Is the Smith Manoeuvre legal in Canada, and what actually breaks it

The question people arrive asking is rarely the question that decides their file. The arrangement is documented, it is old, and its tax treatment comes from the same rules any investment loan is measured against. The thing that goes wrong is downstream of all of that, it has a name here, and almost nobody arrives knowing the word for it.

Contamination is what happens when borrowed money that was tracked to an income-producing use becomes mixed with personal funds or personal spending. The borrowed dollars can no longer be traced to the use that made their interest deductible, so the deduction is at risk for the contaminated portion of the balance.

What the Supreme Court has and has not decided

Three judgments do the work, and it is worth being exact about which part of each one applies.

Singleton v. Canada held that a sequence of transactions is read step by step rather than collapsed into one — a lawyer withdrew capital from his firm to buy a house, then borrowed to replace the capital, and the borrowing was measured on its own direct use. Reordering your own affairs to meet a test in the Act is not, by itself, the problem.

Ludco Enterprises Ltd. v. Canada held that the income the purpose test looks for is income generally, not net profit, and that courts should not weigh the sufficiency of the income expected. A modest distribution against a larger interest cost still meets the purpose test.

Lipson v. Canada is the one usually cited as the cautionary tale, and usually mis-stated. The couple there ran a conversion of much this shape and then bolted a spousal share transfer onto it so the interest landed on the higher-income spouse’s return through the attribution rule. The majority accepted the interest claim itself, on Singleton, and applied the general anti-avoidance rule to the attribution step. The lesson is narrower than “the CRA came for a Smith Manoeuvre”: an added step whose only purpose is to move a deduction to a better return is a different question from the conversion underneath it.

The failure that is actually common

Not a court case. A convenience.

The CRA’s position is set out at ¶1.43 of Income Tax Folio S3-F6-C1, Interest Deductibility: where a single borrowing account is used for eligible and ineligible purposes, a repayment of principal reduces both portions. You cannot direct it at the personal half. The folio’s own worked example, at the same paragraph:

Two things follow from that arithmetic, and they are the whole reason this page exists. The percentage is sticky: it survives every repayment, so the mixed line stays mixed until it is repaid in full and closed. And a claim now rests on a ratio rather than on a list of draws — which is a proportion someone has to compute, evidence, and be able to re-compute in year fourteen.

Elsewhere the folio is generous about mixing. Where borrowed money and other cash sit commingled in one account, ¶1.42 lets a taxpayer choose which uses the borrowed money funded, subject to timing — a use can never be linked to a borrowing that came after it. That flexibility is on the spending side. It does not extend to repayments, which is the asymmetry most spreadsheets are built without.

The five events that mix a balance

Each one is ordinary, and none of them announces itself:

1. A personal draw on the investment line — a vehicle, a renovation, a tax bill.
2. Interest capitalised in a month when cash was short, so the balance stops equalling the sum of the investment draws, and the accumulated part moves to a cash-basis paragraph besides.
3. A draw that paused in a chequing account long enough to be mixed with salary before it was invested, leaving no document tying the borrowed dollar to what it bought.
4. A return of capital reinvested without being recorded as what it was, quietly reducing the amount still standing behind an income-producing use.
5. A sale whose proceeds went somewhere else — spent, or parked, rather than into a replacement holding the borrowing could follow.

The first is the one that changes the shape of the file. The other four change the amount, and the tests they are measured against set out how much.

What can be repaired, and what cannot

The honest answer has two halves, and the second half is the one worth reading.

Going forward, structure is repairable. Folio ¶1.33 says plainly that a taxpayer may restructure borrowings and the ownership of assets to meet the direct-use test, and it works an example through: sell, repay the mixed borrowing, and re-borrow into a clean facility for a qualifying purchase. A second sub-account with a zero balance is the ordinary version of the same move. Nothing about a mixed line prevents a clean one being opened beside it tomorrow.

Backwards, the record is the record. A year that was paid for out of one account is a year whose allocation is a reconstruction, and reconstruction is where tracing gets hard — what a spreadsheet cannot evidence is exactly the material a reviewer asks for. That is a statement about the state of a record, not about anyone’s return, and the distinction matters: what a document says is a fact, and what it means for a particular filing is a professional’s judgment.