RRSP Gross-Up Loan

Objective — contribute the fully grossed-up amount to an RRSP in one step, bridging the difference with a short-term loan that the contribution’s own refund repays within weeks.

Risk level — 1 — Low risk. Risk justification — level 1 on the structural test: measured against its correct counterfactual — the gross-up refund, which reaches the identical end state with no loan at all — this re-arranges the timing of the holder’s own cash and adds no net market exposure, with a downside bounded and identifiable in advance (a few weeks’ interest, plus the risk the refund is smaller than computed). Level 2’s own test fails on its third clause: the outcome does not depend on returns exceeding an after-tax borrowing cost over a long horizon, because the borrowing cost is weeks of interest on a fraction of the contribution, not a hurdle rate to clear for a decade.

⚠ This record is the resolution of the test named in SCHEMA §2.3, and it came back positive. A strategy whose mechanism is borrow money and put it in the market sits at level 1 and is therefore not caught by the risk-level >= 2 compliance filter that SDC/Risks/JOB_DESCRIPTION.md keys on. The filter needs a second criterion. Until that decision lands — Risks first, then the CEO — this record must be routed to Risks by hand and must not be published externally.

Benefit justification — unverified — needs Talbot. Jurisdiction — Canada. No U.S. equivalent — the RRSP is Canada-specific. Decline type — all four, and this is the point of it. The strategy is indifferent to the rate environment, so it survives a Type B (inflationary) decline in which the entire borrower rate channel fails. Almost nothing else on the debt side of this library does.

Prerequisites

Unused RRSP contribution room, confirmed from the CRA notice of assessment.
Taxable income high enough at the stated marginal rate to actually produce the refund. The arithmetic below assumes the full deduction is usable; if it is not, the loan is not fully extinguished.
A short-term loan facility, or any other bridge that survives until the refund arrives.

Mechanism

Let C be the after-tax cash the holder has available for the contribution and t their marginal rate.


Contribute	`X = C ÷ (1 − t)`
Own cash used	`X × (1 − t) = C`
Short-term loan	`t × X`
Refund generated	`t × X` — exactly repays the loan

At t = 40% and C = $6,000: contribute $10,000, borrow $4,000, receive a $4,000 refund, repay the loan. The holder’s own outlay is the $6,000 they started with, and $10,000 is compounding tax-deferred instead of $6,000.

The formula is the source’s own. contribution = after-tax cashflow ÷ (1 − tax_rate) is LevPro’s gross-up refund, strategy 3 of the five refund strategies, implemented at sd_math/strategies/rrsp.py:50. The multiplier is 1 ÷ (1 − t) — 25% more contributed at a 20% marginal rate, 100% more at 50% — and that range is confirmed independently by the source’s own Harmful list, which prices spending the refund as a “guaranteed 25–100% reduction in retirement savings — the inverse of the Gross Up.”

What the loan does, and what it does not. It does not create the extra contribution; the refund does. It compresses into one step what otherwise takes two — and for someone who can reduce tax withheld at source instead, no loan is needed at all. The loan exists for the holder who cannot, or who is contributing after the withholding was already set.

Why the counterfactual matters for the tier. Against a spend-the-refund baseline the strategy plainly adds market exposure. But the source treats spending the refund as the worst of the five options, and market-drop-wins-library-v2 puts it on the Harmful list outright. A tier assigned against a baseline the library itself calls harmful is a tier assigned against the wrong thing.

Benefits

The full grossed-up amount compounds tax-deferred from day one.
The loan is extinguished by an arithmetic certainty rather than by an investment return — which is what distinguishes it from rrsp-catch-up-loan, where the whole thesis is that returns beat the loan rate over a decade.
Rate-environment indifferent. It works in the decline type that closes every other borrower strategy in this library.
The source’s stated reason for calling it the single best bridge from a market-drop conversation to a debt one: “it introduces borrowing-to-invest through the safest possible door.”

Risks

The refund is computed, not guaranteed. It depends on the marginal rate actually applying, on the deduction being usable in the year, and on other credits and deductions. A refund smaller than computed leaves a loan balance that the mechanism does not repay.
Timing. The loan must survive from contribution to refund — weeks, sometimes months. A delayed assessment extends the carry.
Market risk on the contribution, which is larger than it would otherwise be. This risk is identical under the loan-free gross-up refund, which is why it does not raise the tier — but it is real, and it is why the strategy is not level 0.
Over-contribution. Contributing beyond available room triggers a penalty; the gross-up multiplies the contribution, so it also multiplies this error.

Failure modes

Spending the refund instead of repaying the loan. This converts the strategy into its own inverse, which the source lists as Harmful. It is the single failure that matters.
Assuming the interest is deductible. It is not — see Tax considerations.
Applying the gross-up formula to a marginal rate the holder’s income does not actually support.
Running it without confirming room first.

Costs

Short-term interest on t × X for the weeks until the refund arrives, and it is not deductible. Magnitude unverified — needs Talbot.

