Date: 20241128
Dockets: A-199-21
A-200-21
A-201-21
Citation: 2024 FCA 202
CORAM:
|
BOIVIN J.A.
LOCKE J.A.
MONAGHAN J.A.
|
BETWEEN: |
Docket: A-199-21 |
MAGREN HOLDINGS LTD. |
Appellant |
and |
HIS MAJESTY THE KING |
Respondent |
AND BETWEEN: |
Docket: A-200-21 |
2176 INVESTMENTS LTD., (as successor to Grencorp Management Inc. which was the successor to 994047 Alberta Ltd.) |
Appellant |
and |
HIS MAJESTY THE KING |
Respondent |
AND BETWEEN: |
Docket: A-201-21 |
MAGREN HOLDINGS LTD. (successor by amalgamation to 1052785 Alberta Ltd.) |
Appellant |
and |
HIS MAJESTY THE KING |
Respondent |
REASONS FOR JUDGMENT
MONAGHAN J.A.
[1] In 2006, three corporations controlled by James Grenon paid dividends aggregating more than $110 million, and filed elections so that they would be capital dividends. Capital dividends are not taxable to the recipient. However, where a corporation pays a capital dividend in excess of the balance of its capital dividend account, the corporation is liable for tax.
[2] The Minister of National Revenue concluded that the three corporations did exactly that. Accordingly, the Minister issued assessments imposing tax on the basis that all of the capital dividends those corporations paid in 2006 were excess dividends.
[3] For several reasons, the Tax Court of Canada dismissed the appeals of those assessments (2021 TCC 42 per Smith J.). It agreed with the Minister that Mr. Grenon and the corporations did not achieve the results they thought they had. The question we face is whether the Tax Court erred in doing so.
[4] Although I do not agree with all of the Tax Court’s conclusions, and disagree with many of its obiter statements, I would dismiss the appeals.
I. Background
A. Factual Overview
[5] The series of transactions at the centre of this appeal is complex and must be reviewed in some detail. It includes a reorganization of Foremost Industries Income Fund (FMO), a publicly-traded mutual fund trust, in 2005. Before the series began, members of the public owned approximately 42 percent of FMO’s outstanding units. A self-directed registered retirement savings plan (RRSP Trust) owned the remaining 58 percent. James Grenon was the sole annuitant of RRSP Trust.
[6] In late 2005, Mr. Grenon proposed a reorganization to FMO. The stated objectives of the reorganization included increasing the tax cost of FMO’s underlying business assets. To accomplish this, FMO would transfer all of its business operations to a newly established mutual fund trust (FIF) on a taxable basis, and FMO unitholders would exchange their FMO units for FIF units on a one-for-one basis, also on a taxable basis.
[7] For the public FMO unitholders, that is exactly what happened. The business assets were sold to FIF and the public exchanged their FMO units for FIF units. In doing so, the public realized a gain (or loss) if the value of the FIF units they received exceeded (or was less than) their tax cost of the FMO units given in exchange. As far as the public unitholders were concerned, once this occurred the FMO reorganization was largely complete.
[8] While RRSP Trust also owned FIF units at the end of the series, the steps by which it acquired them bore no resemblance to the steps by which the public acquired their FIF units. Along the way, the appellants acquired FMO units and participated in the FMO reorganization in a manner that resulted in no taxable income for them, but an aggregate $113 million addition to their capital dividend accounts. The appellants then paid capital dividends of more than $110 million to their parent corporations, which added those dividends to their capital dividend accounts. This permitted the parent corporations to distribute to Mr. Grenon, as non-taxable capital dividends, approximately $110 million, amounts that they otherwise could only have distributed as taxable dividends.
[9] Before I describe the details of the transactions, a brief summary of certain key income tax principles concerning the taxation of trusts and the operation of the capital dividend account (CDA) is useful.
B. Key Relevant Taxation Principles
[10] Except as expressly noted, in these reasons all references to statutory provisions are to provisions of the Income Tax Act, R.S.C. 1985, c. 1 (5th Supp.), as they read at the time the relevant transactions occurred. The relevant provisions are duplicated in Appendix B to these reasons. Many were amended subsequently and these reasons must be read in that light.
[11] Only 50 percent of a capital gain realized by a taxpayer—the taxable capital gain—must be included in income for tax purposes. Where the taxpayer has a capital loss, 50 percent of that loss—the allowable capital loss—is deductible against taxable capital gains, but is not otherwise deductible in computing income.
(1) Taxation of trusts and their beneficiaries
[12] A trust is a taxpayer and therefore must compute its income and taxable income. However, a trust may avoid a tax liability on its income by paying an equal amount to its beneficiaries, who then must include the payment in their income: ss. 104(6)(b), 104(13). The same is not true of losses. Where a trust realizes a loss, it cannot pass that loss out to its beneficiaries; only the trust may use its losses.
[13] Where a trust pays its income to its beneficiaries, any liability for tax on the trust’s income is determined based on the beneficiary’s circumstances. If the beneficiary is a registered retirement savings plan, or other tax-exempt plan, no tax is payable until the income is withdrawn from the plan, typically many years later. If the beneficiary is itself a trust, it in turn may pay the income to its beneficiaries so it becomes their income rather than income of the trust.
[14] With few exceptions, the character of the trust’s income is not maintained when paid to a beneficiary; rather, the beneficiary has income from a property that is an interest in a trust: s. 108(5)(a). Capital gains are one exception to that general principle relevant to this appeal.
[15] Where a trust realizes a capital gain, and distributes an equal amount to a beneficiary, the trust may make a designation so that one-half of the distribution is deemed to be a taxable capital gain realized by the beneficiary: s. 104(21). The other portion of the distribution—reflecting the non-taxable portion of the trust’s capital gain—is not included in the beneficiary’s income. Moreover, the distribution does not affect the beneficiary’s tax cost—the adjusted cost base (ACB)—of their interest as beneficiary in the trust: s. 53(2)(h)(i.1)(A) and (B). That ACB remains unchanged regardless of the distribution’s effect on the value of that interest.
[16] In this way, provided the appropriate designation is made, the trust’s capital gain is taxed as if the beneficiary had realized it directly.
[17] When a trust purchases for cancellation (redeems) a beneficiary’s interest in the trust, the trust may choose to treat part of the amount it pays for that interest as a distribution of its income, rather than an amount paid to acquire the interest. In that circumstance, only the amount in excess of the income distribution will be proceeds of disposition for the beneficiary’s interest. Consequently, whether the beneficiary has a capital gain (or capital loss) depends on whether those reduced proceeds exceed (or are less than) the beneficiary’s ACB of the repurchased (redeemed) interest.
[18] If the trust’s income includes taxable capital gains, the trust may choose to treat part of the amount it pays on the redemption as a distribution of a capital gain, designating 50 percent as a taxable capital gain. While the distributed capital gain reduces the beneficiary’s proceeds of disposition for the interest in the trust as described in the preceding paragraph, the beneficiary’s overall capital gain (or capital loss) should be the same, albeit comprised of two amounts.
[19] In particular, the capital gain distributed by the trust would reduce the proceeds of disposition for the redeemed trust interest, and would require the beneficiary to include 50 percent as a taxable capital gain in income. The beneficiary would also have a capital gain (or capital loss) on the trust interest redeemed, depending on whether the reduced proceeds of disposition (reflecting the distributed capital gain) exceed (or are less than) the beneficiary’s ACB of the redeemed trust interest. Any resulting allowable capital loss would be deductible against the distributed taxable capital gain.
(2) The capital dividend account
[20] The capital dividend account (CDA) is an important part of what is typically referred to as the integration system in the Income Tax Act. Broadly speaking, the system’s goal is to tax income at the same rate whether earned directly by an individual or by a private corporation of which the individual is a shareholder. To do this, the system integrates—or combines—the taxes paid by the corporation and by the individual shareholder on a dividend of the corporation’s after-tax income to so that they roughly equal the tax the individual would pay had they earned the income directly. The system does this through different mechanisms, of which the CDA is one.
[21] The CDA is a notional account in which a private corporation tracks certain tax-free surpluses that it accumulates over time. A corporation’s CDA balance at a particular time is determined by adding and subtracting specified amounts that have arisen before that time. Two additions related to the non-taxable portion of a capital gain, albeit from different sources, are relevant here. They are:
The positive difference between the non-taxable portion of all capital gains and the non-deductible portion of all capital losses that the corporation has had from dispositions of property before the calculation time: see paragraph (a) of the definition of CDA in s. 89(1); and
The non-taxable portion of any capital gain that a trust distributes to the corporation, as a beneficiary of the trust, before the calculation time: paragraph (f) of the definition of CDA in s. 89(1).
[22] Notably, capital losses the corporation realizes are relevant only to the first CDA addition—that in paragraph (a). If, at the CDA calculation time, a corporation’s cumulative capital losses exceed its cumulative capital gains from dispositions of property, there is no positive amount described in paragraph (a). The resulting deficit precludes any addition to the CDA under paragraph (a) until the corporation realizes sufficient capital gains to eliminate it (i.e., until the corporation has an excess of cumulative capital gains over cumulative capital losses). However, that deficit has no other effect on the corporation’s CDA, including on the addition under paragraph (f) of the definition. This is of particular significance to this appeal.
[23] Where a trust pays an amount equal to its capital gain to the corporation (beneficiary), designating 50 percent of the payment as a taxable capital gain, the corporation adds the other (non-taxable) 50 percent to its CDA under paragraph (f) without regard to any capital losses previously realized by the corporation.
[24] When a corporation has a positive CDA balance, it may pay a capital dividend to its shareholders by making an appropriate election: s. 83(2). Capital dividends paid reduce the corporation’s CDA but are not taxable to the recipient: closing words of the CDA definition and paragraph (a) of the definition of taxable dividend in s. 89(1), 82(1). If the recipient is a private corporation, the capital dividends are added to its CDA: paragraph (b) of the CDA definition in s. 89(1).
[25] With that background, I turn now to the transactions giving rise to the assessments at issue in this appeal.
II. The FMO Reorganization
[26] Some introductory comments are necessary.
[27] First, the three corporations that paid the capital dividends no longer exist as separate entities; the appellants are their successors by amalgamation. As nothing turns on this, I refer to the appellants as if they existed at all relevant times.
[28] Second, the appellants participated in the series of transactions in the same way, albeit for different amounts. For simplicity, rather than describe amounts relevant to each appellant, I describe the amounts on an aggregate basis for the appellants as a group.
[29] Third, certain unit trusts also participated in the series of transactions. In addition to FMO and FIF, Foremost Ventures Trust (FVT) and TOM 2003-4 Income Fund (TOM) played significant roles. References to unitholders and units refer to the beneficiaries and their interests in the trust as beneficiaries, respectively.
[30] Fourth, many transactions are not relevant to the issues in this appeal and I need not describe them in detail. My focus is the most relevant transactions. For this reason, where the details of the transactions undertaken to achieve certain results are not germane to the issues, I may only describe the result. The relevant transactions are described in detail in Annex C attached to the Tax Court’s reasons. In these reasons, references to Steps refer to the Steps as they appear in that Annex.
[31] Finally, I simplify wherever possible. Therefore, I have largely ignored FILP and FULP, two partnerships collectively owned by FMO and FVT. Although they owned the operating businesses transferred to FIF, those partnerships were wound up and FVT acquired their assets: see Steps 6 and 12. To simplify, I typically describe FVT as if it were the party to certain transactions. Similarly, I ignore the trusts and partnerships that comprised the FIF structure, and instead refer simply to FIF. Because I use approximate (i.e., rounded) numbers of units, percentages and values, they do not always add to the “correct”
totals. However, Annex C to the Tax Court’s reasons provides precise amounts. I also largely ignore income and capital gains from other sources. For example, while FMO reported relatively modest capital gains other than as related to the reorganization, for simplicity I treat all the capital gains as realized in the FMO reorganization.
