Is My IFA Tax Shelter (Statutory Review)

This memorandum provides a comprehensive framework for analyzing whether an Immediate Financing Arrangement (IFA) qualifies as a tax shelter under the Income Tax Act (ITA). It systematically examines the statutory definition of a tax shelter, the identification of the relevant property, the connection between the loan and that property, and the classification of the financing as a limited-recourse amount. Furthermore, it outlines the resulting prescribed-benefit calculation, details the consequences of operating an unregistered tax shelter, and addresses common objections from lenders and promoters. Ultimately, the objective of this document is to equip readers with the objective statutory questions required to make that determination themselves.

PDF Download Printable PDF

---

Background: The IFA Structure and Why Deductibility Matters

In a standard IFA structure, a client purchases a permanent life insurance policy and pays the premium out of pocket. Immediately afterward, a bank lends the client money against the policy’s cash value as typically illustrated. The loan proceeds are then invested. In a typical illustration, the net cash outlay in the year of the transaction is shown as $0. The debt is generally intended to remain outstanding for life and to be repaid from the tax-free proceeds of the policy on death, so the illustration typically shows the projected loan subtracted from the death benefit to arrive at a “net death benefit” figure.

While an IFA renders the client cash-neutral in the year of the transaction, it does not make the life insurance free. Instead, it replaces a known, finite premium obligation with a lifelong debt that must be serviced annually.

The cost of carrying that debt is the annual interest charge, which is maintained for the life of the insured. The arrangement’s economics depend heavily on whether this interest charge can be rendered tax-deductible under the Income Tax Act. If the loan were used directly to pay for the insurance premium, that deduction would not be a possibility, as paragraph 20(1)(c) explicitly excludes interest on money borrowed to acquire a life insurance policy.

Without a tax deduction, this lifelong interest burden is highly problematic. If the interest is capitalized and re-collateralized annually, the compounding debt could eventually outrun the policy's cash value, leaving the client with little or no insurance net of the debt. Alternatively, if the interest is paid out-of-pocket annually, the client is exposed to interest rate variability; over a lifetime, this cumulative after-tax cost could easily exceed what it would have cost to purchase the target net death benefit outright using standard, unfinanced insurance products.

To render the interest deductible and thus the plan financially viable, the IFA’s novel representation relies on specific sequencing and the deliberate, represented use of funds. By paying the premium out of pocket first, and then representing that the borrowed funds are used strictly to acquire an income-producing investment (and are not related to the payment of the insurance premium), the resulting tax deductions are represented as achieved, and the illustrated tax savings are integrated directly into the arrangement's stated financial quantum.

This sequencing dictates the financial outcome of the overall plan—and the illustration presents the policy, the loan, the investment, and the tax-deductible interest together as a single, cohesive strategy. Yet, that very design choice—representing that the full amount of the loan was invested to produce a deduction—is what raises the tax-shelter question under the ITA’s tax-shelter definition, as set out below.

The Statutory Definition of a Tax Shelter

Subsection 237.1(1) of the Income Tax Act defines a tax shelter, in relevant part, as an investment in property where it is reasonable to consider—having regard to statements or representations made or proposed to be made in connection with the property—that a purchaser could reasonably be expected to have deductions, losses, or credits within the first four years equal to or exceeding the purchaser’s net cost. Net cost is determined strictly after subtracting any “prescribed benefits” expected to be received directly or indirectly in respect of the property.

This produces three variables that must be established before the test can be applied:

Variable	Question	Statutory Basis
A	What is the cost of the property acquired?	ITA s. 237.1(1)(b)(ii)(A) — cost determined without reference to s. 143.2
B	What prescribed benefit — including any limited-recourse financing — is expected in respect of that property?	ITA s. 237.1(1)(b)(ii)(B); ITR Reg. 3100; ITA ss. 143.2(7) and 143.2(12)
C	What deductions, losses, or credits are represented in the first four years?	ITA s. 237.1(1)(b)(i)

The statutory test under subsection 237.1(1) resolves to a simple equation:

If C ≥ (A − B), the arrangement crosses the tax-shelter threshold and registration is required.

This test is triggered by representations made or proposed to be made — through sales meetings, promotional materials, or financial illustrations — rather than by the arrangement’s actual commercial outcome. The phrase “proposed to be made” is significant: the tax-shelter analysis can be engaged before the representations are actually made to a purchaser, based on what is intended to be represented. Subsection 237.1(4) requires a Tax Shelter ID Number before the arrangement may be sold or issued, regardless of whether the represented deductions are ultimately sustained on audit.

