Shareholder Borrowing Against Corporate Life Insurance: Does It Hold Up?

A private corporation owns a permanent life insurance policy with substantial cash value. Rather than pulling money out as a taxable dividend, the shareholder borrows personally from a bank — and the corporation pledges its policy as collateral. The loan proceeds arrive tax-free. Interest is capitalized rather than paid. When the policy eventually pays out, the proceeds flow into the corporation, generate a capital dividend account credit, and a tax-free capital dividend funds the repayment.

That is the pitch. Every major Canadian insurer publishes material describing it. Canada Life calls it backend leveraging. Equitable Life sells it as the Corporate Preferred Retirement Solution. Sun Life calls it the Corporate Retirement Strategy. BMO Private Banking runs a dedicated lending program built around it.

So the strategy is real, documented, and mainstream. The question is narrower and more useful: does it survive contact with the Income Tax Act, and does the arithmetic still work at today's interest rates?

This article works through both, with every assumption stated so you can substitute your own.

---

Part 1 — Does the guarantee fee actually solve the tax problem?

The problem

When a corporation pledges its assets so its shareholder can borrow personally, the shareholder is using corporate property for a personal purpose. That is the classic fact pattern for a shareholder benefit under subsection 15(1) of the Act.

The CRA has been explicit about the principle since at least 1992:

No person dealing at arm's length with another person would act as guarantor to a loan of that other person unless he received valuable consideration for the risk involved. Consequently, in our view, a benefit may be conferred on the parent notwithstanding that the subsidiary has not given up its beneficial ownership of any of its property.
— CRA Document 9206835, March 18, 1992

Note what that says: a benefit can arise even though the corporation never parts with anything. Merely standing behind the loan is the thing of value.

What the CRA actually conceded

The industry's entire practice rests on one passage, from the 2006 CRA Roundtable:

Where a shareholder of a corporation borrows money ("Loan") from a financial institution and the corporation guarantees the Loan and/or provides security as collateral to the Loan ("Guarantee") … we will not assess a benefit in the situation where the shareholder is dealing at arm's length with the corporation and there is no evidence that the shareholder is, at the time the Guarantee is granted, unable to repay the Loan … In the situation where the shareholder pays a reasonable fee to the corporation as consideration for the granting of the Guarantee, the Guarantee would not, in and by itself, give rise to a benefit.
— CRA Document 2006-0174011C6, June 29, 2006

There are two escape routes in that paragraph, and most people miss that the first one is useless to them.

Route	Condition	Available to an owner-manager?
1	Shareholder deals at arm's length with the corporation, and is demonstrably able to repay	No. You control the company. You are never at arm's length with it.
2	Shareholder pays a reasonable fee for the guarantee	Yes — and it is the only route available

If you own the company, Route 1 does not exist for you. Everything hangs on Route 2, and on the word reasonable.

The one case: Golini

There is exactly one Tax Court decision squarely on this arrangement, and the taxpayer lost.

In Golini v. The Queen, 2016 TCC 174, Paul Golini's holding company bought an offshore annuity and an offshore life insurance policy, pledged both as security, and Golini personally received a $6 million limited-recourse loan. He paid an annual guarantee fee of $40,000 — roughly 0.67% of the loan. The borrowed money circled back to subscribe for high-PUC shares, manufacturing both an interest deduction and a tax-free withdrawal channel.

The Tax Court assessed a shareholder benefit of $5.4 million — calculated, oddly, as total insurance premiums ($6M) less total paid and planned guarantee fees ($600k). Neither the loan amount nor "what a proper fee would have been" featured in the arithmetic. The Court also found several elements of the arrangement to be a sham and said that, if needed, the general anti-avoidance rule would have applied as an abuse of subsection 84(1).

The appeal to the Federal Court of Appeal was withdrawn. The Tax Court decision stands, untested by a higher court.

What Golini does — and does not — tell you

The single most important fact about Golini is the one least often mentioned: there was no bank.

The money reached Golini through a chain — offshore reinsurer → offshore company → Canco, an unrelated Canadian company → Golini personally. Canco was a conduit inside the promoter's own structure. It performed no credit underwriting, which is precisely why it was willing to lend on a limited-recourse basis. A real lender never is.

