 3 is then used to calculate the taxpayer’s taxable income under section 2. Under subsection 2(2), the additions to and deductions from income found in Division C are made when calculating a taxpayer’s taxable income. Subsection 112(3.1) is found in Division C. Therefore, subsection 112(3.1) applies after subsection 39(2).
[27] BMO submits that, because subsection 39(2) applies before subsection 112(3.1), if subsection 39(2) applied to deem the foreign exchange loss on the NSULC common shares to be a capital loss from the disposition of currency, there would be no loss from the disposition of a share to which subsection 112(3.1) could apply. BMO argues that it cannot have received a tax benefit by avoiding the application of an otherwise inapplicable subsection.
[28] The Respondent accepts that BMO’s capital loss ensued from the fluctuation of the US dollar relative to the Canadian dollar. [7] The Respondent also accepts that, if subsection 39(2) applied to deem the foreign exchange loss on the NSULC shares to be a capital loss from the disposition of currency, then BMO did not receive a tax benefit. However, the Respondent says that subsection 39(2) did not apply in that manner. The question is therefore whether subsection 39(2) applied to the capital loss or not. BMO says it did and the Respondent says it did not.
[29] For the reasons that follow, I find that subsection 39(2) applied to deem the foreign exchange loss on the disposition of the NSULC common shares to be a capital loss from the disposition of currency and thus that BMO did not receive a tax benefit. I reach this conclusion based on a textual, contextual and purposive analysis of subsection 39(2).
i. Textual Analysis
[30] At first glance, subsection 39(2) appears to be straightforward. The text indicates that any gains or losses that arise from the fluctuation of the Canadian dollar will be deemed to be capital gains or losses from the disposition of currency. However, a closer examination reveals significant ambiguity.
[31] To understand the ambiguity in subsection 39(2), one must first recognize that a taxpayer can make a gain or sustain a loss on capital account by virtue of a fluctuation in the value of the Canadian dollar in two different ways: disposing of property or satisfying an obligation. [8]
(a) Disposing of property: A taxpayer can make a gain or sustain a loss by virtue of a fluctuation in the value of the Canadian dollar by disposing of capital property that is a foreign currency or is denominated in a foreign currency. For example, a taxpayer who held US dollars in a bank account on capital account could convert those US dollars into Canadian dollars. The conversion would be a disposition of property and the resulting gain or loss would have occurred by virtue of a fluctuation in the value of the Canadian dollar between the time the taxpayer acquired the US dollars and the time he or she disposed of them. In this example, the taxpayer’s gain or loss would be entirely related to the fluctuation in the Canadian dollar. A second example would be a taxpayer who owned shares in a US public company and sold those shares on the market. In this example, part of the taxpayer’s gain or loss would consist of a gain or loss on the underlying shares and the balance would consist of a foreign exchange gain or loss.
(b) Satisfying an obligation: A taxpayer can also make a gain or sustain a loss by virtue of a fluctuation in the value of the Canadian dollar by satisfying an obligation entered into on capital account that is denominated in a foreign currency. For example, a taxpayer who had borrowed money in US dollars to acquire a manufacturing plant could repay the loan at a point in time when the exchange rate had changed. The resulting gain or loss would be a gain or loss arising by virtue of a fluctuation in the value of the Canadian dollar. However, the gain or loss would not be a gain or loss from the disposition of property. This is because a loan payable is an obligation, not an asset.
[32] With this background in mind, the question is whether subsection 39(2) applies to foreign currency gains and losses arising from dispositions of capital property, the satisfaction of obligations or both.
Possible Interpretations
[33] I will refer to the interpretation of subsection 39(2) whereby the subsection catches only dispositions of capital property as the “Property Interpretation”, the interpretation whereby the subsection catches only foreign exchange gains or losses from the satisfaction of obligations on capital account as the “Obligation Interpretation” and the interpretation whereby the subsection catches foreign exchange gains or losses arising both from the satisfaction of obligations on capital account and the disposition of capital property as the “Broad Interpretation”. The Respondent supports the Obligation Interpretation. BMO supports the Broad Interpretation or a more restricted version of the Broad Interpretation (described below) which I will refer to as the “Restricted Broad Interpretation”. Neither party supports the Property Interpretation.