In an IFA, the design directly sets up a potential $0 net cost. In that context, this statutory calculation is not peripheral; overlooking it exposes the arrangement and its participants to severe penalties, regardless of intent.

The sections that follow walk through this three-part statutory cascade as applied to a standard IFA: first 1) identifying the relevant property (A), then 2) establishing whether the loan constitutes a prescribed benefit in respect of that property (B), and finally 3) applying the arithmetic to the represented deductions (C).

1. Identifying the Relevant Property (A)

Variable A is the independently identifiable investment property acquired with the borrowed funds—the very investment assumed in the illustration to qualify for a paragraph 20(1)(c) interest deduction. This follows directly from the structure of paragraph 20(1)(c). Under the Act, interest is deductible only where it is paid on “borrowed money used for the purpose of earning income from a business or property,” while interest on money borrowed “to acquire a life insurance policy” is expressly excluded. Consequently, if an IFA illustration represents interest as deductible, that representation is possible only on the basis that the borrowed funds are used to acquire a qualifying, income-producing investment. The illustration’s own “deductible interest” line is, in effect, a representation that a qualifying investment property exists to which the loan was applied. The illustration need not name the investment or state its cost. The test asks what the representations, objectively read, communicate (Baxter, paras 14 and 42), and a communication that describes tax attributes without quantifying the deduction is sufficient (Krumm v. Canada, 2021 FCA 78).

Where the illustration represents the entire interest charge on the loan as deductible on the basis that the entire loan was used to acquire the investment, that representation identifies the full amount of the loan as financing for that investment.

2. Establishing the Prescribed Benefit (B)

Once the relevant investment property has been identified (Variable A), the next step in the statutory formula is determining whether a "prescribed benefit" (Variable B) exists in relation to that property. Under Regulation 3100(3), treating the bank loan as a prescribed benefit requires establishing two distinct criteria:

1. In Respect Of: The financing must be connected to, or made “in respect of”, the property.
2. The Limited-Recourse Amount: The loan itself must legally qualify as a limited-recourse amount under section 143.2 of the Income Tax Act.

Only when both of these criteria are satisfied does the financing become a prescribed benefit that reduces the net cost of the property in the statutory calculation.

2.1 The “In Respect Of” Connection

It must be shown that the loan is “in respect of” the property, since Regulation 3100 defines a prescribed benefit by reference to amounts reasonably expected in respect of an interest in property. The phrase “in respect of” has been given the widest possible scope by the Supreme Court of Canada in Nowegijick v. The Queen, [1983] 1 S.C.R. 29:

“The words 'in respect of' are, in my opinion, words of the widest possible scope. They import such meanings as 'in relation to', 'with reference to' or 'in connection with'. The phrase 'in respect of' is probably the widest of any expression intended to convey some connection between two related subject matters”

Here, the connection is direct rather than incidental: the loan is represented as the source of the deductible interest, deductible only because the loaned funds were used to acquire the investment. The promoter cannot obtain the paragraph 20(1)(c) deduction without representing that qualifying use of the borrowed funds — the same borrowing tested for limited-recourse status under section 143.2. The represented deduction is thus tied to the loan, and the loan to the investment, satisfying the Nowegijick standard. This tracing is a feature of paragraph 20(1)(c), not unique to an IFA — it is the loan's limited-recourse character under subsection 143.2(7) that does the distinguishing work.

2.2 Is the Loan a Limited-Recourse Amount?

The second element is whether the loan is a “limited-recourse amount.” Industry practitioners often point to outside collateral, personal guarantees, and the lender’s ability to demand repayment as evidence that the debt is “fully recourse.” These features are relevant to a bank’s commercial credit assessment, but they do not answer the statutory question. Subsection 143.2(7) is a deeming provision that overrides ordinary commercial characterization:

“…the unpaid principal amount of the indebtedness shall be deemed to be a limited-recourse amount… unless bona fide arrangements evidenced in writing, at the time the indebtedness arose, for repayment of all principal and interest… within a reasonable period not exceeding 10 years.”