That is what the judge seized on. In his words, the reality was that Golini "would not have to repay the loan," and "everyone's understanding was the annuity and the insurance were the only manner" in which it would be repaid.

A domestic arrangement differs on every axis that mattered:

Golini	A domestic collateral loan
Limited-recourse loan	Full-recourse — the borrower's personal assets are exposed
Offshore insurer and offshore products	Canadian licensed insurer, OSFI-regulated
Death benefit engineered to track the loan balance	Policy values independent of the loan
Policy and loan created simultaneously	Policy typically predates the loan by many years
Money moved in a circle to manufacture PUC	Loan proceeds are simply spent

Two things follow, and they point in opposite directions:

1. The comfort: a conventional domestic structure is a materially different animal, and Golini is weak precedent against it.
2. The warning: the taxpayer in Golini paid a guarantee fee and was assessed anyway. A fee is not an immunity. And if a benefit is found, the quantum is genuinely unpredictable.

The CRA's own commentary since has been limited but pointed. At the 2018 CALU Roundtable it said it remains concerned with "planning arrangements similar to those undertaken in the Golini case" and will continue to act against domestic and cross-border schemes using life insurance to obtain benefits contrary to underlying tax policy (CRA Document 2018-0752971C6).

---

Part 2 — What should the guarantee fee actually cost?

The number quoted most often is 2% of the outstanding loan. Trace it to source and it looks less like a standard and more like the ceiling of the oldest available estimate.

Source	Year	Suggested range
Everett & Ireland, PPI Financial Group	2004	1 – 2%
Glenn Stephens, Estate Planning with Life Insurance, 4th ed., p.171	2008	1 – 1.5%
Glenn Stephens, Estate Planning with Life Insurance, 7th ed., p.271	2019	"usually in the range of 0.5%"
Actually paid in Golini	—	0.67% (and assessed anyway)

The author of the standard Canadian text on the subject cut his own recommendation by roughly two-thirds over eleven years, and added that clients should establish the prevailing acceptable rate at the time the loan is negotiated.

The CRA has never published a percentage

What it has published are two economic tests:

The value of a benefit arising from the right to use someone else's property as security for a loan is a question of fact. One method of calculating the fair market value of such a right might be to compare the difference between the interest rates charged with and without the corporation's collateral security. Another method might be to determine what the shareholder would have to pay a third party to provide a similar collateral security.
— CRA Document 2000-0002575, March 29, 2000

Neither test is a percentage of the loan. The percentage is an industry shortcut, not a rule.

The rate-differential test has judicial support in a different setting. In Canada v. General Electric Capital Canada Inc., 2010 FCA 344 — the leading Canadian guarantee-fee case — the Federal Court of Appeal endorsed the yield approach: price the guarantee at the spread between borrowing with it and borrowing without it. The Court also held that implicit support must be taken into account. Applied here, that cuts in the taxpayer's favour: a shareholder who could have borrowed on decent terms anyway has received something of modest value, and should pay a modest fee.

Sun Life's own guidance points the same way:

A shareholder with a decades-long stellar credit rating, wanting to borrow for a short period of time, could expect to pay a fee at the low end of the range … But a guarantee fee for a shareholder just emerging from their second bankruptcy, and wanting to borrow for a long period of time, could be at the high end.

The practical conclusion

Most people who can qualify for this strategy have strong personal credit and significant net worth — which is exactly the profile that justifies a fee at the low end. Overpaying is not a harmless conservatism: the excess lands in the corporation, gets taxed there, and is taxed again on the way back out.

A defensible position today: price the fee by the rate-differential method, land somewhere in the 0.5% – 1.25% range, put it in a written guarantee agreement, and keep the working papers. The low end has published authority behind it. What matters as much as the number is the evidence that you turned your mind to it — that evidence is what stands between a technical disagreement and a subsection 163(2) gross-negligence penalty.

---

Part 3 — The rule almost nobody mentions

The shareholder-benefit debate absorbs all the attention. The larger exposure is elsewhere.

Since March 22, 2016, the back-to-back shareholder loan rules in subsections 15(2.16) – 15(2.192) provide that where a corporation grants a specified right over property to a lender, and the lender's loan to the shareholder was only possible because of that right, the bank loan is recharacterized as a loan made directly by the corporation to the shareholder — with a full income inclusion under subsection 15(2).

