The Spousal RRSP: Income Splitting Before Retirement, and the Three Year Rule

A spousal RRSP is one of the few pieces of household tax planning that is genuinely simple in concept and genuinely easy to get wrong in practice. The concept: two people will be taxed less in retirement on two moderate incomes than on one large one, so the higher earner funds a plan that the lower earner owns, and the income arrives in the right hands later. That is the whole idea, and it works. The practice is where it goes wrong, because a rule sits underneath it that was written to stop the arrangement being used as a short term income shifting device, and that rule does not care whether the household was trying to do anything clever. Withdraw within a certain window and the money is taxed to the wrong person, at the wrong rate, and there is no way to unwind it afterwards. This article sets out how the plan works, who owns what, exactly what the rule says, and where a spousal plan still earns its place now that pension income splitting exists.

Who owns what, and who deducts what

This is where most confusion starts, so it is worth being exact. In a spousal or common law partner RRSP there are two roles. The contributor is the person who puts the money in. The annuitant is the person who owns the plan.

The contributor uses their own contribution room, not the annuitant’s, and claims the deduction on their own return. That is the point: the deduction lands with the higher earner, where it is worth the most. The annuitant owns the plan, controls the investments, and normally reports the income when money is withdrawn, which is the second half of the point, because that is where the income is intended to be taxed.

A spousal contribution does not create room for the annuitant. If the lower earning spouse also has their own room, it is separate and unaffected, and they can contribute to their own plan as well. The two are not connected beyond the fact that a household has to keep track of which plan is which, because the rule below turns on it.

The three year attribution rule, stated exactly

Here is the rule that decides the outcome, and it deserves to be read slowly. If the contributor made a contribution to any spousal or common law partner RRSP in the year a withdrawal is made, or in either of the two preceding years, then all or part of that withdrawal is included in the contributor’s income rather than the annuitant’s.

Three features of that sentence catch people. It says any spousal plan, so a contribution to a second spousal plan can attribute a withdrawal from the first. It counts calendar years rather than elapsed months, which means a contribution made in December of one year and a withdrawal made in January three calendar years later are further apart on the calendar than they feel, while a contribution in January and a withdrawal in December of the third year is nearly three full years and still inside the rule. And it applies to the withdrawal regardless of intention: nobody has to be doing anything improper for it to bite.

The way to be certain the rule does not apply is stated plainly by the Canada Revenue Agency: make no contribution to any spousal plan in the year of the withdrawal or in either of the two preceding years. That is the test, and it is a calendar exercise rather than a judgment call.

The practical implication for a household using a spousal plan properly is that contributions should stop well before withdrawals are expected to begin. A couple who contribute up to the year they retire and then draw on the plan immediately have built the arrangement and then defeated it.

Where it still earns its place

Pension income splitting, available from age 65 for qualifying income, does much of what the spousal RRSP was invented for. That is a real change and it has narrowed the case. It has not removed it, and the gaps are specific.

The first is early retirement. Income splitting on qualifying pension income generally becomes available at 65, and a couple who stop working earlier have years in between where a spousal plan is doing work that nothing else does.

The second is the shape of the income. Not everything qualifies for splitting, and the rules about what does are technical. Where a household’s retirement income is largely registered savings drawn before 65, the spousal plan matters more.

The third is a durable difference in earnings. Where one spouse has consistently earned much more, the deduction is worth more to them now and the income will be taxed less in the annuitant’s hands later. That arithmetic is unchanged by the existence of splitting.

The fourth is a difference in age. Where the annuitant is younger, the plan can stay registered longer, since the deadline for converting is based on the plan holder’s age. And a contributor past 71 with their own room remaining can still contribute to a spousal plan while the younger spouse is under the age limit, which is the only remaining route to an RRSP deduction at that point.

Where it goes wrong

Withdrawing too soon is first, and the rule above explains it. It is almost always innocent: a household needs money, the spousal plan is the account with a balance, and nobody checks the calendar.

Losing track of which plan is which is second. Over twenty years, with institutions merging and accounts consolidating, spousal and personal plans get combined, and the character of the money follows it. Keeping a spousal plan separate from the annuitant’s own plan is administratively tidier and makes the three year test answerable rather than a research project.

Contributing without coordinating is third, and it is the over contribution route: a spousal contribution uses the contributor’s room, so a couple who each contribute to their own plans and to a spousal plan without adding it up can exceed the contributor’s limit without either of them realising.

And assuming it survives a separation is fourth. Once contributed, the plan belongs to the annuitant, and what happens to it on a separation or divorce is a matter of family law and of the agreement between the parties rather than of who funded it. That is a question for a lawyer or notary, and it belongs in the conversation before the arrangement is built rather than after.

How to use one properly

Decide first whether the arrangement is answering a real difference. If both spouses will have similar retirement income, the spousal plan is added complexity for no benefit.

Keep the accounts separate and labelled, so that three years from now anybody can answer the question the rule asks without reconstructing a decade of statements.

Track contributions against the contributor’s room, together with their own plan and any group plan, once a year, using the limit on the notice of assessment.

And plan the stop date. Work backwards from when withdrawals are expected to begin and stop contributing three calendar years before that. It is the single step that prevents the only serious failure this arrangement has, and it costs nothing but a note in a calendar.

The situations the attribution rule does not reach

The rule as stated above is the working version and the one to plan around. It is not absolute. The Canada Revenue Agency describes a small number of situations in which a withdrawal is not attributed to the contributor even though a contribution was made inside the window.

Separation is the first. Where the spouses or partners are living separate and apart at the time of the payment because the relationship has broken down, the amount is generally the annuitant’s income.

Death is the second. Where the contributor dies in the year the payment is made, the withdrawal is generally taxed in the annuitant’s hands rather than attributed.

Residence is the third. Where the contributor is not resident in Canada at the time of the payment, the rule does not operate in the ordinary way.

Transfers are the fourth, and barely an exception. Amounts moved directly from the plan into another registered vehicle are not received by anyone, and the rule addresses amounts received.

Each of those carries conditions, each is stated by the Canada Revenue Agency in its own words, and none of them is a plan. They are what happens when life happens, and the reason to know they exist is to stop a household concluding that a withdrawal after a separation or a death landed on the wrong return when it did not. Confirm your own facts with a qualified tax professional before the return is filed.

What happens when the plan becomes a RRIF

A spousal plan does not stop being one when it is converted. Moved into a registered retirement income fund, it becomes a spousal RRIF, and the character travels with the money. That is the most useful thing to know about the arrangement in its last phase, because a household that has forgotten the plan was ever spousal is about to be surprised.

The attribution rule follows it, with one adjustment that matters. The minimum amount that must come out of a RRIF each year is not attributed to the contributor. It is the annuitant’s income, which is exactly what the arrangement was built to produce. Anything taken above that minimum is treated like any other withdrawal, and the same calendar test applies: did the contributor put money into any spousal plan in that year or in either of the two before it.

Two consequences follow. A couple who want the income to land with the annuitant should be taking the minimum and not more while the window is still open. And a contributor who is still making spousal contributions while the annuitant has begun drawing above the minimum is working against the arrangement from both ends at once.

The minimum itself is set by a formula rather than chosen, which is why how it is calculated is worth reading before the conversion: everything above that figure is the part the household actually decides.

Frequently Asked Questions

Continue Learning

Read RRSP Contribution Room for the limit a spousal contribution uses, Income Splitting for Couples and Families for the other tools available to a household, and RRIF Minimum Withdrawals for what these plans become. The Wealth Creation guide sets out the whole picture.