The central issue for a standard IFA is the principal-repayment condition: a bona fide written arrangement, made at inception, to repay all principal within 10 years. IFA loans are typically structured as interest-only demand loans or as facilities with a stated term of ten years or less, intended to remain outstanding until death. While s. 143.2(7) also requires interest to be paid annually within 60 days of the tax year-end, satisfying that separate rule does not bypass the 10-year principal requirement.

Promoters are likely to argue that an annually renewable demand loan is by its own terms shorter than 10 years. Subsection 143.2(12) anticipates this argument: where a repayment arrangement can reasonably be considered part of a “series of loans or other indebtedness and repayments” extending more than 10 years, the debtor is treated as not having made arrangements to repay within 10 years.

Where a facility has a stated term of ten years or less, the question becomes whether, when the loan was made, a reasonable person would have contemplated replacement on substantially similar terms (Cassan, 2017 TCC 174, paras 364-368). Where the loan is secured by cash value held by the insurer, a year-ten review is unlikely to function as a genuine requalification, and an illustration showing the loan outstanding for life is relevant evidence on that question. The outcome depends on the actual facility terms.

A taxpayer’s assertion of an honest intention to repay does not establish a bona fide arrangement; the arrangement “should readily be seen to be binding upon the parties” and be “prima facie obvious”, and the surrounding documents must “support and reinforce” the written terms (Tolhoek v. The Queen, 2006 TCC 681, paras 29, 40).

2.3 Completing the Prescribed-Benefit Analysis

Once a loan is established to be both (2.1) in respect of the relevant property, and (2.2) a limited-recourse amount, Regulation 3100(3) applies with statutory automaticity. It provides that a limited-recourse amount in respect of a property, determined under subsection 143.2(1), (7), or (13), is a prescribed benefit in respect of that property.

With the property identified (A) and the loan established as a prescribed benefit in respect of that property (B), both inputs required for the tax-shelter calculation are in place.

3. Identifying the Represented Deductions (C)

The final element of the three-part quantum is determining whether deductions, losses, or credits have been represented in the first four years of the plan (Variable C). In the context of a standard IFA illustration, the presence of these represented deductions—specifically, the paragraph 20(1)(c) interest deduction—is an obvious point that requires no further complex analysis, as it is explicitly printed on the illustration itself.

However, as will be shown in the calculation shortly, the actual quantum of these represented deductions need only be greater than zero. In other words, there simply must be at least some representation of a deduction. As we will see, the specific dollar amount of that deduction becomes legally irrelevant in the mathematical context of a fully financed IFA.

The Tax-Shelter Calculation

With variables A, B, and C established, the final step in determining unregistered tax-shelter status is arithmetic rather than interpretation. The statutory test in subsection 237.1(1) requires registration if the represented deductions (C) equal or exceed the cost of the property (A) minus the prescribed benefits (B). The formula is:

If C ≥ (A − B) → Tax Shelter

Because the IFA illustration represents that the full amount of the investment was acquired with the borrowed funds — meaning the acquisition cost of the represented investment and the borrowed amount are coextensive — the cost of the property (A) and the prescribed benefit against it (B) are identical.

Therefore: (A − B) = $0

Once the net cost is $0, the only remaining question is whether the represented deductions in the first four years (C) exceed $0. Because C > $0 is satisfied the moment C is any amount greater than zero, the size of the represented deduction is legally irrelevant to the outcome of the test: a represented deduction of even $1 satisfies the threshold.

This is the precise mechanism by which a fully leveraged IFA, represented and structured in the manner described above, mechanically satisfies the statutory tax-shelter definition without requiring any additional facts or abusive intent.

A Note on Partial Financing

The analysis does not depend on financing 100% of the insurance premium. The relevant comparison is not between the premium and the loan; it is between the acquisition cost of the property represented as having been acquired with the borrowed funds and the limited-recourse amount treated as a prescribed benefit in respect of that property.

Thus, if $70,000 is borrowed and the entire $70,000 is represented as being used to acquire an investment having an acquisition cost of $70,000, the relevant amounts are A = $70,000 and B = $70,000. The percentage of the insurance premium financed is not itself determinative. What matters is whether the borrowing and the acquisition cost of the represented property are coextensive.

An advisor might argue that an unfinanced premium balance (e.g., paying $30,000 out-of-pocket on a $100,000 premium) injects positive net cost into the arrangement. That argument succeeds by abandoning the investment as the testable property and asserting either that the policy itself is the property, or that the policy and investment form a single composite package. See the Addendum: A Final Point on Property Characterization below.