A $1,000,000 loan becomes $1,000,000 of income in the year.

At the 2018 CALU Roundtable (CRA Document 2018-0745491C6) the CRA drew the line:

Drafting of the security	Specified right?
Bank may realize on the policy only to secure that shareholder debt	No
Bank may use the pledged property to raise capital for itself, or the assignment covers current and future loans including loans to the corporation	Yes

Canada Life's warning is blunt: the second description is standard language in most security agreements, and at the time of writing "some banks that offer collateral loans have loan documentation that creates a specified right under the new rules and are exploring potential solutions."

This is a pass/fail gate, not a risk factor. No one should discuss this strategy with a client before someone has read that specific lender's security agreement line by line.

---

Part 4 — The arithmetic

Everything above is legal risk. This part is whether the structure works economically at today's rates.

Assumptions

Every figure below rests on these. Substitute your own — the conclusions move.

Input	Base case	Note
Cash surrender value at start	$2,000,000	An established policy, not a new one
Policy cash value growth	5.0% / year	A participating dividend scale. Not guaranteed.
Loan interest rate	6.0% / year	Floating, roughly prime + 1.25%
Interest treatment	Capitalized	Not paid in cash; added to the balance
Guarantee fee	0.5% of the outstanding loan balance	Paid annually, funded from the draw
Lending limit	100% of CSV (whole life)	BMO Private Banking's stated margin for whole life; 90% for universal life
Personal marginal rate	48% (Alberta top bracket)
Non-eligible dividend rate	42.3% (Alberta top bracket)	Used for the "what tax did we avoid" comparison
Interest deductibility	None	The borrowed money funds lifestyle, not income-earning property

That last assumption matters and is often glossed over: if the money is spent on living expenses, the interest is not deductible, and neither, in all likelihood, is the guarantee fee. Interest deductibility requires the borrowed funds to be used to earn income from business or property.

New borrowing capacity, and what eats it

The appeal of the strategy is that the policy keeps growing, so new borrowing room keeps appearing:

New capacity each year = lending limit % × growth rate × cash value

At $2M growing 5%, with a 100% whole-life limit, that is $100,000 of fresh capacity in year one — and the figure compounds, reaching $171,034 by year 12 and $265,330 by year 21.

But the loan compounds too, and at a higher rate. What matters is the difference:

Year	New capacity	Interest that year	Net	Loan balance	Limit
1	$100,000	$0	+$100,000	$106,000	$2,100,000
6	$127,628	$36,233	+$91,395	$749,321	$2,680,191
12	$171,034	$97,918	+$73,116	$1,844,537	$3,591,713
21	$265,330	$247,733	+$17,597	$4,504,503	$5,571,925
24	$307,152	$319,855	−$12,703	$5,785,029	$6,450,200
28	$373,346	$439,989	−$66,643	$7,918,003	$7,840,258

Drawing $100,000 per year. The limit is breached in year 28.

The critical loan balance

There is a clean threshold worth committing to memory. Interest consumes all new capacity when:

L* = (lending limit % × growth rate × cash value) ÷ loan interest rate

At $2M, 5% growth, 6% interest, 100% limit: L* = $1,666,667. Past that balance, the arrangement is consuming headroom rather than creating it.

Loan rate	L* at 100% limit	L* at 90% limit
5.0%	$2,000,000	$1,800,000
6.0%	$1,666,667	$1,500,000
7.0%	$1,428,571	$1,285,714

The condition for a perpetual draw

A draw that can continue indefinitely satisfies:

Sustainable draw = (growth rate − loan rate − fee rate) × loan balance

Which means a perpetual draw exists only if the policy grows faster than the loan costs.

	g = 4.5%	g = 5.0%	g = 5.5%	g = 6.0%	g = 6.5%
r = 5.0%	−1.00	−0.50	−0.00	+0.50	+1.00
r = 6.0%	−2.00	−1.50	−1.00	−0.50	±0.00
r = 6.5%	−2.50	−2.00	−1.50	−1.00	−0.50

Net spread in percentage points, after a 0.5% guarantee fee. Positive = a perpetual draw exists.