A. Pre-reorganization Structure
[32] In 2005, FMO was a publicly-traded mutual fund trust. FMO directly owned all of the units in another trust, FVT. FMO and FVT collectively owned 99 percent of two limited partnerships (FULP and FILP), each of which carried on an operating business. At the end of 2005, the fair market value of the operating business assets, most notably the goodwill, exceeded their tax cost by more than $210 million.
[33] RRSP Trust owned 11 million FMO units, representing 58 percent of FMO. The remaining 8 million units, representing 42 percent, were owned by members of the public. Because it is important to distinguish the two groups of FMO units, I will refer to the former as the RRSP FMO units and the latter as the public FMO units. CIBC Trust Corporation was the trustee of RRSP Trust (Trustee).
[34] The simplified structure before the series of transactions commenced may be illustrated as follows:
B. Pre-reorganization Transactions: The Appellants Acquire the RRSP FMO units
[35] In November 2005, James Grenon, annuitant of RRSP Trust and one of FMO’s three trustees, proposed a reorganization to FMO. Its trustees agreed to put the proposal before the unitholders for a vote and, on November 10, 2005, FMO issued a press release announcing the proposed reorganization and unitholders’ meeting: Appeal Book at 10932.
[36] The circular calling the unitholders’ meeting described the proposed reorganization’s objectives as threefold: (i) to increase the cost for tax purposes of the business assets; (ii) to simplify FMO’s organizational and governance structure; and (iii) to attract a wider retail investor base for the FMO units: reasons at paras. 34-35.
[37] To accomplish this, the circular explained, the reorganization would see FMO’s structure largely replicated under a newly established mutual fund trust, FIF. FMO unitholders would exchange their FMO units on a one-for-one basis for FIF units: reasons at para. 33. As the Tax Court explained it, “[a]s far as the public unitholders were concerned, the FMO units would be exchanged on a one-for-one basis for new units of FIF, there would be no change to existing management or the underlying business operations and the new units would continue to trade on the Toronto Stock Exchange under the same ticker symbol.”
: reasons at para. 212.
[38] On November 14, 2005, RRSP Trust subscribed for units in TOM for $153 million and satisfied the subscription price by transferring the RRSP FMO units to TOM: Step 1. Prior to the subscription, TOM had relatively nominal assets, with the result that RRSP Trust became the holder of approximately 99.5 percent of TOM’s outstanding units: reasons at para. 21.
[39] On Friday December 23, 2005, the last business day before the FMO unitholders’ meeting, the appellants acquired the RRSP FMO units from TOM for $161 million: Step 2. As a result, TOM realized a taxable capital gain of $3.9 million: TOM 2005 T3 Trust Income Tax and Information Return, Appeal Book 10716-10730 at 10724. The appellants issued promissory notes, guaranteed by Mr. Grenon, to satisfy the purchase price: reasons at para. 44.
[40] On December 28, 2005, the FMO unitholders approved the FMO reorganization and the steps to effect it commenced and were completed later that same day.
[41] However, as we will see, the manner in which the holders of the public FMO units participated differed significantly from the manner in which the appellants, as holders of the RRSP FMO units, participated.
C. FMO Reorganization: Relevant Transactions
[42] Broadly speaking, the FMO reorganization may be divided into two successive phases.
(1) First phase of the FMO reorganization
[43] The first phase, comprising Steps 3-10, achieved three principal objectives. FIF acquired the operating business assets from the FMO partnerships; FVT acquired one FIF unit for each outstanding FMO unit; and the public exchanged the public FMO units for FIF units on a one-for-one basis. While the transactions comprising the first phase are not relevant, certain results are important.
[44] These transactions were effected on a taxable basis. Therefore when the operating business assets were sold, FVT realized $105 million of income as a result of the $210 million gain realized on disposition of the goodwill. (Although not a capital gain, similarly only 50 percent of this gain was included in income: s. 14(1); see also FVT’s amended T3 Income Tax and Information Return for 2005, Appeal Book 6536-51 at 6541 and FVT-Amended 2005 T3 Income Allocations Schedule at 6532.) FIF acquired the operating assets with a fair market value tax cost, achieving one of the stated objectives of the FMO reorganization. Finally, FVT acquired the 19 million FIF units necessary to effect the one-for-one exchange for FMO units with a tax cost equal to their $277 million fair market value.
[45] The public FMO unitholders realized capital gains (or capital losses) when FVT acquired their FMO public units in exchange for FIF units, depending on whether the value of the FIF units received exceeded (or was less than) their ACB of the FMO units given in exchange: see FMO Notice of Special Meeting of Unitholders and Information Circular and Proxy Statement, Appeal Book 6111-6230 at 6137-8. However, FVT did not realize a gain or loss on the exchange because it both acquired and disposed of the FIF units for the same amount on the same day. FVT acquired the public FMO units with a cost equal to their $115 million fair market value.
[46] At the end of the first phase, FIF owned the operating businesses and the public no longer had any interest in FMO. FVT’s only significant assets were 11 million FIF units, one for each of the RRSP FMO units, and the 8 million public FMO units acquired in exchange for FIF units. FMO’s assets consisted primarily of its FVT units and a $44.5 million promissory note owing by FVT. As a result, FMO and FVT each owned assets with an aggregate $277 million value. The resulting simplified structure at the end of the first phase may be illustrated as follows:
[47] I turn now to describe the key transactions in the second phase.
(2) Second phase of the FMO reorganization
[48] First the appellants acquired the public FMO units from FVT issuing $115 million promissory notes in exchange: Step 11. The appellants thus became holders of all 19 million outstanding FMO units with a total cost of $277 million comprising $161 million paid to TOM for the RRSP FMO units and $115 million paid to FVT for the public FMO units. As we shall see, this $277 million cost is a significant factor in this appeal.
[49] FVT repaid the $44.5 million notes owing to FMO by delivering 3 million FIF units: Step 13. This transaction resulted in no gain or loss and FMO acquired the 3 million FIF units with a cost equal to their $44.5 million value. This left FVT with total assets of $232 million comprising 8 million FIF units and the $115 million promissory notes issued by the appellants.
[50] The resulting simplified structure after this transaction may be illustrated as follows:
[51] FMO then sold its FVT units to TOM for $232 million: Step 14. TOM satisfied the purchase price by transferring the appellants’ $161 million promissory notes (acquired when TOM sold the RRSP FMO units to the appellants), and issuing its own $72 million promissory note, to FMO. Because the $232 million TOM paid far exceeded FMO’s ACB of the FVT units, FMO realized a significant capital gain: FMO T3 Trust Income Tax and Information Return, Appeal Book 12677-12689 at 12681. As we shall see, FMO distributed a $226 million capital gain to the appellants, leading to the $113 million CDA addition that is in dispute in this appeal.
[52] The resulting simplified structure after these transactions may be illustrated as follows:
[53] As is evident, FMO’s only significant assets, with an aggregate value of $277 million, were the $232 million of promissory notes acquired from TOM and the 3 million FIF units acquired from FVT.
[54] Moreover, although the FMO reorganization contemplated that each FMO unit would be exchanged for a FIF unit, the RRSP FMO units had not been exchanged. The 11 million FIF units necessary to make the exchange were owned as to 3 million by FMO and as to 8 million by FVT, in turn wholly-owned by TOM.
[55] FMO then distributed $277 million of its assets to the appellants, the only remaining FMO unitholders: Step 15. As a result, the appellants acquired 3 million FIF units and the $72 million TOM promissory note, and their $161 million promissory notes were cancelled. Thus, they held assets acquired from FMO with an aggregate fair market value equal to their outstanding indebtedness to FVT under the notes they issued to acquire the public FMO units from FVT.
[56] FMO treated $226 million of the $277 million distributed to the appellants as a distribution of its capital gain. To do so, FMO designated $113 million as a taxable capital gain. The balance of FMO’s $277 million distribution reduced the appellants’ ACB of their FMO units: s. 53(2)(h).
[57] The appellants included the taxable capital gain in their income and added the non-taxable portion, also $113 million, to their CDA relying on paragraph (f) of the definition—the non-taxable portion of the capital gain distributed by a trust, FMO.
[58] The distribution left FMO with no remaining income and nominal assets with nominal value. Thus, the value of the appellants’ FMO units was also nominal. The resulting simplified structure after these transactions may be illustrated as follows:
[59] FMO then redeemed substantially all of its outstanding units from the appellants for their nominal value: Step 16. Because the appellants’ ACB of the FMO units was significantly greater, the appellants claimed a $232 million capital loss, and deducted $113 million of the resulting allowable capital loss against their $113 million taxable capital gain resulting from FMO’s distribution. This left the appellants with no taxable income.
[60] The other half of the appellants’ capital loss created a deficit in their CDA under paragraph (a) of the definition. However, that deficit was of no consequence to the paragraph (f) addition from the taxable capital gain FMO distributed in Step 15.
[61] In its 2005 taxation year, FVT’s income was $137 million, comprising $105 million from the sale of the operating businesses (described at paragraph 44 above) and $32 million from the operating businesses’ 2005 operations. FVT distributed substantially all of its assets, consisting of 8 million FIF units and the appellants’ $115 million promissory notes, to TOM: Step 17. FVT treated $137 million of this distribution as a payment of income. TOM therefore included the $137 million in its 2005 income: FVT Statement of Trust Income Allocations and Designations, Appeal Book at 6551. The balance of the distribution would have reduced TOM’s ACB of its FVT units, but because FVT had distributed substantially all its assets to TOM, those units had nominal value.
[62] I pause here to note that the FMO reorganization agreement included in the circular sent to FMO unitholders was amended on the date of the unitholders’ meeting: reasons at para. 34. As amended, it contemplated that FMO would “sell its remaining interest in [FVT] to certain participating remaining [FMO] unitholders”
: section 2.1(n), Reorganization Agreement Amendment, Appeal Book 4936-40, at 4938 (emphasis added). As the diagram at paragraph 50 above illustrates, immediately before FMO sold FVT, the appellants were the only remaining FMO unitholders. But TOM, not the appellants, acquired FVT. Before the Tax Court, the appellants submitted that they assigned their right to acquire FVT to TOM. While conceding that the assignment was not documented, they asserted that the parties agreed to it, there was no contrary evidence, and TOM did in fact acquire FVT: reasons at para. 189.
[63] Had FMO transferred FVT to the appellants—the only participating remaining FMO unitholders—instead of TOM, the tax consequences would have been entirely different. This is so because FVT needed to distribute its $137 million of income if it did not want to be liable for tax. Had the appellants acquired FVT, and FVT distributed that income, the appellants would have had to include the distribution in their income. FVT’s income did not include taxable capital gains and, if distributed to the appellants, would not have led to a CDA addition. Moreover, the appellants would not have been able to deduct any allowable capital losses they realized (when FMO redeemed its units) against any income FVT distributed to them.
[64] Whatever the original plan, it is clear that TOM acquired FVT from FMO so that FMO would have a substantial capital gain, $137 million of which was attributable to FVT’s underlying income, and so that substantially all of FVT’s income could be distributed to RRSP Trust thereby avoiding any current liability for taxes on that income, as described below in paragraph 70.
[65] I return now to the remaining steps in the FMO reorganization.
