The Comparison Question

Paying cash from savings, taking a policy loan, borrowing from an outside lender against the policy, and using ordinary credit are four ways to pay for the same purchase. The usual case for the financing approach known as The Infinite Banking Concept® sets only two of them side by side, a policy loan against a loan from an outside lender, and counts the interest you would not pay that lender as a gain. That comparison is fair for one kind of household: the one that would really have borrowed. For a household that would have paid cash, the fair benchmark is its own savings. Measured that way, the advantage shrinks, and it can disappear.

The reason is simple. A policy loan is an advance from the insurer, made under the contract and secured by the policy's cash value. You owe the insurer, the interest is owed to and paid to the insurer, and the insurer sets the rate and may change it as the contract allows. Paying from savings costs no interest at all. Its cost is what that money would otherwise have earned after tax, and the cushion you no longer hold. Which route costs less depends on figures you can collect: the loan rate, the forgone earnings, the tax on each, and the terms of your own contract. The wider case against the approach, including the parts critics get right, is set out in the honest case against.

Two different questions travel under the same heading. One is how to pay for a purchase: cash, a policy loan, a collateral loan or ordinary credit. The other is where a yearly surplus should go: a permanent policy, term insurance with the difference invested, or something else. They have different answers, so keep them apart.

Who lends, who receives the interest, and who owes whom?

Four routes can pay for the same purchase. For each, settle the chain of money first: who advances it, who is paid the interest, and who owes whom afterwards.

The policy loan and the collateral loan are both described in sales material as borrowing against the policy, and they are different transactions. With a policy loan, the insurer advances the money under the contract, and the cash value stays in the policy as security; it is not withdrawn. The insurer sets the rate and may change it as the contract allows. Depending on the contract, you choose when to repay, and interest you do not pay can be added to the balance. With a collateral loan, an outside lender decides whether to lend, sets its rate and conditions, receives the interest, and takes an assignment of the policy as security. If the loan is still owing when the person insured dies, the lender, as assignee, can be paid from the death benefit before the beneficiary receives the rest.

The difference matters for tax. A policy loan is a disposition of an interest in the policy under the definition of "disposition" in subsection 148(9) of the Income Tax Act, paragraph (b). An assignment of the policy to secure a debt or a loan other than a policy loan is not a disposition (paragraph (f) of the same definition). The two are set side by side in more detail at a policy advance and other credit.

If you lend the money onward, to a child or another relative, count the debts. There are now two: you owe the insurer or the lender, and the relative owes you. The relative's debt does nothing to reduce yours, and if the relative does not repay, you still owe the insurer. How such a loan is documented is covered at capital held within a family.

What does each route really cost?

Each route has a price, and only part of the price is interest.

Paying cash costs the earnings that money would have produced, after the tax those earnings would have borne, until you rebuild the savings. It also costs the cushion: money spent is no longer there for an emergency. Nothing is owed afterwards.

Ordinary credit costs the lender's interest and any fees, on the lender's schedule, while your savings keep earning. Unless your savings earn more after tax than the loan costs after tax, borrowing costs more than paying cash.

A policy loan costs the insurer's interest, at a rate the insurer sets and may change. While the loan is outstanding, the balance and accrued interest are deducted from the death benefit and from any surrender value. Depending on the contract, the dividends on the part of the cash value securing the loan may be adjusted. The loan also uses up room, so a later loan has less value behind it.

A collateral loan costs the lender's interest and conditions, and it gives the lender a right to be paid from the policy if you default or die owing.

Illustrative example. The figures are assumptions chosen to show the arithmetic, not rates from any insurer, lender or account. You need $30,000 for one year. You hold $30,000 in a non-registered savings account and also a participating policy with enough cash value to secure the loan. Your savings would earn 3% for the year, taxed at an assumed marginal rate of 40%. A policy loan is charged 6% for the year and a line of credit 7%, both as simple interest and repaid in full at the end of the year. The purchase is personal, so none of the interest is deductible.

• Pay cash. You give up $900 of interest ($30,000 × 3%), which is $540 after tax ($900 × 60%). You owe nobody.
• Policy loan. You pay $1,800 of interest to the insurer ($30,000 × 6%), while your savings earn their $540 after tax. The loan route costs $1,260 more than paying cash ($1,800 less $540).
• Line of credit. You pay $2,100 of interest to the lender ($30,000 × 7%): $300 more than the policy loan and $1,560 more than paying cash.

Notice what drops out. If your contract does not adjust dividends for loans, the policy grows the same way whether you pay cash or take the policy loan, because in both cases the cash value stays in the policy. So the policy's growth cannot make the policy loan cheaper than paying from savings you already hold. In the most favourable case, under such a contract, it means the loan does not slow that growth. If your contract does adjust dividends for loans, the adjustment is part of the loan's cost, in whichever direction it runs. Real contracts and lenders also charge interest their own way, by the day or compounded at an anniversary, so ask each for its method.