At 5% growth against 6% borrowing, the spread is negative. There is no perpetual draw at today's rates — only a question of how many years the headroom lasts.

This is worth sitting with, because it explains something people find confusing about older illustrations. Between roughly 2010 and 2021, prime sat between 2.45% and 3.95%, collateral loans priced near 4%, and participating dividend scales ran above 6%. The spread was positive. The product has not changed. The spread flipped.

How long does it last?

Years until the loan breaches the lending limit, $2M starting cash value, 0.5% fee, 100% whole-life limit:

	g = 4.5%	g = 5.0%	g = 5.5%	g = 6.0%
r = 4.5%	18	21	27	60+
r = 5.0%	17	19	23	31
r = 6.0%	16	17	19	22
r = 7.0%	14	15	17	18

Drawing $150,000 per year.

And the maximum draw that survives a given horizon, at 6% borrowing and 5% growth:

Horizon	Maximum annual draw
25 years	$108,039
30 years	$93,906
35 years	$83,368
40 years	$75,054

A $2,000,000 policy supports something in the range of $80,000 – $95,000 a year for a working lifetime under these assumptions. It does not support $150,000 a year indefinitely — that draw runs 17 years and stops, typically at the exact moment the client is least able to absorb the news.

Does the fee rate change the outcome?

Only where the plan is otherwise sound:

Annual draw	Fee 0.5%	Fee 1.0%	Fee 1.5%	Fee 2.0%
$60,000	38 years	34	31	28
$100,000	22	20	19	18
$150,000	14	13	13	13

At the more conservative 90% limit and 4.5% growth, to isolate the fee's effect.

At a moderate draw, pricing the fee at 0.5% instead of 2% buys ten additional years. At an aggressive draw it changes almost nothing, because the plan was going to fail regardless. Fee pricing is a real lever — but only inside a plan that was sized correctly in the first place.

Paying the fee versus reporting the benefit

There is a counterintuitive result here that the insurers themselves acknowledge.

Route	Net cost per $1
Pay the guarantee fee	Shareholder pays $1.00. The corporation receives $1.00, is taxed on it as investment income, and returns roughly $0.48 after full distribution → net cost $0.52
Pay no fee and report the benefit	The corporation receives nothing. The shareholder includes $1.00 in income at 48% → net cost $0.48

Sun Life makes the same point in its own words: money the shareholder pays into the corporation "could not come back to the shareholder without tax consequences," while reporting the benefit leaves "less money in the corporation to be taxed upon removal at some future date." Canada Life goes further and states plainly that in some provinces taking the benefit may be more advantageous than paying the fee.

Which means the guarantee fee is not a tax-saving device. It is documentary evidence — proof that consideration was paid for the use of corporate property. Priced properly and papered properly, it is what keeps a technical disagreement from becoming a penalty. Treat it as an insurance premium against assessment risk, not as planning.

---

Part 5 — The strongest argument for the structure

Everything so far has been cost. Here is the offsetting benefit, and it is larger than most of the objections.

The realistic alternative is not "spend nothing." It is: leave the $2,000,000 inside the corporation and invest it in a securities portfolio. That triggers a rule the insurance comparison usually ignores.

Since 2019, adjusted aggregate investment income (AAII) above $50,000 grinds down a Canadian-controlled private corporation's small business deduction by $5 for every $1 of excess, eliminating the $500,000 small business limit entirely once AAII reaches $150,000.

In Alberta, small business income is taxed at 11% and general active business income at 23%. Losing the deduction therefore costs 12 percentage points on up to $500,000 of active income — $60,000 a year.

Assume the $2,000,000 portfolio returns 6.5%, of which 3.5% is currently taxable (interest, dividends, and the taxable half of realized gains):

Year	Portfolio	AAII	Small business limit lost	Extra corporate tax	Cumulative
1	$2,000,000	$70,000	$100,000	$12,000	$12,000
6	$2,536,383	$88,773	$193,867	$23,264	$104,723
12	$3,373,159	$118,061	$340,303	$40,836	$303,378
18	$4,485,994	$157,010	$500,000 — fully lost	$60,000	$622,749
30	$7,934,187	$277,697	fully lost	$60,000	$1,342,749

The grind bites in year one — $70,000 of AAII is already $20,000 over the threshold. It reaches full severity once the portfolio passes $4,285,714. Over thirty years the cost compounds to roughly $1.34 million in additional corporate tax on the operating business, entirely separate from the tax on the investment income itself.