Property Interpretation
[34] I can quickly reject the Property Interpretation. As noted above, neither party supports this interpretation. Furthermore, the case law on subsection 39(2) indicates that the Property Interpretation is wrong. There are a number of reported cases in which subsection 39(2) has been applied to foreign exchange gains or losses that arose on the satisfaction of obligations. [9] Since the Property Interpretation would not allow subsection 39(2) to be applied in that manner, it is clearly incorrect.
Restricted Broad Interpretation
[35] In addition to the Broad Interpretation, BMO also argues in favour of the Restricted Broad Interpretation. I can quickly reject the Restricted Broad Interpretation as well.
[36] Under the Restricted Broad Interpretation, subsection 39(2) would apply to all foreign exchange gains and losses arising from the satisfaction of obligations on capital account but would only apply to foreign exchange gains and losses on the disposition of capital property if those gains or losses arose exclusively from fluctuations in the Canadian dollar. All other foreign exchange gains and losses from the disposition of capital property would be taxed under subsection 39(1).
[37] There is nothing whatsoever in the text of subsections 39(1) or (2) that supports the Restricted Broad Interpretation. Such an interpretation would require very specific language indicating that the gains made or losses sustained must not have been attributable to anything other than the fluctuation in the Canadian dollar. At a minimum, I would expect the word “only” or another word to that effect to have been inserted into the phrase “by virtue of” in subsection 39(2) such that it read “by virtue only of”.
[38] BMO’s sole support for the Restricted Broad Interpretation comes from a position that the Canada Revenue Agency took at a conference of the Association de planification fiscale et financière in 2005. The CRA stated the following: [10]
The CRA’s position, in respect of a gain or loss resulting from the disposition of a property that is a capital property, is that subsection 39(2) of the ITA will apply if, and only if, the gain or loss, as calculated in accordance with section 40 . . . , is solely attributable to the fluctuation in the value of a foreign currency relative to the Canadian currency. If the aforesaid gain or loss is not solely attributable to the fluctuation in the value of a foreign currency relative to the Canadian currency, subsection 39(1) . . . and not subsection 39(2) . . . , must be used to calculate the capital gain or loss resulting from the disposition of the property.
[Emphasis added.]
[39] The CRA provided no textual, contextual or purposive explanation of why it believed subsection 39(2) operated in this manner. I find this published position to be of no value in interpreting subsection 39(2).
[40] Based on all of the foregoing, I reject the Restricted Broad Interpretation. Having done so, I can now focus the textual analysis on the two remaining interpretations: the Obligation Interpretation and the Broad Interpretation.
“made a gain or sustained a loss”
[41] Subsection 39(2) uses the phrase “made a gain or sustained a loss”. This phrase supports both the Obligation Interpretation and the Broad Interpretation.
[42] Subsection 39(1) deals with capital gains and losses. It describes a capital gain as a “gain . . . from the disposition of any property” and a capital loss as a “loss . . . from the disposition of any property”. These phrases do not appear in subsection 39(2). Instead, subsection 39(2) refers to a taxpayer who has “made a gain or sustained a loss” without specifying how those gains were made or those losses were sustained. This different language in subsection 39(2) clearly indicates that Parliament did not intend that subsection 39(2) would only apply to dispositions of property. Had Parliament intended that result, it would have used the same language that it used in subsection 39(1).
[43] While the phrase “made a gain or sustained a loss” is certainly textual support for the rejection of the Property Interpretation, it does not assist me in choosing between the Obligation Interpretation and the Broad Interpretation. Parliament may have used different language because it wanted to cover gains or losses that did not arise from dispositions of property (i.e. obligations). Alternatively, Parliament may have used different language because it wanted to cover both obligations and dispositions of property.