Tax considerations

The contribution is deducted from taxable income; growth is tax-deferred; withdrawals are taxed as income.
Interest is not deductible. CRA’s position under ITA §20(1)(c) is unqualified: money borrowed to contribute to an RRSP is not used to earn income from business or property. Folio S3-F6-C1 and the former IT-533 confirm it. Because the loan here lasts weeks rather than years, the non-deductibility costs far less than it does in rrsp-catch-up-loan — but it is the same rule.
Contribution room is 18% of prior-year earned income to the annual maximum, less the pension adjustment.

Who it may suit

A Canadian with confirmed room, a marginal rate high enough for the gross-up to be worth the step, and income that reliably produces the refund. The benefit scales with the marginal rate, so it is worth most to exactly the people for whom 100% more contributed is available.

Who should avoid it

Anyone without confirmed contribution room.
Anyone whose taxable income will not support the full deduction at the assumed rate.
Anyone who would spend the refund. For that holder this strategy is not neutral — it leaves them with a loan.
Anyone in a low bracket now who expects a materially higher one later, for whom deferring the deduction is worth more than grossing it up today.

Implementation outline

Confirm available room from the CRA notice of assessment.
First check whether reducing tax withheld at source achieves the same thing with no loan. If it does, stop here and use that.
Compute X = available after-tax cash ÷ (1 − marginal rate), and verify taxable income supports the deduction at that rate.
Arrange the short-term loan for t × X, sized to the computed refund and no more.
Contribute X, file promptly, and apply the refund in full to the loan the day it arrives.

Evidence status

documented — the gross-up formula is implemented and verified in sd-math as part of the rrsp-catch-up-loan comparison, but the loan-bridged version is explicitly not modelled: LevPro’s own text calls the top-up loan “generally accepted” and sd-math/docs/design.md records it as out of scope for the catch-up comparison. So the arithmetic is verified and the strategy is not.

The strategy is attributed to Talbot in market-drop-wins-library-v2’s Smart Debt Bridge. That attribution is a claim this library records rather than verifies.

⚠ The authoritative source for this strategy is missing. return-to-office-plan-v1 §5 points at “the Gross-up document” for full state. No such document is in the vault — searched vault-wide 2026-09-22, nothing matches. Everything above is reconstructed from the gross-up formula in sd-math plus Market Drop Wins’ one-paragraph description. This is the blocking gap, and it is the reason evidence-status is documented rather than anything stronger.

Counterarguments

“The source calls this Can’t Lose; this record calls it level 1.” Both are right about different scales. Market Drop Wins’ Tier 0 measures availability and certainty of mechanism; SCHEMA §2.2’s risk-level measures magnitude of risk to the holder. The mechanism here is certain — the refund arithmetic is arithmetic. The contribution is still at market risk, which is what level 1 records. The two scales do not map by formula, and this is the clearest case of it in the fold-in.
“If the same end state is reachable without a loan, why borrow at all?” Conceded, and the implementation outline says to check that first. The loan is a convenience for a holder whose withholding is already set, not an improvement on the loan-free route.

Variants

The name is a known collision and is provisional. Talbot’s own corpus uses gross-up refund for the loan-free variant, and top-up loan for the loan. market-drop-wins-library-v2 uses RRSP Gross-up for the loan version. “RRSP Gross-Up Loan” is this record’s provisional title, assigned to disambiguate; confirming or replacing it is a CEO decision (SCHEMA §1.1). Both prior labels stay here as redirects.
The five refund strategies from the source: spend refund · reinvest refund · gross-up refund (this record’s loan-free twin) · top-up loan (this record’s mechanism) · catch-up loan (rrsp-catch-up-loan).
A withholding-reduction version, which reaches the same place with no borrowing.

rrsp-catch-up-loan · borrow-to-fund-tfsa · hbp-llp-early-repayment · interest-only-investment-loan · claim-investment-interest-deduction

Sources

market-drop-wins-library-v2 — the Smart Debt Bridge table (“the single best bridge”, rate-environment indifference, attribution to Talbot); and the Harmful list entry “spending the RRSP refund — guaranteed 25–100% reduction in retirement savings — the inverse of the Gross Up”
~/projects/monorepo/packages/sd-math/docs/design.md — the five refund strategies; gross-up refund defined as after-tax cashflow ÷ (1 − tax_rate) with no loan; top-up loan defined and recorded as not modelled
~/projects/monorepo/packages/sd-math/src/sd_math/strategies/rrsp.py:50 — the implemented gross-up formula
Talbot Stevens, Dispelling the Myths of Borrowing to Invest — “Borrowing for RRSPs” (RRSP-Strats.docx), the origin of the five refund strategies
Core/_WorkingOn/Research/canada-investment-taxation.md §6.5 — interest on RRSP borrowing not deductible; §11.1 — contribution room

Open questions

Where is the Gross-up document? return-to-office-plan-v1 §5 names it as this strategy’s full-state source and it is not in the vault. Blocking — everything here is reconstructed without it.
The name. CEO decision, and it blocks publication.
Risks review of the level-1 assignment and of the compliance-filter gap it exposes. Blocking under the IP charter’s inherited gate.
Whether the loan-bridged version should be modelled in sd-math alongside the catch-up loan, which would close the last gap between “arithmetic verified” and “strategy verified”.
Whether a refund smaller than computed should be handled as a variant or as a prerequisite check.