[66] TOM set off its obligation under its $72 million promissory note against an equal amount owing under the appellants’ $115 million promissory notes: Step 17. This left the appellants owing TOM an amount equal to the value of their 3 million FIF units.
[67] The resulting simplified structure after these transactions may be illustrated as follows:
[68] As is clear, despite the promised one-for-one exchange of FMO units for FIF units, the appellants, which owned 11 million RRSP FMO units when the FMO reorganization began, received only 3 million FIF units. The other FIF units that they “should have”
received were held by TOM.
[69] In January 2006, FMO was dissolved, completing the FMO reorganization: Step 18(b).
D. Post FMO Reorganization Transactions: Capital Dividends and other 2006 Transactions
[70] In March 2006, TOM purchased 1.4 million FIF units from the appellants and paid by issuing a promissory note: Step 19. TOM’s 2005 income was $143 million (which included the $137 million FVT distributed to TOM and the $3.9 million taxable capital gain TOM realized on selling the RRSP FMO units to the appellants). TOM paid its 2005 income to its unitholders by distributing its 9.4 million FIF units with a value of $140.5 million and paying cash for the balance: Step 20. However, only RRSP Trust received FIF units; the other TOM unitholders received their distribution in cash.
[71] The resulting simplified structure may be illustrated as follows:
[72] As is clear, the public who owned 42 percent of FMO when the reorganization commenced, own 42 percent of FIF, consistent with the one-for-one exchange contemplated by the FMO reorganization. On the other hand, neither RRSP Trust, owner of the other 58% of FMO when the reorganization was proposed, nor the appellants, owner when the reorganization commenced, own the other 58 percent of FIF. Rather, RRSP Trust and the appellants collectively own that 58 percent. Moreover, the appellants’ net indebtedness to TOM is substantially equal to the value of their FIF units, and TOM is owned as to 99.5 percent by RRSP Trust. (The value of the FIF units in March 2006 when TOM purchased them from the appellants to distribute its income would have reflected the FIF trading value at that time, rather than on December 28, 2005.)
[73] In 2006, the appellants each paid more than one capital dividend, to their parent corporations. The appellants’ dividend payments resulted in deficits in an almost equal amount: Appellants’ Balance Sheets, Appeal Book at 12279, 12378, 12405. Two of them paid stock dividends, such that their assets remained unchanged: Resolutions declaring capital dividends, Appeal Book at 11816, 11819, 11822, 11844, 11847, 11850.
[74] The parent corporations in turn added the capital dividends received to their CDA and paid Mr. Grenon capital dividends. The parent corporations’ capital dividends were not paid as stock dividends. Consequently, to pay those dividends, the parent corporations would have distributed assets to Mr. Grenon, and the value of their assets would have been correspondingly reduced: Resolution declaring capital dividend, Appeal Book at 11884.
E. Summary of Key Tax Results Relevant to the Appeal
[75] It is perhaps useful here to summarize the four key tax results relevant to this appeal that the appellants assert arise from the transactions I have just described.
[76] First, the appellants included the $113 million taxable capital gain designated by FMO in income, but deducted the same amount as an allowable capital loss from the disposition of their FMO units when FMO redeemed them: see paragraphs 55-57 and 59 above. As a result, the appellants reported no net taxable capital gain, no resulting income, and no resulting tax liability.
[77] Second, the appellants added $113 million to their CDA, relying on paragraph (f) of the CDA definition. While the capital losses from their disposition of the FMO units resulted in a deficit in their CDA under paragraph (a) of the CDA definition, that deficit had no effect on the addition under paragraph (f).
[78] Third, the capital dividends paid by the appellants to their parent corporations resulted in additions to the parent corporations’ CDA. As a result, the parent corporations could pay $110 million of non-taxable capital dividends to Mr. Grenon.
[79] Fourth, substantially all of FVT’s $137 million of income (from the sale of the operating businesses and their 2005 business operations) was paid to RRSP Trust, a tax-exempt taxpayer. As a result none of that income was subject to tax. (Whether RRSP Trust is taxable on that income is addressed in Grenon v. His Majesty the Queen, 2021 TCC 30 (RRSP TCC), described in paragraph 84 below, an appeal of which decision to this Court remains pending.)
[80] To summarize, no part of FVT’s 2005 income (including from 2005 business operations while the public owned 42 percent of FMO), nor any portion of the capital gains FMO realized when it sold FVT to TOM, were paid to the public FMO unitholders. And, despite the significant income realized in the course of the FMO reorganization ($105 million on sale of operating business assets and $113 million in taxable capital gains) and $110 million of dividends being paid to Mr. Grenon, no tax was payable by the appellants, FMO, FVT, TOM, RRSP Trust or Mr. Grenon. Only the public FMO unitholders and public TOM unitholders had a potential tax liability—the former on the taxable exchange of their public FMO units for FIF units, and the latter on TOM’s 2005 income distribution.
F. The Resulting Assessments
[81] The Minister of National Revenue disagreed with these results. Accordingly, in 2013, the Minister reassessed the appellants under Part I of the Income Tax Act. Relying on the general anti-avoidance rule (GAAR), the Minister eliminated the appellants’ capital gains and capital losses, and issued notifications no tax was payable, also known as nil assessments (the Part I assessments). The Part I assessments could not be appealed: Okalta Oils Ltd. v. Minister of National Revenue, [1955] S.C.R. 824, [1955] 5 D.L.R. 614 at 826; Canada v. Interior Savings Credit Union, 2007 FCA 151, [2007] 4 C.T.C. 55 at para. 17; and Canada v. 984274 Alberta Inc., 2020 FCA 125, [2020] 4 F.C.R. 384 at para. 59, leave to appeal to SCC refused 39355 (29 April 2021).
[82] In 2014, the Minister assessed the appellants for tax under Part III of the Income Tax Act. The assessments were premised on the appellants not having any capital gains from the FMO reorganization, and thus no CDA, with the result that all of the 2006 capital dividends they paid were excess dividends. In 2016, the Income Tax Act was amended to reduce the rate of Part III tax, and the Minister issued new assessments reflecting that lower rate (the Part III assessments).
[83] The appellants appealed the Part III assessments to the Tax Court.
[84] The Minister also assessed RRSP Trust and Mr. Grenon in connection with RRSP Trust’s acquisition of the TOM units, and the units of several other unit trusts established at Mr. Grenon’s initiative. Those assessments also were appealed to the Tax Court. All of the appeals were heard on common evidence, but the Tax Court issued separate judgments and reasons for judgment. RRSP TCC is the decision allowing Mr. Grenon’s appeal, but dismissing RRSP Trust’s appeal, of those assessments. A third decision awarded costs of all appeals to the respondent: Grenon v. The Queen, 2021 TCC 89.
[85] RRSP Trust has appealed RRSP TCC and all but the respondent have appealed the costs decision. While this Court heard the three appeals together, we too will issue separate judgments and reasons.
[86] Although RRSP TCC shares some common facts with this appeal, and the two decisions are related, the focus of this appeal is the Tax Court’s decision dismissing the appellants’ appeal of the Part III assessments.
[87] I turn to that decision now.
III. Tax Court Decision
A. Validity of the Part III Assessments
[88] Before the Tax Court, the appellants advanced two arguments attacking the validity of the Part III assessments, asserting that they should be vacated on that basis. First, they argued that each capital dividend election made by an appellant required a separate assessment under subsection 185(1). Here, a single assessment assessed Part III tax in respect of more than one capital dividend election made by each appellant. Secondly, the appellants argued that the assessments were invalid because they had not been issued with all due dispatch as subsection 185(1) requires.
[89] The Tax Court rejected both of these submissions. It concluded that the first was without merit, particularly as the capital dividend elections made by each appellant concerned capital dividends it declared and paid in a single taxation year. As to the second, the Tax Court concluded that any failure to assess the appellants with all due dispatch did not permit it to vacate or invalidate the assessments, citing Carter v. The Queen, 2001 FCA 275, [2001] 4 C.T.C. 79 [Carter] and Ginsberg v. Canada, [1996] 3 F.C. 334, 50 D.T.C. 6372 (F.C.A.) [Ginsberg].
[90] Having dismissed the appellants’ challenges to the validity of the Part III assessments, the focus turned to the correctness of the Part I assessments. Although the appellants could not appeal the Part I assessments, the Part III assessments depended on the correctness of those Part I assessments because they eliminated the appellants’ capital gains that led to the CDA addition.
B. Correctness of the Part I Assessments
[91] The respondent advanced three alternative arguments in support of the Part I assessments. First, the respondent asserted that the series of transactions that resulted in the capital gains and capital losses were not legally effective because, despite RRSP Trust’s transfer of the RRSP FMO units to TOM, followed by TOM’s transfer of those units to the appellants, beneficial ownership of those units never changed.
[92] Alternatively, the respondent said that the series of transactions that gave rise to the capital gains and capital losses were a sham and a misrepresentation.
[93] In the further alternative, the respondent said that the GAAR applied. The respondent asserted that the transactions were abusive of the capital gain and capital loss provisions and the capital dividend provisions of the Income Tax Act.
[94] The appellants challenged each of these grounds. They claimed the transactions were legally effective, beneficial ownership of the FMO units changed, none of the transactions was a sham, and that the conditions for the application of GAAR were not met.
[95] Again, the Tax Court disagreed with the appellants. It concluded that the appellants never acquired beneficial ownership of the RRSP FMO units because they acquired and held legal title as agents. Alternatively, it found that the transactions giving rise to the capital gains and capital losses were a sham. Finally, it said that, if it was wrong on the first two grounds, GAAR would apply and the elimination of the capital gains and capital losses by the Part I assessments, and the imposition of Part III tax by the Part III assessments, were reasonable tax consequences to deny the tax benefits.
[96] The appellants now appeal each of those findings.
IV. Standard of Review
[97] It is indisputable that the appellate standard of review applies to this appeal. Any question of fact or mixed fact and law (excluding an extricable question of law) is reviewed for palpable and overriding error; questions of law are reviewed on the standard of correctness: Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235.
V. The Appeal
A. Validity of the Part III Assessments
[98] Before us, the appellants challenge the Tax Court’s conclusion that the assessments cannot be vacated on the basis they were not issued with all due dispatch. However, the Tax Court cited and properly considered itself bound to follow Ginsberg and Carter, two decisions of this Court that directly address that very issue.
[99] Absent “exceptional circumstances”
, decisions of a panel of this Court also bind future panels of this Court: Miller v. Canada (Attorney General) 2002 FCA 370, 220 D.L.R. (4th) 149 at para. 9; Feeney v. Canada, 2022 FCA 190, 2022 A.C.W.S. 5833 at para. 16; Chen v. Canada, 2023 FCA 146, 2023 A.C.W.S. 2685. Exceptional circumstances exist “where the decision was ‘manifestly wrong’, in the sense that the Court overlooked a relevant statutory provision, or case that ought to have been followed”
, “when the decision has been overtaken by subsequent Supreme Court jurisprudence”
, or “where there are compelling reasons to [depart from the previous decision] and correctness prevails over certainty”
: Tan v. Canada (Attorney General), 2018 FCA 186, [2019] 2 F.C.R. 648 at para. 31.
[100] No exceptional circumstances exist here. Therefore, the appellants’ submission that the Tax Court erred in refusing to vacate the Part III assessments because they were not issued with all due dispatch must fail.
[101] While the appellants’ memorandum of fact and law asserts that the Tax Court erred in concluding that a single assessment for more than one capital dividend is valid, the memorandum contains no submissions regarding that alleged error. Although the appellants did not say they had abandoned this issue, neither did they advance any arguments in support during their oral submissions.