Consequences of Unregistered Tax-Shelter Status

Where an arrangement meets the statutory definition of a tax shelter, the promoter is required, before selling, issuing, or accepting consideration in respect of the arrangement, to obtain a Tax Shelter Identification Number from the CRA and to file the required prescribed information, including a list of subscribers. Obtaining an identification number does not signify CRA approval of the arrangement or of any represented deduction; it is a disclosure mechanism, and registration typically results in increased audit scrutiny of the arrangement. If an arrangement meeting the tax-shelter definition is sold without registration, or if the required filings are not made, the following consequences may apply:

• Denial of deductions. Under subsection 237.1(6), no amount may be deducted or claimed in respect of a tax shelter unless the prescribed form containing the tax-shelter identification number has been filed.
• Reduction of cost basis. Section 143.2 contains rules that can reduce the capital cost or other expenditure amount of a tax-shelter investment by the amount of any specified limited-recourse debt associated with it.
• Promoter and third-party penalties. Under subsection 237.1(7.4), selling or accepting consideration in respect of an unregistered tax shelter can result in a penalty equal to the greater of $500 and 25% of the consideration received or receivable from a particular person in respect of the tax shelter. Separately, section 163.2 provides for third-party civil penalties where a person makes or participates in a false statement in connection with a taxpayer’s affairs.
• Extended reassessment period. Where a tax shelter or limited-recourse amount is involved, subsection 143.2(15) permits reassessment outside the normal reassessment period (Liang v. Canada, 2013 TCC 336).

A Diagnostic Test for Advisors

An advisor or promoter who disagrees with the application of this analysis to a particular arrangement does not need to accept or refute the analysis in the abstract. The following can be applied directly to the specific illustration and loan documentation at issue:

1. What property does the illustration represent as being acquired with the borrowed funds? If the answer is a qualifying income-producing investment intended to satisfy paragraph 20(1)(c), the relevant property has been identified.
2. Why is the interest on the loan represented as deductible? If the answer depends on the borrowed funds having been used to acquire that income-producing investment, the loan-to-property relationship has been expressly represented.
3. Was there a bona fide, written arrangement, made at the time the debt arose, requiring repayment of all principal and interest within 10 years? If yes, that arrangement should be identified and reviewed. If no, subsection 143.2(7) must be addressed.
4. If the loan is a limited-recourse amount, on what basis is it not a prescribed benefit in respect of the investment identified in Question 1, given Regulation 3100(3)?
5. If the loan is a prescribed benefit in respect of the investment, what is the resulting statutory net cost under the formula in Section 2?
6. If the represented deductions in the first four years are greater than $0, and the net cost calculated in Question 5 is $0, how does the arrangement avoid satisfying the numerical test in subsection 237.1(1)?

Conclusion

A mechanical tax-shelter analysis of an IFA does not require any speculative recharacterization of the arrangement or resort to the General Anti-Avoidance Rule (GAAR). Following the promoter’s own representations and the taxpayer’s own tracing of borrowed funds for paragraph 20(1)(c) purposes is sufficient to raise the statutory questions addressed above: what property was acquired, is the loan financing that property a limited-recourse amount, is that amount a prescribed benefit in respect of the property, and are the represented deductions in the first four years equal to or exceed the net cost.

Where those facts converge — borrowing represented as fully invested to generate a represented deduction, and lacking a bona fide 10-year repayment arrangement — the arrangement would satisfy the numerical test in subsection 237.1(1), and the registration requirements of subsection 237.1(4) would apply before the arrangement may lawfully be sold. Any professional reviewing a specific IFA arrangement should be able to identify the precise point in this statutory chain at which the facts of that particular arrangement diverge from this analysis. Where they do not diverge, the statutory consequences apply regardless of the intentions of the parties involved.

David Kakon, B.Sc. (Math Hons.)
Financial security advisor
[email protected]

---

Anticipating the Objections

“This would make every investment loan a tax shelter.” It would not. An ordinary investment loan typically involves an independently selected investment, borrowing sized to the investor’s own financing needs, a bona fide repayment arrangement, and no representation that the arrangement produces a $0 net cost. Those facts do not produce the same result. The argument advanced in this memorandum is narrower and fact-specific: it applies where the borrowed amount is represented as invested, the resulting interest is represented as fully deductible, and the principal is intended to remain outstanding beyond 10 years. That is a specific fact pattern, not a description of investment borrowing generally.