Calculations of Foreign Exchange Gains and Losses
[44] Under the Broad Interpretation, taxpayers would have to separate their gains or losses on the property itself from their gains or losses arising from fluctuations in the Canadian dollar. The former would be taxed under subsection 39(1) and the latter under subsection 39(2). The Respondent submits that such treatment would be redundant. The Respondent further submits that the lack of instructions from Parliament as to how subsections 39(1) and (2) should interact argues against this interpretation. For the reasons that follow, I disagree with both of these arguments.
[45] The Respondent submits that the application of subsection 39(2) to foreign exchange gains and losses arising from the disposition of property would be redundant. The amount of those gains and losses is calculated under section 40. Pursuant to subsection 261(2), that calculation is required to be done in Canadian dollars and thus already takes into account foreign currency gains and losses. [11] As a result, the Respondent submits that the portion of any gain or loss arising from the fluctuation of the Canadian dollar is already factored into the overall gain or loss in subsection 39(1). I agree that, absent subsection 39(2), foreign exchange gains and losses from the disposition of capital property would be taxed under subsection 39(1). However, that does not mean that Parliament could not have decided to separate those gains and losses and instead tax them under subsection 39(2). As discussed in more detail below, subsection 39(2) applies notwithstanding subsection 39(1).
[46] The Respondent also points out that there are no instructions in subsections 39(1) and (2) indicating how taxpayers should separate their gains and losses on underlying property from their gains and losses arising from fluctuations in the Canadian dollar. The Respondent submits that the absence of such instructions indicates that Parliament did not intend subsection 39(2) to cover dispositions of property and thus argues in favour of the Obligation Interpretation. I disagree.
[47] It could be, as the Respondent argues, that there are no instructions in subsection 39(1) and (2) because Parliament did not intend subsection 39(2) to apply to dispositions of property. However, it could also be that there are no instructions because Parliament did not think it was necessary to include them. It does not strike me that it would be particularly difficult to separate one’s gain or loss on an underlying asset from one’s foreign exchange gain or loss in most situations. [12]
[48] Additionally, section 261 was only introduced in 2007. With some minor exceptions, it sets out how foreign currency transactions are to be handled for all purposes of the Act. Prior to the introduction of section 261, the Act contained no rules describing how foreign currency transactions were to be handled. Therefore, from a historical perspective, the absence of guidance from subsections 39(1) or (2) is neither troubling nor revealing.
[49] Based on all of the foregoing, I find that there is no redundancy under the Broad Interpretation and that the lack of instructions in subsections 39(1) and (2) does not weaken the Broad Interpretation nor is it inconsistent with it.
“Notwithstanding subsection (1)”
[50] Subsection 39(2) opens with the phrase “[n]otwithstanding subsection (1)”. This phrase could be interpreted in three different ways. Two of those interpretations support the Property Interpretation and, consequently, should be rejected. The third interpretation of the phrase supports either the Obligation Interpretation or the Broad Interpretation.
(a) The phrase “[n]otwithstanding subsection (1)” could mean “despite the fact that this gain or loss is already taxed under subsection (1)”. Since obligations are not taxed under subsection 39(1), this interpretation of the phrase would only work under the Property Interpretation and must therefore be rejected.
(b) The phrase “[n]otwithstanding subsection (1)” could also be interpreted to mean “instead of taxing these foreign exchange gains or losses under subsection (1)”. Again, since obligations are not taxed under subsection 39(1), this interpretation of the phrase would only work under the Property Interpretation and must therefore be rejected.
(c) Finally, the phrase “[n]otwithstanding subsection (1)” could be interpreted to mean “despite the definitions in subsection (1)”. This interpretation is consistent with both the Obligation Interpretation and the Broad Interpretation. [13] Under the Broad Interpretation, the phrase would indicate that, despite the fact that foreign exchange gains and losses from the disposition of property already meet the definition of capital gains and losses in subsection 39(1), Parliament is going to deem them to be a different type of capital gain or loss: a capital gain or loss from the disposition of currency. Under the Obligation Interpretation, the phrase would indicate that, despite the fact that foreign exchange gains and losses from the satisfaction of obligations would not otherwise fit the definition of capital gains and losses in subsection 39(1), Parliament is going to deem them to be capital gains or losses.