[102] No error being identified by the appellants, I need not address that question further.
B. Correctness of the Part I assessments
[103] The correctness of the Part I assessments is at issue.
[104] The appellants submit that the Tax Court erred with respect to the three alternative grounds on which it determined that those assessments should be sustained and so dismissed their appeals. To succeed, the appellants must establish that is so; demonstrating an error on one or two grounds is not sufficient to overturn the Tax Court’s decision.
[105] Accordingly, to decide this appeal, I must answer the following questions:
Did the Tax Court err in concluding that certain transactions were legally ineffective such that the appellants never acquired beneficial ownership of the RRSP FMO units?
Did the Tax Court err in concluding that the transactions giving rise to the capital gains and capital losses were a sham?
Did the Tax Court err in concluding that GAAR applies and that the tax consequences of it applying as reflected in the Part I and Part III assessments were reasonable to deny the tax benefits?
[106] I have concluded that the answer to the first two of these questions is yes, the Tax Court erred. Nonetheless, I would dismiss the appeal.
[107] I agree with the Tax Court that GAAR applies. There were tax benefits and avoidance transactions, and the avoidance transactions abused the object, spirit and purpose of the capital gain and capital loss provisions and the capital dividend rules. Finally, I am satisfied that the resulting tax consequences are reasonable in the circumstances.
VI. Analysis
[108] Because GAAR is a provision of last resort, I must address each of the three grounds on which the Tax Court dismissed the appellants’ appeals. However, before I do, I wish to reiterate that I disagree with many obiter statements in the Tax Court’s reasons and my failure to address many of them is not an endorsement.
A. Underlying Concern with Beneficial Ownership and Sham Analysis
[109] As an introductory matter, I must note a significant concern underlying the Tax Court’s reasons leading it to conclude that the appellants did not acquire beneficial ownership of the RRSP FMO units and that certain transactions were a sham.
[110] The position advanced by the respondent, and accepted by the Tax Court, failed to consider or address the ramifications of those conclusions on other aspects of the FMO reorganization. The approach appears to have been that those ramifications were not a relevant consideration.
[111] I disagree.
[112] With respect, this is not a case in which two transactions in the series (RRSP Trust’s transfer of the RRSP FMO units to TOM and TOM’s transfer of those units to the appellants) can be considered in isolation and the analysis proceed as if RRSP Trust can “simply”
be substituted for the appellants. The appellants and TOM participated in other transactions in the FMO reorganization, and after it was completed. The most obvious example, but not the only one, is that the appellants acquired the public FMO units from FVT in the course of the reorganization.
[113] The Tax Court described the transaction by which the appellants acquired the public FMO units, the respondent’s assumption those units had no value once the public exchanged them for FIF units, and the appellants’ submissions to the contrary: reasons at paras. 46, 72, 188. But, nowhere does the Tax Court make any findings regarding these points; nowhere does it tell us what happened to those units under its view of the transactions.
[114] The Tax Court often glossed over or ignored details. For example, with respect to FMO’s transfer of FVT to TOM, it said that certain “transactions would have been inconsequential for tax purposes”
: reasons at para. 178. This is incorrect. The details are important.
[115] While necessary to identify this concern, which undermines the Tax Court’s conclusions on beneficial ownership and sham, it is not the focus of my analysis of the Tax Court’s reasons. My focus is on the role of the Trustee.
[116] With those introductory comments, I turn now to the first ground on which the Tax Court dismissed the appellants’ appeals.
B. Did the Tax Court Err in Concluding that Certain Transactions were Legally Ineffective such that the Appellants Never Acquired Beneficial Ownership of the RRSP FMO Units?
[117] The Tax Court found that despite RRSP Trust transferring the RRSP FMO units to TOM (at Step 1) and TOM in turn transferring those units to the appellants (at Step 2), beneficial ownership of those units never changed. As a result, the Tax Court concluded that the appellants did not realize capital gains or capital losses in the course of the series of transactions.
[118] The appellants say the Tax Court erred in coming to these conclusions. I agree.
[119] The Tax Court summarized the respondent’s position in paragraphs 156 to 164 of its reasons as follows.
[120] Disposition as defined in the Income Tax Act excludes “any transfer of the property as a consequence of which there is no change in the beneficial ownership”
so that a transfer of mere legal title is not a disposition: paragraph (e) of the definition of disposition in s. 248(1). While that exclusion does not apply to transfers between a trust and its beneficiaries, “an arrangement under which the trust can reasonably be considered to act as agent for all the beneficiaries under the trust with respect to all dealings with all of the trust’s property”
is deemed not to be a trust: definition of trust in s. 104(1). Therefore, a transfer to a bare trust is not a disposition. A bare trust exists “when the trust acts entirely as agent of the beneficiary”
, citing Fourney v. The Queen, 2011 TCC 520, [2012] D.T.C. 1019 at paragraph 23; De Mond v. The Queen, 99 DTC 893, [1999] 4 C.T.C 2007 at paragraphs 36-37. Finally, “the test to determine beneficial ownership is the point in time when a person possesses the three key attributes of ownership, namely, risk, use and possession”
quoting Smedley v. The Queen, 2003 DTC 501, [2003] 2 C.T.C. 2658 at para. 10 (Smedley).
[121] Having summarized the respondent’s position, the Tax Court proceeded to analyze whether beneficial ownership of the RRSP FMO units changed.
[122] With respect to TOM’s acquisition of the RRSP FMO units from RRSP Trust, the Tax Court found that “TOM was the legal and registered owner of the [RRSP] FMO units but [Mr.] Grenon remained as beneficial owner because the units were still part of the RRSP Trust”
: reasons at para. 167 (emphasis added). With respect to the appellants’ acquisition of the RRSP FMO units from TOM, the Tax Court similarly concluded that “[Mr.] Grenon retained beneficial ownership of the [RRSP] FMO units in his capacity as annuitant”
notwithstanding that “[l]egal title to the [RRSP] FMO units may have been transferred from TOM to the Appellants”
: reasons at para. 176 (emphasis added). These conclusions led the Tax Court to further conclude that FMO’s transfer of the FVT units to TOM was “a circuitous transaction within the RRSP Trust”
, and did not result in a disposition or change of beneficial ownership: reasons at para. 178.
[123] I cannot agree.
[124] I commence with the Tax Court’s findings regarding the RRSP FMO units.
(1) Mr. Grenon as beneficial owner of the RRSP FMO units
[125] The Tax Court’s focus on Mr. Grenon’s status as the annuitant (i.e., sole beneficiary), his role as a TOM trustee, and his ability to direct RRSP Trust’s investments, led it to conclude that TOM was the legal and registered owner but Mr. Grenon remained beneficial owner. The Tax Court said “[the Trustee] was the legal and registered owner of those [FMO] units as plan administrator and trustee and … [Mr.] Grenon was the beneficial owner [of the RRSP FMO units] as the annuitant”
: reasons at para. 165 (emphasis added).
[126] In other words, the Tax Court appears to have concluded that RRSP Trust was a bare trust holding legal (registered) title to the RRSP FMO units for their beneficial owner, Mr. Grenon.
[127] This is an extricable error of law.
[128] The part of the trust definition relied on by the respondent that deems an agency relationship not to be a trust does not apply to a registered retirement savings plan: s. 104(1). Therefore, RRSP Trust was not a bare trust; it was a “real”
trust and a separate taxpayer from Mr. Grenon. It was the beneficial owner of the RRSP FMO units, not Mr. Grenon.
[129] It is true that Mr. Grenon was “beneficially interested”
in RRSP Trust’s assets: s. 248(25). But the conclusion that there was no change in beneficial ownership of the RRSP FMO units when TOM acquired them from RRSP Trust cannot be founded on Mr. Grenon’s status as sole beneficiary of RRSP Trust with a right to instruct the Trustee to dispose of the RRSP FMO units to TOM.
[130] RRSP Trust is a trust. Where a trust transfers property, it is a disposition. This is true even if the transfer is to a beneficiary or another trust with identical beneficiaries: subparagraphs (e)(i) to (iii) of the definition of disposition in s. 248(1). The only exception is where legal title is transferred to a person that holds it as bare trustee and acts as agent for the owner with respect to all dealings with the property.
[131] Here the owner was RRSP Trust. Thus, in order to conclude beneficial ownership of the RRSP FMO units did not change, TOM and the appellants must have been agents, holding the legal title to those units on a bare trust basis, for RRSP Trust.
(2) TOM and the appellants as agents of RRSP Trust
[132] Although the Tax Court did not explicitly find that TOM acquired legal title to the RRSP FMO units as RRSP Trust’s agent, its reasons sometimes suggest it drew that conclusion: “TOM was the legal and registered owner of the [RRSP] FMO units”
but they “were still part of the RRSP Trust”
and “continued to be held by TOM for the RRSP Trust”
: reasons at paras. 167, 178.
[133] In contrast, the Tax Court left no doubt that it considered the appellants to be agents, although for whom is less clear. The Tax Court said that it “must conclude that the Appellants were mere agents or nominees for the RRSP Trust or [Mr.] Grenon as principal”
: reasons at para. 174 (emphasis added). The appellants, as holders of the RRSP FMO units, cannot have been agents and nominees for Mr. Grenon and RRSP Trust at the same time, unless RRSP Trust is a bare trust, and so Mr. Grenon’s agent. I have already explained why that cannot be the case.
[134] The Tax Court’s lack of clarity about who the appellants are agents for is not isolated to a single statement. In its reasons, the appellants are considered nominees or agents for Mr. Grenon in paragraphs 172, 174, 176, 235, and 258, and for RRSP Trust in paragraphs 174, 178 and 179. Despite this lack of clarity, I am prepared to proceed on the basis that the Tax Court concluded that TOM and the appellants, in turn, held legal title to the RRSP FMO units as agent or nominee for RRSP Trust, which at all times retained beneficial ownership.
[135] Having said that, the Tax Court failed to consider certain key facts that lead me to conclude the Tax Court erred in coming to that conclusion.
[136] To qualify for registration as a registered retirement savings plan for purposes of the Income Tax Act, a plan must satisfy certain conditions, including, for a plan like RRSP Trust, that the trustee is a corporation licensed or otherwise authorized to carry on the business of offering its services to the public as trustee: ss. 146(1) definition of retirement savings plan, 146(2). Because RRSP Trust qualified as a registered retirement savings plan, Mr. Grenon was not, nor could he be, a trustee of RRSP Trust.
[137] RRSP Trust was established and governed by a declaration of trust: Retirement Savings Plan Declaration of Trust, Appeal Book at 4446-49. The declaration expressly appointed Trustee as trustee and CIBC Wood Gundy (Agent) as the Trustee’s agent. While the declaration authorized the Trustee to delegate performance of certain of its “duties, obligations and discretion”
to the Agent, the Trustee retained ultimate responsibility for the administration of RRSP Trust: sections 1 and 7 of the declaration.
[138] Under that declaration, the Trustee had significant rights, duties and obligations, including investing and reinvesting the RRSP Trust assets as Mr. Grenon directed, but subject to the Trustee’s discretion to not make particular investments if they and related documentation did not comply with the Income Tax Act or the Trustee’s requirements. The Trustee had the right to modify its requirements from time to time. The Trustee’s obligations extended to “generally exercising all powers or rights of an owner with respect to all [investments which are part of the plan]…including…the right to vote or give proxies to vote”
: section 8 of the declaration.