“The bank treats the loan as fully recourse.” This observation, while accurate as a description of the bank's commercial position, does not answer the statutory question. A personal guarantee, additional collateral, a right of set-off, or the lender's ability to demand repayment may all be relevant to a bank's internal credit risk assessment, but none of them constitutes a bona fide written arrangement made at the time the debt arose to repay the principal within 10 years, as subsection 143.2(7) requires. Nor does a stated 10-year term settle the matter: under subsection 143.2(12), a facility retired only by drawing a new loan against the same policy is treated as part of a single series extending beyond 10 years, regardless of how the individual instrument is papered — and a stated term is not a bona fide arrangement if the surrounding documents and conduct do not support it (Tolhoek, para 29). The commercial and statutory characterizations of the same loan can diverge, and it is the statutory characterization that governs the tax-shelter test.

“The broker sold a policy, not the investment, so they are not promoters of the tax shelter.” The definition of “promoter” under subsection 237.1(1) is functional rather than transactional. It does not require a person to be the legal vendor of the underlying investment property. Instead, it captures anyone who, in the course of business, promotes the acquisition of the tax shelter, acts as an adviser or agent in respect of it, or accepts consideration in connection with it—and multiple parties can be promoters of the same arrangement. The case law supports a functional reading. In Baxter (2007 FCA 172, paras 44-45), the Court treated both the issuer and the independent agents who distributed its materials as promoters, while noting that the identity of the representor was not the focus of argument (para 44). In Kinglon Investments Inc. v. Canada, 2015 FCA 134 (para 20), the Court treated the comments in Baxter (paras 9 and 44) as obiter and, at the pleadings stage, declined to strike the Crown’s position that neither definition requires the representor to be the seller.

"IFAs are private, individually advised plans." The statutory definition of a tax shelter contains no public-marketing requirement. In Krumm (2020 TCC 7, aff’d 2021 FCA 78), a single private sale was held to be a tax shelter. Furthermore, the statutory test is strictly objective. It evaluates the representations, not the individual purchaser’s subjective understanding of the transaction or the nature of the advisory relationship (Baxter, para 54).

"The illustration says deductibility is assumed, calls itself a concept presentation, and says the policy is separate from the lending." Courts look to the substantive communication of the materials, not their boilerplate labels. In Baxter, the Federal Court of Appeal held that a tax opinion explicitly subject to assumptions still constituted a representation for the purposes of the Act (paras 18 and 38). Likewise, scattering the tax claims and the lending details across separate documents does not break the statutory connection if they are presented to the client as a coordinated strategy (para 39). Slapping an "assumed" disclaimer on a deduction, or calling a fully integrated financial illustration a mere "concept," does not shield the arrangement from the tax-shelter test.

“The lender requires us to requalify for the loan before the 10-year mark, satisfying the bona fide arrangement test.” As a threshold matter, this defense assumes the credit agreement contains a stated 10-year term; many standard IFA bank facilities are structured as open-ended demand loans with no fixed maturity date, and may not contain a bona fide written repayment arrangement required by paragraph 143.2(7)(a). Where a facility does contain a stated term of ten years or less, relying on Cassan v. The Queen (2017 TCC 174) remains problematic. In Cassan, loans with stated maturities inside ten years were held to be limited-recourse amounts on two grounds: first, the repayment arrangements were not bona fide when measured against arm’s-length commercial practice (paras 344–355); and second, subsection 143.2(12) applied because a reasonable person would have contemplated replacement on similar terms (paras 364–368). The second test is objective and does not require an express promise of renewal. While Cassan arose under the charitable gift rules on facts involving thin credit underwriting—meaning a facility with documented commercial underwriting may be distinguishable on the first ground—it remains vulnerable under the second. Where the facility is secured primarily by policy cash values held by the insurer, and the sales illustration models the debt remaining outstanding until death to be repaid from policy proceeds, a year-ten review is unlikely to function as a genuine commercial requalification. Under an objective test, the arrangement reasonably forms part of an extended series of loans under subsection 143.2(12). The ultimate determination depends on the specific terms of the credit agreement.

Scope and Limits of This Analysis

This analysis does not conclude that all IFAs—or all insurance-backed investment loans—are tax shelters. Nor does it validate the underlying tax deductions. The statutory test remains highly fact-specific and requires a review of the actual loan agreements, fund tracing, and representations in each individual case.