[51] Based on all of the foregoing, I find that, from a textual perspective, the phrase “[n]otwithstanding subsection (1)” provides support for either the Obligation Interpretation or the Broad Interpretation.
Case Law
[52] As noted above, there are a number of decisions in which subsection 39(2) has been applied to obligations. Unfortunately, these decisions do not assist me in determining the correct approach since both the Obligation Interpretation and the Broad Interpretation cover obligations. I am not aware of any case where a court has applied subsection 39(2) to a disposition of property. [14] However, that does not mean that the subsection could not be applied in that manner.
CRA Positions
[53] The CRA has changed its mind about subsection 39(2) so many times that its views are hardly helpful. At best CRA’s conflicting administrative positions underline the textual ambiguity of the provision. The first interpretation bulletin that the CRA issued on the topic indicated that subsection 39(2) only applied to dispositions of foreign currency. [15] I cannot see any textual interpretation of the subsection that would support that view. An updated version of that interpretation bulletin indicated that subsection 39(2) applied to both the satisfaction of “capital debt obligations” and the disposition of foreign currency by any taxpayer. [16] Later, in the position noted in the discussion of the Restricted Broad Interpretation above, the CRA interpreted subsection 39(2) to apply to the satisfaction of obligations, the disposition of foreign currency and the disposition of any other property so long as the gain or loss arose solely from a fluctuation in the Canadian dollar. As described above, there is nothing in the text of the subsection that would support this view.
Commentary
[54] The ambiguity of subsection 39(2) has not gone unnoticed in the tax community. In describing the subsection, the authors of Timing and Income Taxation stated that “it was unclear whether subsection 39(2) was the operative provision for including in income all of a taxpayer’s foreign-currency gains and losses, or whether it was intended only to capture those foreign-currency gains and losses that were not otherwise included under subsection 39(1)”. [17] Other commentators have similarly noted the ambiguity of subsection 39(2). [18]
Summary
[55] In summary, the text of subsection 39(2) is ambiguous. It could support either the Obligation Interpretation or the Broad Interpretation. As the text does not support the Property Interpretation or the Restricted Broad Interpretation, I will not consider those interpretations in the contextual or purposive analyses.
ii. Contextual Analysis
[56] A contextual analysis of subsection 39(2) reveals weaknesses in the Obligation Interpretation and supports the Broad Interpretation.
Subsection 39(3)
[57] Subsection 39(3) deals with gains or losses that may arise when a taxpayer who has issued bonds, debentures or similar instruments repurchases those obligations. It deems such gains or losses to be capital gains or capital losses. This deeming provision is necessary because the taxpayer is not disposing of property and thus is not caught by the normal capital gain and loss provisions in subsection 39(1).
[58] The relevant portions of subsection 39(3) read:
Where a taxpayer has issued any bond, debenture or similar obligation and has at any subsequent time in a taxation year and after 1971 purchased the obligation in the open market, . . .
(a) the amount, if any, by which the amount for which the obligation was issued by the taxpayer exceeds the purchase price paid or agreed to be paid by the taxpayer for the obligation shall be deemed to be a capital gain of the taxpayer for the taxation year from the disposition of a capital property, and
(b) the amount, if any, by which the purchase price paid or agreed to be paid by the taxpayer for the obligation exceeds the greater of the principal amount of the obligation and the amount for which it was issued by the taxpayer shall be deemed to be a capital loss of the taxpayer for the taxation year from the disposition of a capital property,
. . .
[Emphasis added.]
[59] An analysis of subsection 39(3) reveals two key points. First, when dealing with gains and losses, if Parliament wants to describe something that is an obligation, it uses the word “obligation”. The word appears multiple times in subsection 39(3). [19] In light of this, the absence of the word “obligation” in subsection 39(2) significantly undermines the Obligation Interpretation. It suggests that Parliament did not use that word because it wanted subsection 39(2) to catch more than just obligations. It wanted the subsection to also catch foreign exchange gains and losses from dispositions of capital property.