When is the outside-lender comparison the fair one?

When you would genuinely have borrowed. A business that has always financed its equipment cycle, a purchase too large to pay from cash, a household that has in fact carried a vehicle loan or a credit line for years: in each case the real alternative is a lender, and comparing one source of credit with another is the right analysis, not a trick.

The test is your own record. If you have never borrowed for this kind of purchase, do not accept an argument that assumes you would have.

When you would have borrowed, compare the whole of each offer, not only the rate. On the lender's side, that means its approval and conditions, whether it can change the rate or call the loan, the repayment schedule, and what happens if you cannot pay. On the policy side, it means the insurer's rate and how it can change, the loan limit the contract sets, what unpaid interest does to the balance, and what the balance does to the death benefit.

Tax enters when the money earns income. Interest on money borrowed to earn income from a business or property can be deductible under paragraph 20(1)(c) of the Income Tax Act. Money borrowed to acquire a life insurance policy is excluded, and interest on money used for personal spending is not deductible. To claim policy-loan interest, the Canada Revenue Agency asks you to have the insurer complete Form T2210 (CRA, line 22100); Quebec residents also use Revenu Québec's Form TP-163.1-V. Policy-loan interest you pay in cash and do not deduct counts as a premium under paragraph (a) of the definition of "premium" in subsection 148(9), so it raises the policy's adjusted cost basis. That is our reading of the provisions, not a ruling, and your accountant confirms which applies to you.

What does a policy loan do to the policy, and to your tax?

The cash value is not paid out when you borrow. It stays in the policy as security, and the loan sits beside it as a debt to the insurer. Three things follow.

The death benefit shrinks while the balance stands. When the person insured dies, the insurer deducts the balance and the accrued interest from the amount paid to the beneficiary.

Repayment is flexible, and the debt is real. Depending on the contract, there may be no fixed repayment schedule, and interest you do not pay can be added to the loan, where it bears interest in turn. The Financial Consumer Agency of Canada warns that a loan against the cash value has to be repaid, or what your beneficiaries receive can be reduced (FCAC, life insurance). Both statements hold together: the contract may let you choose when to repay, but the balance is settled out of the policy's values if you never do.

The loan can end the policy. If the balance and accrued interest grow until they overtake the value securing them, the contract can terminate under its terms, and the coverage ends with it. The contract sets that point and any notice the insurer gives. Ask for the loan provision in writing, and when the margin is narrow, check it more often than the annual statement allows.

On tax, a policy loan is a disposition, but it produces income only to the extent the loan's proceeds exceed the policy's adjusted cost basis immediately before the loan (subsection 148(1)). The proceeds are generally the amount advanced, except any part the insurer applies directly to a premium. The loan lowers the adjusted cost basis. If you later repay a loan that was partly included in income, you can deduct the repayment in the year you make it, up to the amount previously included (paragraph 60(s)). That is a deduction in the year of repayment, not a refund of the earlier tax, and it does not arise without repayment. The adjusted cost basis can fall over time, for example as the net cost of pure insurance is subtracted each year, so a loan that creates no income in one year can create some in a later one. Ask the insurer for the figure before each loan. The federal rules apply everywhere in Canada, and Quebec residents also report to Revenu Québec.

If the policy ends with a loan outstanding, by surrender or by lapse, that is also a disposition. For tax, the proceeds are the cash surrender value less the policy loans owing, so the small cheque, or no cheque at all, is not the figure that decides the result. Income arises only to the extent those proceeds exceed the adjusted cost basis, and it can arrive in a year with no cash to pay the tax. A lapse caused by unpaid premiums is not a disposition if the policy is reinstated no later than 60 days after the end of the calendar year of the lapse (paragraph (g) of the definition of "disposition"). That exception is written for premium lapses. If a contract ends because the loan overtook its value, ask the insurer and your accountant in writing whether it can apply, and do not assume it does. Reinstatement itself is the insurer's decision under the contract. Before letting a policy with a loan end, ask the insurer in writing for the disposition proceeds, the loan settlement, the adjusted cost basis and the tax slip it expects to issue, and have an accountant review them even if little money changes hands. This is our reading of section 148, not a ruling, and the full mechanics are at when a policy loan becomes taxable.

What does a policy offer that savings do not, and at what price?

A participating policy is insurance, not a savings or investment account. Some things do survive the concession, though. They are narrower than the usual claim, and each comes with a condition.

Dividends and loans. A participating policy does not credit interest at a stated rate. The insurer's board declares dividends, which are not guaranteed, based on the participating account's results as the insurer's dividend policy describes. Some contracts do not adjust the dividend for an outstanding loan; others adjust it on the part of the cash value securing the loan, up or down. That is a feature of a specific contract, not a rule. Ask for the insurer's dividend policy and the loan provision for your contract, in writing, before you rely on either.