What it does establish is highly specific: where an IFA’s borrowing is represented as fully invested to generate deductible interest, and lacks a bona fide 10-year principal repayment arrangement, the statutory net cost resolves to $0. Consequently, any represented deduction in the first four years mechanically triggers the tax-shelter registration requirement under subsection 237.1(4). This is strict statutory arithmetic—independent of the parties’ intentions and entirely separate from the General Anti-Avoidance Rule (GAAR).

Industry and Regulatory Context

This analysis is not an isolated legal theory; the statutory tension it highlights has been noted by a major Canadian life insurance carrier. In its September 2026 published guidance on leveraged life insurance ("Leveraged life insurance under personal ownership"), Manulife stated that the tax shelter rules "may be a consideration" where leveraging occurs within the first four years for these arrangements. While Manulife observes that there are no reported court decisions applying tax-shelter rules to leveraged life insurance planning, and suggests that the mathematical test would likely not be met for a traditional strategy if the policy is included in the property or if the debt is not deemed limited-recourse, it nevertheless highlights the exact statutory triggers:

• The Net Cost Math: Manulife notes that limited-recourse debt reduces the cost base, noting that deductions "may exceed that 'cost', and therefore may trigger the tax shelter rules."
• The Representation Threshold: Manulife notes that this requirement may be met if interest or collateral insurance deductions under paragraph 20(1)(e.2) are verbally discussed or illustrated in connection with a tax shelter property.
• The 4-Year Trigger: The carrier explicitly notes that these rules apply "where there is a leveraging within the first 4 years of the arrangement," a common feature for an Immediate Financing Arrangement.
• Broad Definition of Property: Manulife highlights that under subsection 248(1), "property is broadly defined to include 'property of any kind'," meaning the property requirement under the tax shelter test is "easily satisfied."

If carrier guidance is not enough to prompt caution, advisors must look to the Canada Revenue Agency’s current enforcement posture. At the 2026 CRA Roundtable (Reference: 2026-1089291C6), the CRA stated generally that it audits arrangements using insurance products to obtain a tax advantage—auditing both the taxpayers and the promoters:

“We continue to seek out, audit and re-assess all arrangements which attempt to utilize insurance products or insurance-like products to gain a tax advantage not envisioned by tax policy and legislation. Within CRA, the High Net Worth Compliance Directorate (HNWCD) develops, implements, and coordinates business intelligence and compliance initiatives to address non-compliance in the High Net Worth population through the Audit Program Division, which includes the Aggressive Tax Planning audit program (ATP), and the Tax Promoter and Advisor Compliance program (TPAC).

While the ATP Audit Program addresses aggressive tax planning by taxpayers who participated in such abusive insurance-related tax schemes, the TPAC program addresses non-compliance associated with promoters and advisors who promote or sell aggressive tax schemes. TPAC is accountable for identifying and combating aggressive tax schemes and holding promoters accountable through audit activities.”

For any advisor or promoter, the convergence of these two realities requires close attention. One carrier has noted that the tax shelter rules 'may be a consideration' and identified the statutory triggers, and the CRA is deploying dedicated audit teams to scrutinize promoters; thus, the application of these rules is a development worth monitoring.

Addendum: A Final Point on Property Characterization

It is useful to briefly test alternative characterizations of the relevant “property” under subsection 237.1(1) to confirm that shifting the framework does not bypass the statutory prescribed-benefit analysis.

1. The Investment as Property

If the illustration represents that the borrowed funds were invested and the resulting interest is deductible, the income-producing investment is independently identifiable as the testable "interest in property" under Regulation 3100 to which the represented deduction relates. By presenting the investment as the sole engine for the tax deduction, the illustration isolates it as the primary property against which the limited-recourse financing and the resulting tax benefits must be mathematically measured.

2. The Policy as Property

The policy route is advanced only in the alternative and requires its own showing. This memorandum does not count the same borrowing against both properties. For the policy to be the property tested, the loan must be a prescribed benefit 'in respect of' the policy. 'In respect of' has the widest possible scope (Nowegijick), and a loan secured by the policy's cash value and repaid from its death benefit is connected to the policy in the ordinary sense of those words. The paragraph 20(1)(e.2) collateral insurance deduction represented on the policy supplies Variable C for that property. The route has limits: the paragraph 20(1)(c) deduction is possible only because the borrowed funds are represented as used to earn income, so the illustration itself represents the loan as financing the investment. If the policy is established as the property, the prescribed benefit reduces its net cost and any represented deduction satisfies the test. If not, the inquiry returns to the investment as property (item 1 above).