Growth inside an exempt life insurance policy is not AAII. It is not taxed annually, it does not count toward the $50,000 threshold, and it does not erode the small business deduction. The grind cost is zero.

This reframes the whole comparison. The question is not "does a 5% dividend scale beat a 6.5% equity portfolio" — on that framing the policy loses. The question is:

5% clean, versus 6.5% net of annual investment tax and up to $60,000 a year of small-business-deduction erosion on the operating company.

On that framing the policy is competitive, and for a company with a full $500,000 of active income it is frequently ahead. This — not the tax-free loan — is the strategy's real engine.

One caveat that reinforces the loan structure specifically: if the corporation instead withdraws from the policy, the resulting policy gain is taxable and is AAII. Borrowing against the policy rather than withdrawing from it is what preserves the advantage.

---

Part 6 — The exit, which must be designed on day one

The most common failure point is not the tax analysis. It is the ending.

If the corporation's policy proceeds are used to pay off the shareholder's personal loan directly, the entire loan amount is a taxable shareholder benefit. This is not a grey area.

The sequence that works has to be agreed with the lender before the first advance:

1. The insurer is instructed not to pay the proceeds directly to the bank
2. The estate representative asks the lender to accept substitute collateral and release the policy
3. The proceeds are received tax-free by the corporation
4. The excess over the policy's adjusted cost basis credits the capital dividend account
5. The corporation pays a tax-free capital dividend to the estate
6. The estate repays the loan

Two related exposures round out the picture:

• Subsection 70(5.3) deems a corporate-owned policy on the shareholder's life to be worth its cash surrender value when valuing the shares. These are high-cash-value policies by design, so this inflates the final capital gain. Note that a loan taken by the corporation reduces share value and offsets this; a loan taken personally does not.
• If the lender realizes on the collateral after a default, three things happen at once: the corporation is taxed on the policy gain, the shareholder is assessed a benefit equal to the debt discharged, and the coverage is gone.

---

The verdict

The structure is legitimate, mainstream, and documented by every major insurer in the country. It is not, however, a standard product — it is a bespoke arrangement that fails badly when any one component is wrong. Five conditions have to hold together:

1. A full-recourse loan from an arm's length Canadian bank. This is the primary distinction from Golini, and it is not optional. No lender will accept an offshore policy as collateral in any event — Canadian programs are insurer-specific arrangements requiring the insurer to acknowledge the bank's priority lien.
2. Security documentation that does not create a specified right. Read the agreement. Pass or fail.
3. A policy that predates the loan by years and exists for a genuine estate or succession purpose, not a policy purchased in order to borrow against it.
4. A guarantee fee priced by the rate-differential method, documented in a written agreement — or a deliberate, calculated decision to report the benefit instead.
5. A repayment path agreed with the lender in writing at the outset, so the proceeds never touch the personal debt directly.

And the arithmetic imposes its own discipline: at current rates there is no perpetual draw. Size the withdrawal to a horizon, stress test it at loan rate +2% and dividend scale −0.5% simultaneously — the two move against you asymmetrically, because a floating loan rate reprices immediately while a dividend scale lags by years — and revisit the numbers annually rather than trusting a projection prepared at inception.

What this article does not tell you

Two things could not be verified from public sources and have to be confirmed case by case:

• Which Canadian lenders have amended their security documentation to avoid creating a specified right. Public material confirms the problem existed and that lenders were seeking solutions; it does not confirm who has solved it. Ask the specific lending desk for the specific agreement.
• The current market rate for a guarantee fee. The published range spans 0.5% to 2% — a fourfold spread — and no public benchmark exists. In practice this needs a valuator's opinion supported by written quotes from the lender with and without the corporate security.

---

This article is general information about Canadian tax rules, not tax or legal advice, and no part of it is a recommendation to enter into any particular arrangement. Rates and thresholds cited are Alberta figures and change. Anyone considering this structure needs their own CPA and tax counsel to review their specific facts and their specific lender's documents before proceeding.

Wallace Wang Financial Services · LLQP-licensed in five provinces