[139] The declaration also governed Mr. Grenon’s rights as annuitant/beneficiary. He had the right to direct the RRSP Trust’s investments (subject to the Trustee’s discretion to refuse), to withdraw funds from RRSP Trust, and to appoint the Agent as his agent for purposes of giving investment instructions. However, he had no right to act on RRSP Trust’s behalf or to appoint an agent to act on its behalf.
[140] It is obvious that the Trustee and Agent played a critical role and had significant responsibility in any actions RRSP Trust undertook. Indeed, in RRSP TCC the Tax Court recognized this: RRSP TCC at paras. 42, 52, 89-99, 474, 475, 537. Yet, in deciding that the appellants were RRSP Trust’s agent, the Tax Court does not turn its mind to the role played by the Trustee or Agent, nor the terms of the declaration governing RRSP Trust. The Tax Court describes the Trustee as trustee and plan administrator, the registered and legal holder of the RRSP FMO units, and the person Mr. Grenon could instruct regarding RRSP Trust’s investments, but says nothing more about it. Obviously, the Trustee’s role with respect to RRSP Trust was far more significant than that.
[141] In my view, the Tax Court erred in failing to consider this evidence.
[142] To put it plainly, neither the appellants nor TOM (nor anyone else, including Mr. Grenon) could be agents for RRSP Trust unless the Trustee or Agent — the only ones authorized to act on RRSP Trust’s behalf — appointed them as such. Yet there is no suggestion, nor any finding, that they did. Nothing in the record supports such a finding.
[143] The Tax Court made no findings the Trustee or Agent had knowledge concerning Mr. Grenon’s “CDA plan”
or the transactions in issue here. I see no suggestion that the respondent asserted otherwise. Rather, the respondent asserted, and the Tax Court accepted, that Mr. Grenon remained beneficial owner of the RRSP FMO units throughout, seemingly treating Mr. Grenon and RRSP Trust as if they were one and the same so that Mr. Grenon could direct TOM and the appellants to act as RRSP Trust’s agent.
[144] This is not so.
[145] Mr. Grenon could provide investment instructions, but those instructions were subject to the Trustee or Agent’s approval. As the Tax Court recognized in RRSP TCC, the Trustee required legal opinions that the TOM units were qualified investments for RRSP Trust before it agreed to RRSP Trust’s subscription. If RRSP Trust was not disposing of the RRSP FMO units when it subscribed for TOM units, and instead continued to hold the RRSP FMO units through its agent TOM, why would opinions be necessary?
[146] As a final point, the Tax Court’s conclusion that TOM was agent for RRSP Trust is entirely inconsistent with its conclusion that RRSP Trust was liable for tax based on its cost of the TOM units being $153 million: RRSP TCC at paras. 629, 637. RRSP Trust acquired those units in exchange for the RRSP FMO units valued at $153 million. If TOM acquired only legal title to the RRSP FMO units, RRSP Trust’s cost of the TOM units would not be $153 million. It would be nominal, limited to the value of the legal title to the RRSP FMO units.
[147] In conclusion, the Tax Court’s findings, the respondent’s assumptions, the evidence in the record, and the Tax Court’s analysis, cannot sustain a conclusion that TOM or the appellants were agents for RRSP Trust so that beneficial ownership of those units did not change.
(3) The attributes of ownership
[148] I acknowledge that the Tax Court also said that “it cannot be said that the Appellants enjoyed ‘the three key attributes of ownership, namely, risk, use and possession’”
: reasons at para. 175, quoting Smedley. In saying this, the Tax Court points out that the appellants had no discretion regarding disposal of the RRSP FMO units once they acquired them, it being understood that the appellants would not take any other steps, and that all of the transactions were pre-ordained.
[149] In the context of a reorganization like the FMO reorganization, steps are invariably pre-ordained once the reorganization is approved and proceeds. Thus, those statements without more cannot sustain the Tax Court’s conclusion that the appellants did not acquire beneficial ownership of the RRSP FMO units.
(4) Beneficial ownership of the FVT units
[150] As noted in paragraph 122 above, the Tax Court also found that beneficial ownership of the FVT units did not change when TOM acquired them from FMO. That conclusion was premised on the Tax Court’s finding that TOM and the appellants were agents who did not acquire beneficial ownership of the RRSP FMO units. My conclusion that the Tax Court erred in those findings is sufficient to conclude this too is an error. I need say no more about the Tax Court’s analysis on this finding.
(5) Conclusion on beneficial ownership
[151] To summarize, I agree with the appellants that the Tax Court’s conclusions that the appellants did not acquire beneficial ownership of the RRSP FMO units, and that TOM did not acquire beneficial ownership of the FVT units, cannot be sustained.
C. Did the Tax Court Err in Determining that the Transactions that Gave Rise to the Capital Gains and Capital Losses were a Sham?
[152] I turn now to the Tax Court’s alternative conclusion for sustaining the Part I assessments, that certain transactions were shams. In particular, the Tax Court concluded each of TOM’s transfer of the RRSP FMO units to the appellants, FMO’s transfer of the FVT units to TOM, and FMO’s purchase for cancellation of the FMO units from the appellants, was a sham: reasons at paras. 221, 223, 226.
[153] The first of these transfers resulted in the appellants acquiring the RRSP FMO units from TOM. The second resulted in the disputed capital gain, and the third the disputed capital loss.
[154] Notably, the Tax Court did not find that RRSP Trust’s transfer of the RRSP FMO units to TOM was a sham. Rather, it said, that “TOM continued to hold beneficial ownership of those units for the benefit of the RRSP Trust and ultimately for the benefit of [Mr.] Grenon, as the annuitant thereof”
: reasons at para. 221. In other words, the Tax Court concluded TOM was RRSP Trust’s agent. I have already explained why I cannot agree with that conclusion.
[155] Turning to the sham findings, the Tax Court relied on its analysis from the RRSP TCC: see RRSP TCC at paras. 364-373.
[156] As the Tax Court there described, a sham exists when acts are done or documents are executed with the intention of giving the appearance of creating legal rights and obligations that differ from the actual legal rights and obligations that the participants intend to create. A sham involves an element of deceit in that the participants know that their actual legal rights and obligations differ from those presented to others. The necessary “element of deceit…generally manifests itself by a misrepresentation by the parties of the actual transaction taking place between them”
: 2529-1915 Québec Inc. v. Canada, 2008 FCA 398, [2009] 3 C.T.C. 77 at para. 59 [Faraggi].
[157] As this Court has explained, the concepts of sham and abuse are not the same: Faraggi at paras. 54-55. The Tax Court recognized this, stating that neither a tax motivation nor taking steps to implement a “tax plan”
by itself constitutes a sham, citing Cameco Corporation v. The Queen, 2018 TCC 195, [2019] 1 C.T.C. 2001 (aff’d 2020 FCA 112, [2020] 4 F.C.R. 104, leave to appeal to SCC refused, 39368 (18 February 2021) at paragraph 605: reasons at para. 220.
[158] Despite this, reading the Tax Court’s sham analysis, the only conclusion I can draw is that its finding of sham was founded on the tax motivation, i.e., the purpose of the transactions: see reasons at paras. 213, 214, 218, 219, 220. Thus, the Tax Court erred.
[159] The Tax Court concluded that the “sole purpose”
of RRSP Trust transferring the RRSP FMO units to TOM was “implementing the series of transactions that would lead to the creation of the subject capital gains and capital losses and the payment of the alleged capital dividends”
: reasons at para. 214. Then, after saying that RRSP Trust had a $125 million accrued gain on the RRSP FMO units before they were transferred to TOM, the Tax Court described a second “sole purpose”
of the transfer of those units to TOM as “extract[ing] the gain…that had accrued in the RRSP Trust”
: reasons at para. 219.
[160] From this, the Tax Court concluded that the transactions involving the appellants were undertaken to create the illusion that Mr. Grenon had extracted this gain, although the “[RRSP] FMO units had not been withdrawn from the RRSP Trust”
, “TOM continued to hold beneficial ownership of those units for the benefit of RRSP Trust”
and “[Mr.] Grenon, and by extension the Appellants, knew or must be taken to have known [this]”
: reasons at paras. 220, 221.
[161] It is clear that when the Tax Court referred to a withdrawal from RRSP Trust, it meant a withdrawal by the annuitant—in this case Mr. Grenon—which, as the Tax Court properly observed, results in the withdrawn amount being added to the annuitant’s income: reasons at paras. 23, 24, 166, 170, 180, 215, 240, 242, 243, 258. Thus, the Tax Court can only have meant that Mr. Grenon wanted to be seen as having made a withdrawal from RRSP Trust (i.e., creating an illusion he had done so) when in fact he did not.
[162] However, the Tax Court did not explain why Mr. Grenon wanted to create that illusion and how it is consistent with the first purpose the Tax Court identifies—creating capital gains and capital losses and paying capital dividends. As to the second purpose, extracting the gain in RRSP Trust, how is that purpose achieved by creating an illusion it was achieved?
[163] In my view, the Tax Court’s focus on the accrued gain on the RRSP FMO units is misplaced. RRSP Trust’s exchange of one asset (the RRSP FMO units) for another (initially TOM units, and ultimately TOM units and FIF units) did not result in the extraction of any gain from RRSP Trust. The value of RRSP Trust’s assets was not reduced by $125 million and the respondent does not appear to have suggested otherwise. The Tax Court itself recognized that the exchange of RRSP FMO units for TOM units had no effect of the value of RRSP Trust’s assets: RRSP TCC at paras. 449-50, 614, 626.
[164] Before the series of transactions commenced, the RRSP FMO units had a value of $151 million. At the end of the series of transactions RRSP Trust owned FIF units with a value of $140.5 million as well as its interest in TOM, with a value of more than $20 million: see Step 20. Where is the extracted gain?
[165] The Tax Court’s sham analysis suffers from other problems.
[166] The Tax Court described FMO’s transfer of the FVT units to TOM, and FMO’s repurchase of its units for cancellation, as “mere paper transactions as described in Faraggi”
: reasons at para. 223. With respect, the Tax Court appears to have misunderstood that case.
[167] In Faraggi, this Court specifically criticized the Tax Court’s conclusion (in the decision under appeal there) that transactions were mere paper transactions and thus shams: Faraggi at paras. 69-70. This Court concluded that the capital dividend elections, not the transactions, were the shams: “[i]n making these [capital dividend] elections, the subsidiaries misrepresented to the Minister and to all those affected by these elections that the disposition…had resulted in capital gains”
; “[d]espite the impression given, no capital gains were made”
: Faraggi at paras. 77-78.
[168] In Faraggi, third parties would have had to have been participants in the sham the Tax Court identified. This Court expressed skepticism that that was the case. I am of the same view here. Finding TOM was RRSP Trust’s agent equates with finding the Trustee or Agent was a participant in the sham. The respondent did not allege that was the case, and the Tax Court’s findings on knowledge about the “deceit”
do not extend to the Trustee or Agent.
[169] I agree with the Tax Court that the appellants enjoyed no economic gain or loss from their acquisition and subsequent disposition of the FMO units. However, that does not lead to a conclusion that the transactions were shams. Here the capital gain arose on the disposition of the FVT units. There was no allegation that those units were not capital property. There can be no dispute that before the series of transactions began, FMO had an accrued gain on the FVT units attributable to the gain on the underlying operating business assets.
[170] Finally, although the Tax Court concluded that the transactions that purported to create the capital gains and capital losses were shams that should be disregarded, it does not explain what the “real”
transactions were, or the consequences flowing from them. It merely repeats its assertion that FMO’s transfer of the FVT units to TOM is “a circuitous transaction within the RRSP Trust”
that “cannot be said [to have] triggered ‘real’ capital gains”
: reasons at para. 223.