3. The Integrated Arrangement ("Composite Property")

Promoters and carriers may argue that an IFA is an integrated package, merging the policy and the financed investment into a single “composite property” to inflate the cost denominator. This packaging argument faces commercial and statutory hurdles:

Distinct Commercial Reality vs. Hexalog: In Hexalog (2005 TCC 67), a single promoter sold a bundled arrangement where neither component had standalone commercial value—the nominal land was useless without the prepaid development contract. An IFA is the exact opposite. A permanent life insurance policy must be medically and financially underwritten on its own merits; the carrier requires the client to demonstrate the independent financial capacity to justify the coverage absent any loan. The investment is equally distinct in ownership, custodian, and commercial character. Merging two independently viable assets acquired from unrelated institutions into a single property distorts commercial reality.

One Property at a Time (Seica): In Seica (2021 QCCA 1401), persuasive authority from the Quebec Court of Appeal applying Quebec's provincial counterpart rules, the Court confirmed that Hexalog was a fact-specific decision and rejected the idea that marketed packages become a single property as a general rule (paras 41–57, with specific emphasis on para 55). The tax-shelter calculation under subsection 237.1(1) evaluates “a property” and tests deductions and net cost against “the property”. The arithmetic must be applied to each property individually. Having two distinct properties in a plan means testing each property separately—it does not authorize aggregating an unleveraged policy to inflate the cost base of a leveraged investment.

Paragraph 20(1)(c) / 20(1)(e.2) Tension: Maintaining this position creates a tension that is difficult to reconcile. Paragraph 20(1)(c) allows an interest deduction only where borrowed money is traced to an income-producing investment, explicitly barring deductions for money borrowed to acquire life insurance (though paragraph 20(1)(e.2) may allow a distinct deduction for collateral insurance premiums where specific conditions are met). If the policy and investment truly form an inseparable "single property," asserting that the loan proceeds did not finance the acquisition of the policy becomes difficult to maintain.

Appendix A: The Statutory Chain

• ITA s. 237.1(1) ("Tax Shelter"): Defines a tax shelter by comparing represented deductions, losses, or credits in the first four years against the property's cost less any "prescribed benefits."
• ITA s. 237.1(1) ("Promoter"): Defines a promoter functionally to include anyone who, in the course of a business, sells, issues, promotes, or acts as an agent/adviser in respect of a tax shelter.
• ITA s. 237.1(4): Requires a Tax Shelter Identification Number to be obtained before the arrangement is sold, issued, or funded.
• ITA s. 237.1(6): Prohibits taxpayers from claiming any deductions in respect of a tax shelter unless the prescribed form containing the ID number has been filed.
• ITR Regulation 3100(3): Defines a "prescribed benefit" for tax-shelter purposes to include any amount that qualifies as a limited-recourse amount under section 143.2.
• ITA s. 143.2(7): Deems the unpaid principal of an indebtedness to be a limited-recourse amount unless there is a bona fide written arrangement, made at the time the debt arose, to repay all principal and interest within a reasonable period not exceeding 10 years (along with annual interest-payment terms).
• ITA s. 143.2(12): Defeats superficial 10-year maturity dates by extending the analysis to any "series of loans" or expected refinancing that, considered together, extends beyond 10 years.
• ITA s. 143.2(6): Contains rules that reduce the capital cost or expenditure amount of a tax-shelter investment by the specified limited-recourse amounts.
• ITA s. 237.1(7.4): Imposes severe financial penalties on promoters who sell or accept consideration in respect of an unregistered tax shelter.
• ITA s. 163.2: Provides for third-party civil penalties for individuals who make or participate in making false statements (such as invalid tax deductions) under a culpable-conduct standard.