[60] The second point that the contextual analysis of subsection 39(3) reveals involves the phrase “notwithstanding subsection (1)”. That phrase does not appear in subsection 39(3). Subsections 39(1), (2) and (3) were all introduced at the same time. The presence of the phrase “[n]otwithstanding subsection (1)” in subsection 39(2) and the corresponding absence of the same phrase in subsection 39(3) must therefore be presumed to have been intentional. So why does the phrase not appear in subsection 39(3)? Subsection 39(3) deals with obligations. It deems the gains or losses from those obligations to be capital gains or losses. Under the Obligation Interpretation, this is exactly what subsection 39(2) would do to foreign exchange gains and losses on obligations. So if both provisions deem gains and losses on obligations to be capital gains and losses, why would Parliament think it needed to clarify that subsection 39(2) applied notwithstanding subsection 39(1) but not make the same clarification in subsection 39(3)? The absence of a logical explanation suggests that the Obligation Interpretation is not the correct interpretation.
[61] In the textual analysis, I concluded that the phrase “[n]otwithstanding subsection (1)” could support either the Obligation Interpretation or the Broad Interpretation if one interpreted the phrase as meaning “despite the definitions in subsection (1)”. However, the contextual analysis of subsection 39(3) now reveals that Parliament did not consider this phrase necessary when deeming a gain or loss from an obligation to be a capital gain or loss. If subsection 39(2) only applied to obligations there would be no need for the phrase. This undermines the Obligation Interpretation and supports the Broad Interpretation. It indicates that Parliament inserted the phrase “[n]otwithstanding subsection (1)” into subsection 39(2) because Parliament intended subsection 39(2) to cover things that would otherwise be caught by subsection 39(1) – dispositions of property. [20]
Summary
[62] Based on all of the foregoing, I find that the contextual analysis of subsection 39(2) supports the Broad Interpretation and significantly undermines the Obligation Interpretation.
iii. Purposive Analysis
[63] The purposive analysis of subsection 39(2) supports the Broad Interpretation.
[64] Both parties agree that subsection 39(2) catches foreign exchange gains and losses arising from the satisfaction of obligations. There is a good policy reason why Parliament would want the subsection to do so. As set out above, such gains and losses are not caught by subsection 39(1) as they do not result from the disposition of property. If they were not caught by subsection 39(2), they would be left untaxed. The question is whether Parliament intended to limit the application of subsection 39(2) to obligations or whether it also wanted subsection 39(2) to catch foreign exchange gains and losses from the disposition of property?
Deeming Provision
[65] Subsection 39(2) is a deeming provision. Deeming provisions are designed to create legal fictions that differ from reality. [21] In the case of subsection 39(2), the subsection deems a foreign exchange gain or loss to be a capital gain or loss that arose from the disposition of currency. There are two parts to the deeming provision. The first part deems the gain or loss to be a capital gain or loss. The second part deems that capital gain or loss to be from the disposition of a certain type of property, namely a foreign currency. I will examine these two parts of the deeming provision separately.
(a) Deemed to be a capital gain or loss: The first part of the deeming provision deems a foreign exchange gain or loss to be a capital gain or loss. When dealing with obligations, the purpose of this part of the deeming provision is clear. As discussed above, obligations are not property so a gain or loss arising from satisfying an obligation is not caught by the definitions of capital gain or capital loss in subsection 39(1). Therefore, if Parliament wanted to tax the gain or loss as a capital gain or loss, it needed to deem it to be a capital gain or loss. By contrast, when dealing with a disposition of property, the first part of the deeming provision is entirely redundant. A disposition of property already falls within the definitions of capital gain or loss under subsection 39(1) so there is no need to deem it to be a capital gain or loss. In summary, the first part of the deeming provision supports both the Obligation Interpretation and the Broad Interpretation. Both interpretations cover obligations, therefore both interpretations need to deem gains and losses from the satisfaction of those obligations to be deemed to be capital gains or losses. The fact that the first part of the deeming provision only affects some of the gains and losses caught by the Broad Interpretation does not mean it is unnecessary.
(b) Deemed to arise from the disposition of a foreign currency: The second part of the deeming provision deems the capital gains and losses to have arisen from the disposition of a foreign currency as opposed to some other type of property. I cannot see any purpose to th