[171] My comments in paragraphs 109 to 115 above apply here. When confronted with a sham, courts are to “consider the real transactions and disregard the one that was represented as being the real one”
: Faraggi at para. 59. Here, the Tax Court disregarded transactions, but did not consider—or even identify—the real ones.
[172] In my view, the Tax Court’s conclusion that the appellants’ acquisition of the RRSP FMO units, TOM’s acquisition of the FVT units from FMO, and FMO’s redemption of the FMO units from the appellants, were shams cannot stand.
[173] Having determined that the Tax Court’s conclusions on beneficial ownership and sham are not sustainable, I turn now to application of GAAR.
D. Does GAAR Apply?
[174] I am satisfied GAAR applies.
[175] To sustain the assessments based on GAAR, three questions must be answered affirmatively:
Is there a tax benefit?
If so, is one or more of the transactions giving rise to the tax benefit an avoidance transaction?
If so, is the tax avoidance abusive?
Canada Trustco Mortgage Co. v. Canada, 2005 SCC 54, [2005] 2 S.C.R. 601 at para. 66 [Canada Trustco]; Copthorne Holdings Ltd. v. Canada, 2011 SCC 63, [2011] 3 S.C.R. 721 at para. 33 [Copthorne]; Deans Knight Income Corp. v. Canada, 2023 SCC 16, [2023] 4 C.T.C. 25 at para. 51 [Deans Knight].
[176] While “tax benefit”
and “avoidance transaction”
are defined in the Income Tax Act, to determine whether the avoidance is abusive, a court must identify the rationale—the object, spirit and purpose—of the provisions relied on to obtain the tax benefit. Frustration of that object, spirit and purpose leads to a finding of abuse.
[177] There is no dispute that the Tax Court addressed all three questions in concluding that GAAR applied. However, the appellants say the Tax Court erred and that GAAR does not apply.
[178] I disagree. There were tax benefits, the series of transactions included several avoidance transactions, and the tax avoidance was abusive.
[179] I turn now to address each of these elements in turn. For this purpose, I have accepted that all of the transactions were legally effective and occurred largely as described in Annex C to the Tax Court’s reasons.
(1) Was there a tax benefit?
[180] The Tax Court identified more than one tax benefit from the series of transactions. The first was in connection with RRSP Trust: reasons at para. 239. However, the Tax Court described as a second tax benefit the appellants being “able to record capital gains and off-setting capital losses in more or less equal amounts, as a result of transactions that occurred within minutes of each other on the same day and then [purporting] to make additions to their capital dividend accounts and [declaring] tax-free capital dividends of about $110,000,000”
: reasons at para. 248.
[181] Before us, the appellants say the Tax Court erred in its tax benefit analysis in two respects.
[182] First, they say, to sustain GAAR assessments against the appellants, the appellants must have had the tax benefit, but they had no benefit here. Neither the addition to the appellants’ CDA nor an avoidance of Part III tax qualifies as a tax benefit: 1245989 Alberta Ltd. v. Canada (Attorney General), 2018 FCA 114, [2019] 3 C.T.C. 1 at paras. 31-32 [Perry Wild]; Gladwin Realty Corporation v. Canada, 2020 FCA 142, [2020] 6 C.T.C. 185 at para. 47 [Gladwin]. The appellants’ capital losses were neither relevant to the Part III assessment, nor identified as a tax benefit by the Minister in connection with the Part III assessment. And, they say, Mr. Grenon’s capital dividends, and any tax benefit related to RRSP Trust, were not the appellants’ tax benefits.
[183] Second, the appellants say, a tax benefit is to be identified by comparison to an alternative arrangement that satisfies certain criteria. Here the Tax Court erred in failing to identify an appropriate alternative.
[184] In my view, both of these submissions must fail.
(a) The appellants had a tax benefit
[185] Before the Tax Court, the respondent asserted the appellants’ avoidance of Part III tax was a tax benefit from the series of transactions: see Corporate Appeals Fresh as Amended Replies Amended Replies [sic], Appeal Book at 1988-2059 (Amended Replies) at para. 26. The Tax Court did not consider this tax benefit separately. Rather it described a number of elements as collectively constituting a tax benefit without reference to the definition of tax benefit. This was an error.
[186] The Supreme Court teaches that where a series of transactions gives rise to more than one tax benefit, the legal analysis must focus on the individual tax benefits: Lipson v. Canada, 2009 SCC 1, [2009] 1 S.C.R. 3 at para. 22. The series of transactions at issue here undoubtedly involved many tax benefits, several of which may have been perfectly acceptable. GAAR’s focus is denial of one or more particular tax benefits.
[187] Although the existence of a tax benefit is a question of fact, given the Tax Court’s error, as in Lipson, this Court must identify the tax benefit.
[188] The definition of “tax benefit”
includes “a reduction, avoidance or deferral of tax”
: s. 245(1). As the respondent points out, nothing precludes a corporation from paying a capital dividend despite the corporation’s CDA balance being nil. The consequence of doing so is liability for Part III tax. Conversely, paying a capital dividend while avoiding that tax is a tax benefit. There is no value judgment at the tax benefit stage—the only question is whether there was a reduction, avoidance or deferral of tax or other amount payable, or an increase in a refund of tax or other amount, under the Income Tax Act. Undoubtedly avoiding or reducing Part III tax is a perfectly acceptable tax benefit in many circumstances, but it nonetheless is a tax benefit.
[189] Citing Gladwin, the appellants submit that Part III tax could not become payable until GAAR applied to reduce their CDA, and accordingly avoidance of Part III tax is not a tax benefit. I disagree. Whether there is a tax benefit is a question of fact that must be decided in each case based on the particular facts and circumstances.
[190] Gladwin concerned a notice of determination of CDA based on GAAR, not an assessment of tax. There, the parties agreed there was a tax benefit but did not specify what it was: Gladwin Realty Corporation v. The Queen, 2019 TCC 62, [2020] 2 C.T.C. 2184 at para. 8, and Appendix A, Partial Agreed Statement of Facts [Gladwin TCC]. The Tax Court surmised that if the appeal of the CDA determination failed, the Minister would assess Part III tax unless the taxpayer elected to treat the excessive capital dividend as a taxable dividend, and so avoidance of Part III tax was the tax benefit: Gladwin TCC at para. 6. But that was not a tax benefit the parties identified.
[191] On appeal, the parties disagreed on the tax benefit, the taxpayer claiming it was the addition to CDA (not a tax benefit following Perry Wild), and the Crown the avoidance of Part III tax. While the tax consequence imposed under GAAR to deny the tax benefit was a reduction of the taxpayer’s CDA by a notice of determination, nothing in the record suggests any tax had been assessed against anyone. Whether the Minister intended to assess Part III tax or to assess the dividend recipients as having received taxable dividends, or sought to deny some other tax benefit, was unknown. Similarly, whether the taxpayer would avoid any Part III tax liability by making the taxable dividend election immediately after the appeal is unknown. In those circumstances, a notice of determination that reduced CDA might be viewed as having brought about the means to avoid Part III tax. But that factual finding in that case does not govern here.
[192] The tax benefit the Minister seeks to deny by the assessments at issue here is clear—the appellants’ avoidance of Part III tax. To deny that tax benefit, the Minister issued assessments, not notices of determination of the appellants’ CDA. The Part I assessments reduced the appellants’ capital gains and capital losses to nil, thus eliminating their CDA, and the Minister assessed the appellants for the Part III tax avoided.
[193] It is true that Part III tax was not payable by the appellants until GAAR applied. However, where the tax benefit at issue is the avoidance of tax, it is no surprise that the consequence of GAAR applying is liability for the tax avoided. That does not mean the avoidance of that tax was not a tax benefit.
[194] Many cases illustrate this point. In Deans Knight the tax benefit was the reduction of Part I tax by deducting losses. Until GAAR applied, the losses were deductible and the reduced Part I tax was not payable. GAAR applied to deny that tax benefit by denying the losses so the Part I tax became payable. In Lipson, the deduction of interest resulted in a reduction of Part I tax and the reduced Part I tax was not payable until GAAR was applied.
[195] Copthorne might be viewed as more analogous to this case. There the tax benefit was the avoidance of withholding tax; that tax did not become payable until GAAR applied to reduce the paid-up capital of the Canadian corporation. By reducing the paid-up capital and assessing the Canadian payer corporation for withholding tax, the tax benefit—the avoidance of withholding tax—was denied. Here the tax benefit is avoidance of Part III tax. By applying GAAR to eliminate the appellants’ capital gains and capital losses, and thus their CDA, and assessing the Part III tax, the appellants’ tax benefit was denied.
[196] That is not to say that there were not other tax benefits from the series of transactions at issue in this appeal. There were. For example, the capital dividends Mr. Grenon received were a tax benefit. The Minister might have assessed Mr. Grenon on the basis his capital dividends were a tax benefit that should be denied. However, nothing in the record suggests the Minister did so and I am not persuaded that the Minister was obligated to assess Mr. Grenon to deny his tax benefit, rather than the appellants to deny theirs. GAAR does not dictate which tax benefit or tax benefits from avoidance transactions must be denied.
[197] The assessments at issue in this appeal address a particular tax benefit—avoidance of Part III tax. That tax benefit is the appellants’ tax benefit.
(b) Did the Tax Court err with regard to the alternative transaction analysis?
[198] The appellants submit the Tax Court erred in not identifying an alternative transaction that might reasonably have been carried out but for the existence of the tax benefit. Moreover, say the appellants, for the purposes of a comparative analysis, it is unreasonable to construct an alternative that “involves a conscious decision to declare an excess capital dividend”
: Appellants’ Memorandum of Fact and Law at para. 72.
[199] There is no dispute that alternative transactions may play a role in a GAAR analysis. They may assist in identifying a tax benefit or an avoidance transaction: Canada Trustco at para. 20; Copthorne at para. 35. They may help to determine whether the object, spirit and purpose of the relevant provisions have been abused: see, for example, Lehigh Cement Limited v. Canada, 2010 FCA 124, [2011] 4 F.C.R. 66, leave to appeal to SCC refused, 33794 (4 November 2010) at paras. 40-41; Birchcliff Energy Ltd. v. Canada, 2019 FCA 151, [2020] 1 C.T.C. 1 at paras. 31, 48, leave to appeal to SCC refused, 38761 (14 November 2019); Univar Holdco Canada ULC v. Canada, 2017 FCA 207, [2019] 2 F.C.R. 569 at paras. 19-20; 3295940 Canada Inc. v. Canada, 2024 FCA 42, [2024] 3 C.T.C. 93 at para. 58.
[200] While “in some instances, it may be that the existence of a tax benefit can only be established by comparison with an alternative arrangement”
, in other circumstances “the existence of a tax benefit is clear”
: Canada Trustco at para. 20; Fiducie financière Satoma v. Canada, 2018 FCA 74, [2019] 2 C.T.C. 33 at para. 45, leave to appeal to SCC refused, 38146 (28 March 2019); Copthorne at para. 38.
[201] Therefore, it is not always necessary to identify an alternative transaction to establish the existence of a tax benefit—here the appellants’ CDA was increased, they paid capital dividends and avoided Part III tax. Avoidance of the tax is clearly within the definition of tax benefit.