Appendix B: Case Law and CRA Ruling Resources

• CRA Technical Interpretation 2006-0196251C6 and Advance Income Tax Ruling 2009-0340381R3 (2010): CRA positions discussing demand loans and limited-recourse characterization under section 143.2 in the context of leveraged insurance arrangements.
• Nowegijick v. The Queen, [1983] 1 S.C.R. 29: Supreme Court of Canada authority establishing that the phrase “in respect of” has the widest possible scope in tax legislation, requiring only a connection between two subject matters.
• R. v. Baxter, 2007 FCA 172: The tax-shelter test is objective, forward-looking, and applied before any sale. Baxter asks whether representations were made by or on behalf of a promoter (paras 44–45), treating both the issuer and independent distributing agents as promoters. Paragraph 44 notes the identity of the representor was not the focus of argument.
• Kinglon Investments Inc. v. Canada, 2015 FCA 134: At the pleadings stage, the Court treated comments regarding the identity of the representor in Baxter (paras 9 and 44) as obiter and declined to strike the Crown's position that neither definition requires the representor to be the seller (para 20).
• Krumm v. Canada, 2021 FCA 78: Cited for the proposition that the tax-shelter test is strictly objective and applied by reference to what the representations communicate, with no requirement that they explicitly quantify the deduction.
• Tolhoek v. The Queen, 2006 TCC 681 (aff’d 2008 FCA 128): Evaluates “bona fide arrangements” under 143.2(7)(a) as a question of fact. An honest intention to repay is insufficient; the arrangement must be commercially genuine, binding, and prima facie obvious from the surrounding documents and conduct.
• Cassan v. The Queen, 2017 TCC 174: Arose under the charitable gift rules. Applies an objective test under subsection 143.2(12): what a reasonable person would contemplate at inception (para 367). Where replacement on similar terms would be contemplated, the arrangement triggers the "series of loans" deeming rule.
• Seica v. Agence du revenu du Québec, 2021 QCCA 1401: Quebec Court of Appeal applying the provincial counterpart to subsection 237.1(1) (persuasive authority). Upheld a separate-properties finding on deferential review, noting that Hexalog is a fact-specific finding and not a general rule (paras 41–57, specifically para 55). Cost excludes interest on the unpaid balance.
• Hexalog Ltée c. La Reine, 2005 TCC 67: Treated a vendor-marketed package (raw land plus prepaid development fees) as a single composite property on its unique facts; characterized in Seica as a fact-specific finding where the components formed an "inseparable whole."
• Liang v. Canada, 2013 TCC 336: Applies the mathematical tax-shelter test, confirms that obtaining an ID number does not validate the deductions, and upholds reassessments issued beyond the normal period.
• CALU Practice Note, “Best practices for insured leveraging programs” (May 30, 2023): Industry guidance advising insurance professionals on interest deductibility, the collateral insurance deduction, and the application of tax-shelter rules to financing arrangements.

Appendix C: Characteristics of Typical IFA Bank Loans

In industry commentary and loan documentation, IFA bank facilities are commonly described as exhibiting the following features, each of which directly informs the statutory tax-shelter analysis:

• Demand-loan structure lacking fixed maturity: The facility is commonly structured as an interest-only demand loan, which may not contain a bona fide written repayment arrangement required by paragraph 143.2(7)(a) to repay principal within 10 years.
• Death as the anticipated repayment event: The death of the insured is generally treated as the trigger for repaying the outstanding balance, confirming the intended term spans the life of the insured rather than a fixed commercial period.
• Linkage to reinvestment: Loan advances are frequently conditioned on evidence that the proceeds will be reinvested at a value at least equal to the loan amount, linking the financing directly to the investment rather than treating them as independent transactions.
• Controlled tracing restrictions: Loan terms commonly prohibit using the facility—or the policy's accumulated cash value—to pay the insurance premium directly. This enforces the specific "pay premium, then borrow and invest" sequence required to claim a paragraph 20(1)(c) deduction.
• Limits tied to policy metrics, not general credit: Annual advances and overall facility limits are commonly tied to the policy’s cash surrender value, rather than being underwritten as a general-purpose, standalone credit facility.
• Integrated product branding: These facilities are frequently documented and marketed as a named, integrated product tied directly to a specific insurance policy, reinforcing that the arrangement is sold as a single package rather than unrelated financial products.
• Lender tax disclaimers: Loan agreements typically include a disclaimer stating the lender makes no representation regarding the tax consequences of the arrangement. This may leave the advisor/broker as the party making the representations that trigger the tax-shelter analysis.

Any review of a specific IFA arrangement should be conducted against the complete documentation for that arrangement: the actual loan agreement, policy illustration, financing conditions, and marketing materials, rather than by reference to these general characteristics.