RRSP Catch-Up Loan

Objective — borrow a lump sum to fill unused RRSP contribution room immediately, then repay the loan over several years, so the full amount compounds tax-deferred from day one.

Risk level — 2 — Moderate risk. Risk justification — borrowed money is placed at market risk, so level 2 at minimum. It is not level 3: the loan is not callable and the position sits inside a registered plan rather than a margin facility. Benefit justification — unverified — needs Talbot. Jurisdiction — Canada. No U.S. equivalent — RRSPs are Canada-specific, stated as such in the audit.

Prerequisites

• Unused RRSP contribution room. The implemented model assumes the loan does not exceed available room — partial-refund cases are explicitly not modelled.
• Capacity to make the annual after-tax loan payment for the full amortization.
• A marginal tax rate high enough for the refund to matter.

Mechanism

Verified against Talbot’s own published worked example (“Bob”, from Dispelling the Myths of Borrowing to Invest):

1. Borrow the unused room — say $20,000 at a 40% marginal rate.
2. The contribution generates a refund: 20,000 × 0.40 = $8,000.
3. The refund is applied to the loan, leaving $12,000 to repay.
4. That balance amortizes over the term on a beginning-of-period (annuity-due) basis — the convention that reproduces the source’s stated $1,656/yr payment, where ordinary end-of-period amortization gives $1,788.24.

The engine compares the catch-up loan against the three no-loan ways of handling an RRSP refund — spend, reinvest (contribution × (1 + tax rate)), and gross-up (contribution ÷ (1 − tax rate)) — across a range of returns.

Benefits

• The whole amount compounds tax-sheltered from year one instead of being contributed in instalments.
• Verified comparison, not a sales claim: at 12% return over 10 years the source’s own table gives Catch-Up $62,120 against Gross-Up Refund $54,240 and Spend Refund $32,550.
• The honest sanity check is in the same table: when the RRSP return equals the loan rate (8%), Catch-Up and Gross-Up Refund are exactly equal at $43,180. The strategy only wins if the return beats the borrowing rate — and the model asserts this identity as a test.

Risks

• The 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, so the entire interest cost is after-tax. Folio S3-F6-C1 and the former IT-533 confirm it, with no advance ruling departing from it.
• Market risk inside the plan, with the loan payable regardless.
• Withdrawals to service the loan are taxed as income and the contribution room is lost permanently.

Failure modes

• Assuming the interest is deductible. The single most common misconception in this strategy.
• Borrowing more than available contribution room — not modelled, and creates an over-contribution penalty problem.
• Spending the refund instead of applying it to the loan, which breaks the arithmetic above.

Costs

Non-deductible interest for the full term. Compare against gross-up refund, which reaches a similar place with no loan at all when returns are modest.

Tax considerations

Contribution deducted from taxable income; growth tax-deferred; withdrawals taxed as income. Interest not deductible (§6.5). Contribution room is 18% of prior-year earned income to a maximum of $31,560 (2024) / $32,490 (2025), less the pension adjustment.

Who it may suit

A Canadian with material unused room, a high current marginal rate, stable income to repay, and an expectation that the plan’s return exceeds the loan rate.

Who should avoid it

• Anyone expecting a return at or below the loan rate — at equality the gross-up refund does as well with no debt.
• Anyone in a low marginal bracket now who expects a higher one later; the deduction is worth less today.
• Anyone who would need to withdraw to service the loan.

Implementation outline

1. Confirm available contribution room from the CRA notice of assessment.
2. Size the loan to room, not to appetite.
3. Apply the refund directly to the loan.
4. Compare against the gross-up refund alternative at a realistic return before deciding.
5. Amortize on a schedule that survives an income interruption.

Evidence status

implemented-and-verified — sd-math/strategies/rrsp.py::calculate_catch_up, verified against all 16 values of the author’s own published worked example (tests/test_rrsp.py, tolerance ±$6 against the source’s nearest-$10 rounding).

One provenance caveat, carried openly: unlike every other implemented strategy, this is not a VB6 line-trace. LevPro’s modMath.bas preserved only the mechanical shape (four compound-growth projections) — the refund and gross-up business logic lived in a UI form file that was not preserved, confirmed by exhaustive grep across all three .bas files. Talbot supplied the authoritative source instead: his own booklet chapter. That is an author-supplied derivation, not an inference from PDFs — but it is a different provenance class from the other two.

Not implemented: the Top-Up Loan (refund strategy 4, out of scope per the source’s own framing), partial-refund cases where the loan exceeds room, and combining this with the taxable-account leverage strategies.

Counterarguments

• Non-deductible interest to fund a tax-deferred account is a worse tax structure than borrowing to invest in a non-registered account. Structurally true; the offsetting argument is the shelter and the deduction on contribution. The model exists so the comparison is computed rather than argued.

Variants

The five refund strategies from the source: spend refund · reinvest refund · gross-up refund · top-up loan · catch-up loan (this record).

The top-up loan now has its own record — decided 2026-09-22

This record’s Open questions asked “whether the Top-Up Loan warrants its own record; the source calls it generally accepted.” Answered: yes. It is rrsp-gross-up-loan (IL-6), added in the Market Drop Wins fold-in, where Market Drop Wins names it “the single best bridge” into Smart Debt.

The two records are not duplicates, and the distinction is the whole point:

	this record (IL-4)	rrsp-gross-up-loan (IL-6)
What repays the loan	Payments over a multi-year amortization	The refund the contribution itself creates, within weeks
What the outcome depends on	Plan return exceeding the loan rate over ~10 years	Arithmetic. No hurdle rate to clear.
risk-level	2	1

That is also why the gross-up-refund comparison in this record’s Benefits section is the right benchmark rather than a rival: at an 8% return equal to the loan rate, catch-up and gross-up refund are exactly equal, so the multi-year loan adds nothing without excess return. IL-6 is the version where that dependency does not exist, because the loan is gone before the market matters.

A naming collision travels with this, and is flagged in both records: Talbot’s own corpus uses gross-up refund for the loan-free variant (strategy 3, implemented in sd-math), while Market Drop Wins uses RRSP Gross-up for the loan version. IL-6’s title is provisional pending a CEO decision.

Related strategies

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

Sources

• ~/projects/monorepo/packages/sd-math/src/sd_math/strategies/rrsp.py — implementation
• ~/projects/monorepo/packages/sd-math/docs/design.md — “RRSP Catch-Up Loan — Algorithm”, derivation, worked example, provenance caveat
• Talbot Stevens, Dispelling the Myths of Borrowing to Invest — “Borrowing for RRSPs” (RRSP-Strats.docx)
• Core/_WorkingOn/Research/canada-investment-taxation.md §6.5 (interest not deductible), §11.1 (contribution room)

Open questions

• Modelling partial-refund cases where the loan exceeds available room.
• Whether the Top-Up Loan warrants its own record. Closed 2026-09-22 — it is rrsp-gross-up-loan. Whether that record should be modelled in sd-math alongside this one is its own open question, recorded there.