The death benefit. A permanent policy pays a death benefit when the person insured dies, while the policy remains in force, under its terms, less any loan and interest owing. A savings account pays no such sum. That belongs in any fair comparison, and it cuts both ways: the death benefit is a large part of what the premium pays for, so a household that does not want permanent coverage is paying for something it does not want.

Tax on growth. Growth inside a policy that stays exempt under Regulation 306 of the Income Tax Regulations is not taxed each year the way interest in a non-registered account is. The conditions are real. The policy must stay exempt; a policy that ceases to be exempt is deemed to have been disposed of (paragraph 148(2)(d)); and loans, surrenders and lapses carry the tax consequences described above. The exempt test explains the limit.

Forced saving, not forced repayment. The contract enforces the premium, not the loan. If a premium goes unpaid, the contract responds by its terms: depending on the contract, dividends may pay it, an automatic premium loan may pay it from the cash value, the policy may be converted to reduced paid-up insurance, or it may lapse. Loan repayment is the opposite. Depending on the contract, nobody sets a schedule for you, and unpaid interest can be added to the loan. If you would not keep up a voluntary repayment plan on a line of credit, a policy loan will not make you. That is a point about habits, and it is fair to weigh it only against your own.

The price of all this. Early cash value sits well below the premiums paid, because the cost of putting the policy in force falls heaviest in the first years. The premium is a long commitment, and leaving early returns less than you paid. The policy also needs underwriting before it exists, and the insurer can decline, rate or postpone an application. Each cost is set out at the real costs.

The corporate case. For an incorporated owner, the arithmetic changes. Corporate surplus is taxed while it is held in the company. When a private corporation receives life insurance proceeds because of a death, its capital dividend account is credited with the proceeds less the policy's adjusted cost basis immediately before the death (subsection 89(1), definition of "capital dividend account", paragraph (d)). A capital dividend is paid out of that account only with the election under subsection 83(2). Who owns the policy, who is the beneficiary, and any loan or collateral assignment in place can change the result, so the corporation's accountant calculates it before anyone relies on it. The payment itself is described at paying a capital dividend after a death.

What has to be in place before a policy loan is possible?

A policy loan needs a policy with value in it, and that takes time and meets conditions along the way:

• Insurability. The insurer underwrites the application and can accept it, rate it, postpone it or decline it. Having surplus money does not show that you can be insured.
• Premiums paid on schedule. The premium commitment runs for years, and the design chosen at issue sets it.
• Cash value. The contract sets the maximum loan, based on the cash value, and early cash value is low against the premiums paid.
• The loan provision. The contract sets how interest is charged and added, and when an unpaid balance can end the policy. The insurer may require signatures, for example where a beneficiary designation is irrevocable.
• The adjusted cost basis. Ask the insurer for it before each loan, because it decides how much of the loan could be income.
• Separate reserves. An emergency fund held in cash does a job that early cash value cannot.

The request itself, and what the insurer needs from you, is covered at how policy loans work.

Does buying term and investing the difference win?

The second question readers bring is a different comparison, and it deserves a straight answer. Term life insurance costs much less than permanent coverage for the same death benefit. So buy term, invest the difference in a low-cost portfolio, and after decades you hold more, with full access and visible fees.

It can be right. Where the need for coverage is temporary, such as a mortgage or children who will become independent, term does the insurance job for a fraction of the cost. On growth alone, a portfolio can end with more than a policy's cash value. Whether it does depends on the premiums, the returns after fees and tax, how much you actually contribute, and the dates chosen. A presentation that says a policy wins as a growth vehicle has told you something it cannot support.

It is weaker than it sounds where its assumptions are. It assumes the difference is invested every year and left invested through bad markets. Test that against your own record: if you have invested surplus steadily for years, buy-term-and-invest is the benchmark to beat; if your surplus has tended to be spent, compare against what you actually did. Test the policy side just as strictly, because a policy funded without a gap for decades is also an assumption.

It also assumes the term coverage does what you need when it ends. Term coverage ends at the end of its term unless the contract lets you renew it, at a higher premium, or convert it to permanent coverage. Some contracts allow conversion without new medical evidence up to an age or a date the contract sets. The FCAC notes that you might be able to renew certain term policies and that premiums may rise on renewal. Read the renewal and conversion clauses before assuming either way; term life insurance and converting term life insurance explain both. If you do not convert and the need turns out to be permanent, replacing coverage at an older age can be costly or unavailable, as risks and failure modes explains.

A permanent need does not mean participating whole life. Term to 100 and non-participating permanent coverage are designs without a participating account, at a different cost, and participating and non-participating whole life compares them. The fair comparison for a permanent need is between permanent designs.