[202] That said, as described at paragraphs 210 to 212 below, I am satisfied that in an appropriate alternative series of transactions, the appellants would have no role. It is clear that the only purpose for the appellants’ participation in the series of transactions was to pay capital dividends to their parent corporations and to avoid Part III tax in doing so. But for that tax benefit, the appellants would not have paid any dividends, capital or otherwise. A tax benefit finding cannot be challenged on the ground “had I known GAAR would apply, I would not have taken advantage of the tax benefit”
.
(2) Was there an avoidance transaction?
[203] There were avoidance transactions.
[204] The appellants do not take issue with the Tax Court’s finding that “the transfer of the FMO units from the RRSP Trust to TOM was undertaken for the sole purpose of implementing the series of transactions that would lead to the creation of the subject capital gains and capital losses and the payment of the alleged capital dividends”
: reasons at para. 214. Rather, they submit that because there is no tax benefit, there is no avoidance transaction.
[205] But, there are tax benefits and I see no error in the Tax Court’s conclusion that TOM’s acquisition of the RRSP Trust FMO units is an avoidance transaction. It is not the only one.
[206] If a transaction results in a tax benefit, or is part of a series of transactions that would result in a tax benefit, the transaction is an avoidance transaction unless it may reasonably be considered to have been undertaken primarily for bona fide purposes other than to obtain the tax benefit: s. 245(3).
[207] However, “a bona fide non-tax purpose for a series of transactions does not exclude the possibility that the primary purpose of one or more transactions within the series is to obtain a tax benefit”
: Canada v. MacKay, 2008 FCA 105, [2008] 4 F.C.R. 616 at para. 25, leave to appeal to SCC refused, 32616 (15 January 2009) (emphasis in original). Consequently, while the FMO reorganization may have had bona fide purposes, about which I express no view, that does not preclude a finding that there is an avoidance transaction.
[208] The stated objectives of the FMO reorganization were threefold: increasing the tax cost of the operating businesses’ assets, simplifying the organizational structure, and making units more attractive to retail investors: reasons at para. 35.
[209] Before the Tax Court, the respondent submitted that the FMO reorganization could have been achieved in fewer steps and without the involvement of the appellants. The Tax Court did not address alternatives in its GAAR analysis. While it referenced the respondent’s submissions in its sham analysis, it found it unnecessary to review those submissions choosing to “highlight the submissions made on the subject capital gains and the effect of the capital losses on the calculation of the capital dividend account”
: reasons at para. 196.
[210] The respondent’s pleading in the Tax Court described the transactions that the respondent considered unnecessary to complete the FMO reorganization: see Amended Replies, Appeal Book at 1988-2059. In particular, the respondent asserted that the following transactions were unnecessary to effect the FMO reorganization:
i. Step 1: RRSP Trust’s transfer of the RRSP FMO units to TOM.
ii. Step 2: TOM’s sale of the RRSP FMO units to the appellants.
iii. Step 11: the appellants’ purchase of the public FMO units from FVT (FULP).
iv. Step 14: FMO’s sale of FVT to TOM.
v. Step 15: FMO’s distribution of FIF units and promissory notes to the appellants to effect payment of its $226 million capital gain.
vi. Step 16: FMO’s repurchase of all but 100 of its units from the appellants.
vii. The appellants’ payment of capital dividends to their parent corporations and those corporations’ payment of capital dividends to Mr. Grenon.
I agree that each of these “transactions [was] undertaken or arranged primarily to provide tax benefits to [the appellants] and [Mr.] Grenon”
: Amended Replies, at para. 15, Appeal Book at 2007-2009. I agree with the respondent that each is an avoidance transaction.
[211] Having said that, I am not suggesting that simply eliminating those transactions from the series would by itself achieve the stated objectives of the FMO reorganization. However, I am satisfied that there is at least one “alternative arrangement … that ‘might reasonably have been carried out but for the existence of the tax benefit’”
: Copthorne at para. 35 [citations omitted]. Describing an alternative arrangement allows a court to “isolate the effect of the tax benefit from the non-tax purpose of the taxpayer”
: Copthorne at para. 35. This, in turn, assists in identifying the avoidance transactions.
[212] Appendix A to these reasons describes one possible alternative series of transactions that eliminates or varies the transactions described in paragraph 210 above, while achieving the FMO reorganization’s stated objectives. That alternative eliminates the appellants’ participation in the FMO reorganization entirely. Thus, the alternative isolates the non-tax purpose—the FMO reorganization’s objectives, from the tax benefit at issue here—avoidance of Part III tax on capital dividends the appellants paid to their parent corporations. That particular tax benefit resulted, directly or indirectly, from one or more avoidance transactions in the series.
[213] I am satisfied that the Tax Court and the respondent correctly identified avoidance transactions.
[214] I turn now to the final element of GAAR—abuse.
(3) Was there an abuse?
[215] The avoidance transactions that resulted in the tax benefits were abusive.
[216] Determining whether the tax avoidance—obtaining the tax benefits—was abusive involves two steps. First, a court must identify the object, spirit and purpose of the relevant provisions. Once it has done so, it must decide whether the result of the transactions frustrates that object, spirit and purpose: Deans Knight at para. 56, citing Canada Trustco at para. 44 and Copthorne at paras. 69-71.
[217] To identify the object, spirit and purpose of the provisions alleged to be abused, courts must have “reference to the provisions themselves, the scheme of the [Income Tax Act] and permissible extrinsic aids”
: Deans Knight at para. 58, citing Canada Trustco at para. 55.
[218] Before the Tax Court, the respondent alleged that the avoidance transactions resulted directly or indirectly in an abuse of paragraphs 38(b), 39(1)(b) and 40(1)(b) of the Income Tax Act—concerned with capital gains and capital losses—and subsections 83(2) and 89(1) – concerned with capital dividends.
[219] The Tax Court said that the abuse analysis “relate[d] in part to the capital gains regime”
as it had described earlier in its reasons: reasons at para. 253. By this I understand the Tax Court to mean its statement that the appellants did not enjoy “a ‘real’ economic gain nor a real economic loss”
, quoting from Triad Gestco Ltd. v. The Queen, 2011 TCC 259, aff’d 2012 FCA 258, [2014] 2 F.C.R. 199 [Triad Gestco FCA]: reasons at para. 224.
[220] As to the capital dividend provisions, the Tax Court concluded that “the ‘object, spirit and purpose’ of the CDA regime is to ensure that it mirrors the tax treatment of capital gains for an individual and that the Minister can only seek to tax gains that give rise to ‘real’ economic gains”
, but “[b]y the same token, only one half of the ‘real’ economic gains realized by a corporation can be added to the CDA”
: reasons at para. 255.
[221] While the appellants submit that the Tax Court erred in its abuse analysis, they address only the Tax Court’s comments concerning the CDA definition, and not the abuse of the capital gains and capital loss provisions. They assert that “[t]here is no reason to believe that Parliament did not intend a taxpayer to be able to rely on the calculation of the capital dividend account exactly as it read”
: Appellants’ Memorandum of Fact and Law at para. 79.
[222] That, of course, only goes so far. Where “a taxpayer does not satisfy the statutory requirements of a provision on which [it] relies, the Minister need not resort to GAAR”
: Canada v. Imperial Oil Ltd., 2004 FCA 36, [2004] 2 C.T.C. 190 at para. 30; see also Copthorne at para. 66; Deans Knight at para. 62. And, “there is no bar to applying the GAAR in situations where the [Income Tax Act] specifies precise conditions that must be met to achieve a particular result”
: Deans Knight at para. 71.
[223] The Tax Court’s discussion of the object, spirit and purpose of the capital gain and capital loss and the CDA provisions is undeniably short. And, it must be said, I disagree with a number of statements that the Tax Court made about the definition of CDA as it read in 2006.
[224] That said, this Court has already determined the object, spirit and purpose of the provisions said to be abused. The Tax Court was not obliged to start afresh. In light of that jurisprudence, I am satisfied that the series of transactions frustrated the object, spirit and purpose of the capital gain and capital loss provisions, as well as the CDA provisions.
[225] The abuse of the capital gains and capital loss provisions is what enabled the appellants to achieve their tax benefit. As I will explain later in these reasons, abusing the CDA provisions enabled Mr. Grenon to achieve his tax benefit— the receipt of capital dividends on which he paid no tax. However, as I have already stated, Mr. Grenon’s tax benefit is not the tax benefit denied by the appellants’ assessments.
[226] I start with the abuse of the capital gain and capital loss provisions.
(a) Abuse of the object, spirit and purpose of the capital gain and capital loss provisions
[227] Triad Gestco FCA involved a series of transactions undertaken to create a loss to avoid tax on a substantial capital gain. There, this Court observed that “the capital gain system is generally
understood to apply to real gains and real losses”
: Triad Gestco FCA at para. 41. This Court cited as “entirely apposite”
the following passage from the House of Lords decision in W.T. Ramsay Ltd. v. Inland Revenue Commissioners, [1981] UKHL 1 (BAILII), [1981] 1 All E.R. 865 at 873:
The capital gains tax was created to operate in the real world, not that of make-believe. As I said in Aberdeen Construction Ltd. v. Inland Revenue Comrs, [1978] 1 All ER 962 at 996, [1978] AC 885 at 893, [1978] STC 127 at 131, it is a tax on gains (or, I might have added, gains less losses), it is not a tax on arithmetical differences.
[228] As the Court further observed in Triad Gestco FCA, “the capital gain system has been understood, since a time that pre-dates its creation, to be aimed at taxing increases in ‘economic power’”
: para. 42 [citation omitted].
[229] The same must be said of a capital loss—it must be understood as being aimed at providing relief where there is a decrease in economic power. It is not a relief from tax based on an arithmetic difference.
[230] This Court has said that “[i]n Triad Gestco, this Court distinguished a ‘paper loss’ from an ‘economic’ or ‘true’ loss and held that, given the object, spirit and purpose of paragraphs 38(b), 39(1)(b)
and 40(1)(b), a paper loss does not give rise to an allowable capital loss”
: 2763478 Canada Inc. v. Canada, 2018 FCA 209, 2018 D.T.C. 5130 at para. 53 [2763478]. It there explained:
Sections 39 and 40 provide the method for calculating the gain or loss. A loss is incurred when property is disposed of for “proceeds of disposition” that are lower than its “adjusted cost base”. The “adjusted cost base” is the purchase price of a capital property adjusted in accordance with section 53, and the “proceeds of disposition” is the price for which the property is sold or is otherwise compensated for, as provided in section 54. The difference between the adjusted cost base and the proceeds of disposition of a given property provides a measure of its change in value, and the corresponding increase or decrease in the owner’s economic power …
2763478 at para. 55, citing Triad Gestco at paras. 42 and 50 (emphasis added). While 2763478 concerned capital losses, the same is true of capital gains – the difference between the proceeds of disposition and the adjusted cost base of a given property provides a measure of its change in value and the corresponding increase in the owner’s economic power.
[231] These principles apply here.
[232] The appellants had neither an economic gain nor an economic loss; there was absolutely no change in their economic power as a result of their participation in the FMO reorganization. There was no change in the net value of their assets as a result. The appellants paid $277 million to purchase the FMO units ($161 million to TOM and $116 million to FVT). To do so, they incurred indebtedness of $277 million, issuing promissory notes to TOM and FVT.
[233] Immediately after acquiring the FMO units, the appellants exchanged those units for property with exactly the same value, albeit in two steps: FMO first distributed the appellants’ promissory notes ($161 million), TOM’s promissory note ($72 million) and 3 million FIF units ($44.5 million) to the appellants: Step 15. Immediately thereafter FMO redeemed the FMO units from the appellants for nominal consideration: Step 16. After these transactions, the appellants had TOM’s promissory note and the 3 million FIF units, and were indebted to TOM in a substantially equal amount. Thus, they were in exactly the same “economic”
position before and after the two-step exchange of their FMO units for FMO’s assets; and they were in the same economic position as before the FMO reorganization commenced. They had no change in net assets, or liabilities—they incurred no liability for Part I tax.
[234] All of these transactions resulted in only one thing from the appellants’ perspective: their $113 million CDA addition that enabled the appellants to pay capital dividends, while avoiding Part III tax.
[235] Clearly then, the avoidance transactions were undertaken with the following objectives:
(i) maximizing the value of the FVT units, so that FMO’s capital gain realized on the disposition of those units would be maximized, by causing the appellants to purchase the public FMO units from FVT and by delaying FVT’s income distribution;
(ii) distributing that significant capital gain exclusively to the appellants, notwithstanding that they had no economic gain, with the sole objective of maximizing their CDA addition, so they could pay capital dividends to their parent corporations without incurring Part III tax;
(iii) the parent corporations paying capital dividends to Mr. Grenon so that they could distribute to Mr. Grenon, as tax-free dividends, amounts that could otherwise only be distributed to him as taxable dividends;
(iv) maximizing the appellants’ cost of the FMO units so that they could incur a loss on FMO’s redemption of those units that would offset their taxable capital gains from FMO’s distribution, eliminating any Part I tax liability for the appellants; and
(v) distributing FVT’s income to TOM so that the income could be distributed to RRSP Trust, a tax-exempt.
None of these transactions was necessary to achieve the objectives of the FMO reorganization. A review of the alternative arrangement in Appendix A makes this abundantly clear. They are all avoidance transactions.
[236] As in Triad Gestco, I am satisfied avoidance transactions frustrated the object, spirit and purpose of the capital gain and capital loss provisions in the Income Tax Act. They resulted, directly or indirectly, in the appellants’ tax benefit—the avoidance of Part III tax on capital dividends the appellants paid to their parent corporations—because they served no purpose other than to create CDA. That CDA in turn enabled the appellants to pay capital dividends while avoiding Part III tax despite having no real capital gains.
[237] In my view, that abuse is sufficient to sustain the Part III assessments.
(b) Abuse of the object, spirit and purpose of the capital dividend provisions
[238] However, I also agree that the series of transactions abused the CDA provisions and that abuse led to at least one other tax benefit, that is not the subject of this appeal, Mr. Grenon’s capital dividends. That said, the appellants played a role in that abuse.
[239] In Gladwin, the taxpayer used the CDA it created to distribute as tax-free capital dividends amounts “other than qualifying amounts”— in that case, the
taxable portion of its actual capital gain which cannot be distributed tax-free—in circumstances in which the offsetting negative CDA would never be realized. The Court determined using the CDA in that way was abusive of the capital dividend provisions.
[240] Unlike the taxpayer in Gladwin which had a real economic gain and income that it could distribute only as taxable dividends, the appellants had no income or real capital gains to distribute. Distributing their income that otherwise could only be distributed as a taxable dividends was not the purpose of creating their CDA and paying capital dividends.
[241] Indeed, two of the appellants owned nominal net assets before acquiring the FMO units, and had nominal equity and retained earnings. Mr. Grenon confirmed they carried on no activity before participating in the FMO reorganization: see Undertakings, Answers, and Updates from Examination for Discovery of James T. Grenon, Appeal Book at 12274. With no net assets, those appellants paid their capital dividends by issuing stock dividends. While the third appellant had had some other activities before the FMO reorganization, it too had nominal retained earnings and equity, and the more than $47 million capital dividends it paid to its parent corporation created an almost equal retained earnings deficit. There was no change to the appellants’ net assets as a result of their participation in the FMO reorganization.
[242] However, the parent corporations added the capital dividends they received to their CDA and they in turn paid capital dividends to Mr. Grenon. The obvious inference is that this enabled the parent corporations, like the taxpayer in Gladwin, to distribute as capital dividends amounts that they otherwise could have distributed to Mr. Grenon only as taxable dividends. But, any tax benefit Mr. Grenon had is not the subject of the assessments at issue here.
[243] Here to deny the appellants’ tax benefit, it is sufficient to establish that the capital gains and capital loss provisions were abused leading to that tax benefit—avoidance of Part III tax on payment of significant capital dividends. But for the abuse, the appellants would have been liable for Part III tax on those capital dividends.
(4) Are the tax consequences reasonable in the circumstances?
[244] Where GAAR applies, the tax consequences to a person shall be determined as are reasonable in the circumstances in order to deny a tax benefit that would result, directly or indirectly, from a series of transactions that includes an avoidance transaction: s.245(2).
[245] As defined, “tax consequences”
includes the tax or other amount payable under the Income Tax Act or any other amount that is relevant for purposes of computing that amount (i.e., the tax): s. 245(1). In determining the tax consequences as are reasonable in the circumstances to deny a tax benefit that would result from an avoidance transaction, the tax effects that would otherwise result from the application of other provisions of the Income Tax Act may be ignored: s. 245(5)(d).
[246] The appellants argue that Part III tax is not a reasonable tax consequence for three reasons.
[247] First, they submit that Part III tax does not remove any tax benefit from the corporate appellants.
[248] Second, they argue that Part III imposes a penalty, not a tax, and so the Minister was precluded from assessing on that basis, citing the Tax Court decision in Copthorne Holdings Ltd. v. The Queen, 2007 TCC 481, [2008] 1 C.T.C. 2001 at para. 77 [Copthorne TCC]. They note that Gladwin did not address this issue because that appeal concerned a notice of determination of the taxpayer’s CDA, not a Part III assessment.
[249] Finally, the appellants submit that even if Part III tax is not a penalty, it is not a reasonable tax consequence because “[t]he 60% levy is arbitrary”
and “[s]imply because there is a 60% levy in the [Income Tax Act] does not make that the benchmark for reasonableness”
: Appellants’ Memorandum of Fact and Law at para. 100.
[250] I disagree.
[251] The assessments remove the appellants’ tax benefit by eliminating the appellants’ capital gains and capital losses, and the commensurate CDA addition. The Part III assessments do not operate alone. Rather they result from tax consequences—ignoring the tax effects that would otherwise result from the application of other provisions of the Income Tax Act—that eliminate the capital gains and capital losses. Those tax consequences are in turn reasonable to deny the tax benefit because the capital gains and capital losses were created, in circumstances where the appellants had no economic gain or loss, for only one purpose: creating the CDA that enabled the appellants to pay capital dividends while avoiding Part III tax.
[252] Again this is not unique. In Copthorne, the tax benefit was the avoidance of withholding tax. There, like here, the tax consequences determined as reasonable in the circumstances to deny that tax benefit was not only the assessment of withholding tax but also the reduction of the Canadian corporation’s paid-up capital—an amount relevant for the purposes of computing the withholding tax. The assessment of the avoided withholding tax followed from that reduction of paid-up capital; the tax benefit was thereby denied.
[253] I turn now to the appellants’ second and third arguments: that Part III tax is a penalty and the rate of Part III tax is not reasonable in the circumstances.
[254] It is true that the Tax Court in Copthorne TCC decided that a penalty under subsection 227(8) for failure to withhold tax could not be imposed where the tax consequence of a GAAR assessment was the imposition of withholding tax. That said, the definition of “tax consequences”
includes tax or other amounts payable under the Income Tax Act and thus does not on its face preclude the imposition of a penalty as a tax consequence. For purposes of this appeal, however, it is sufficient to say that I do not consider Part III tax to be comparable to a penalty for failure to withhold the avoided tax. Part III tax is the tax avoided; it is the tax benefit denied.
[255] Finally, where the tax benefit is the avoidance of tax, an argument that a reasonable tax consequence does not include liability for the avoided tax is without merit. GAAR does not invite an inquiry into whether the rate of tax payable, once the tax benefit is denied, is reasonable in a quantitative sense. The rate of Part III tax applied to the appellants’ excess capital dividends is exactly the same as the rate applicable to any excess capital dividend, including those paid through genuine mistake in non-abusive circumstances. Imposing exactly the same tax in a GAAR context is therefore clearly reasonable.
VII. Additional Comments: Appellants’ Elections to Treat the Dividends as Taxable
[256] Part III tax is the primary tax liability that arises when an excess capital dividend is paid. However, shareholders who receive such a dividend are jointly and severally liable for a portion of the corporation’s Part III tax, based on their proportionate share of the excess capital dividend: s. 185(4).
[257] A corporation that is liable for Part III tax may elect to treat the excess capital dividend as a separate taxable dividend paid to the shareholders who received it, provided the conditions for the taxable dividend election are met: s. 184(3). In that event, the corporation does not have an excess dividend and is not liable for Part III tax. However, its shareholders are then treated as having received a taxable dividend and thus may be liable for tax. In addition, penalties and interest may be payable as a result of the election: ss. 184(3), (4) and (5).
[258] The respondent’s Amended Replies stated that, in assessing the appellants’ tax liability under Part III, the Minister assumed that, on March 21, 2014—after the initial Part III assessments—the appellants “made protective [taxable dividend] elections under subsection 184(3) to treat the excess dividends as taxable dividends”
: Amended Replies at para. 15 (jjj), Appeal Book at 2006. Neither the relevance of this assumption to the correctness of the Part III assessments, nor what is meant by the elections being “protective”
, is explained. (My assumption is that the appellants asked that the elections be held in abeyance and not processed pending the appeal of the Part III assessments.)
[259] In its reasons, the Tax Court said that, as a result of “[its] finding that the additions to the capital dividend accounts were a sham and a misrepresentation, it follows that the Appellants are not entitled to rely on the [taxable dividend] elections filed pursuant to subsection 184(3) to treat the excess dividends as ordinary taxable dividends”
: reasons at para. 236.
[260] The appellants submit that neither the validity nor invalidity of the taxable dividend elections was pleaded and therefore was not before the Tax Court. Consequently, they say, the Tax Court erred in addressing their validity and determining that the appellants could not rely on the taxable dividend elections.
[261] I first observe that the Tax Court’s statements appear to be based on its findings regarding sham. I have explained why I do not agree with those findings. Nonetheless, I agree that the validity of the taxable dividend elections was not properly before the Tax Court.
[262] The Tax Court’s role on the appeal was to address the correctness of the assessments. Nothing in the record suggests that the appellants relied on the elections to support their position that those assessments were incorrect. And nothing in the record suggests that the respondent sought to have those elections declared invalid. The only relief the respondent sought was the dismissal of the appeals of the Part III assessments.
[263] Having said that, the Tax Court’s views about the taxable dividend elections are entirely obiter. Notably, the Tax Court’s judgments say nothing about the validity of those elections, and are appropriately limited to dismissing the appeals from the assessments. Any dispute about the validity of the elections cannot be resolved on this appeal.
VIII. Conclusion
[264] I am satisfied that GAAR applies. The avoidance transactions that led to the tax benefit—the appellants’ avoidance of Part III tax—frustrated the object, spirit and purpose of the capital gain and capital loss provisions of the Income Tax Act. I am satisfied that the elimination of the appellants’ capital gains and capital losses by the Part I assessments, with the commensurate elimination of the additions to their CDA and the resulting Part III assessments, are reasonable tax consequences to deny their tax benefit.
[265] Accordingly, I would dismiss the appeal with costs.
“K.A. Siobhan Monaghan”
“I agree.
Richard Boivin J.A.”
“I agree.
George R. Locke J.A.”
Email this Content