On risk, a portfolio's value on the day you draw from it depends on the market that day. A policy's guaranteed values follow a contractual schedule and do not. Its non-guaranteed values depend on future dividends, and the loan rate can change, so what you can draw or leave at death still depends on premiums, dividends, loans and interest. The policy is smoother, not immune.

For some households the answer is both: term coverage for the temporary need and a smaller permanent policy for the lifelong one, in the shape described at whole life insurance.

What must a fair comparison hold constant?

Six conditions. A comparison missing any one of them leans in some direction.

1. The same period. Any two products can be made to win by choosing when to start and when to stop.
2. The same fee treatment. After fees against after fees, or before against before. Mixing them produces a gap in whichever direction was chosen.
3. The same certainty. Guaranteed against guaranteed, projected against projected. A contractual floor and a projected average are different quantities.
4. The same tax treatment and account type. Money in a non-registered account, a TFSA, an RRSP and a policy is taxed differently, so a comparison must say which it uses. These do different jobs from life insurance. This practice is licensed to place insurance of persons and is not registered to advise on registered plans, so it gives no ordering between them and a policy; take registered-plan questions to a professional licensed for them.
5. The same cash outlay. Each side puts in the same dollars in the same years, or the difference in outlay is shown.
6. Everything each provides. A death benefit while the policy is in force, less amounts owing; access to money in years one to five; the cost of leaving early.

Applied fairly, these conditions can produce a mixed result rather than a clean one: one product costs less as a way to grow money, and the other provides something the first does not. A comparison that shows a clean win for whatever the presenter sells has either failed one of the six or compared different things. Ask which.

Which comparisons are built wrong?

Five constructions appear on both sides of the argument. Each one looks like method:

• Setting a policy's cash value against a portfolio's full value. One includes the cost of a death benefit. The other provides none.
• Setting gross investment returns against net policy values, or the reverse. The direction depends on who is presenting. The error does not.
• Quoting an internal rate of return on a policy as if it were an investment return. It includes the price of coverage.
• Setting the insurer's dividend scale interest rate beside a loan rate or a savings rate. That rate is one input to the scale. It is not the policy's return. A policy's growth also reflects the cost of insurance, expenses and the contract's design.
• Choosing the period after the fact. The comparison then starts in a market trough or ends at a peak.

None of these needs bad faith, and each produces a conclusion the facts do not support.

What should you compare against: your own record?

Not a market index, which needs assumptions nobody can defend, and not a disciplined ideal. Compare against your own recent history:

• What did you do with surplus money over the last five years?
• What did you finance, and at what rate?
• What did you spend that you meant to save?
• In a tight year, what did you stop paying first?

A household that invested its surplus steadily should compare against that. A household that spent it is comparing against spending, and the result changes. Neither answer is a judgement. It is the input that decides which alternative is real for you, and no presentation can supply it.

The same test applies to the policy side. A policy funded without interruption for decades, at a level a strong year supported, is also an idealised path. A household that stops funding in year four has not reached the illustrated outcome either.

What should an illustration and a lender's quote show before you compare?

A projection built on assumed dividends, rates and tax proves whatever its assumptions chose. A dated insurer illustration and a lender's written quote are different: they are documents you can check line by line. Ask for these, in writing, before comparing anything.

Then put four questions to the advisor, before any application:

1. Who should not buy this, and could I be one of them?
2. What does the guaranteed column show at year three?
3. What happens if I can pay for five years and then cannot?
4. How are you paid on this policy, and when?

Reading is free. Canadian Wealth Creation Centre Inc., the firm behind this site, is paid by insurer commission if a policy is bought through it, and any advisor you consult should tell you plainly how they are paid.

Who is this comparison likely to go against?

A policy loan is hard to justify on cost when:

• you would have paid cash, and you have no need for permanent coverage;
• the money may be needed in the first years, when cash value is low;
• the funding depends on a good year rather than an ordinary one;
• high-rate debt is outstanding;
• you would not keep up a voluntary repayment plan.

Being on that list is not a judgement of anyone. Other tools suit that household better today, and circumstances change. The case is stronger for a household with a documented need for permanent coverage, durable surplus, separate reserves, a record of repaying what it borrows, and purchases it would really have financed.

What does the comparison come down to?

Two statements, both true. Permanent insurance is an expensive way to grow money. It provides something a portfolio does not. A comparison that offers only one of them has chosen its conclusion before it began.

Hold any comparison you are shown to the six conditions, ours included. If one of ours fails them, write to us; the contact details are at the foot of every page, and we would rather hear about a failure than leave it standing. For the wider case against the approach and what it gets right, start at the honest case against. For the contract itself, see how a participating policy works.