Marginal vs average tax rate

Two numbers describe the same tax bill, and people reliably reach for the wrong one. Your marginal rate is what the government takes from your next dollar. Your average rate — often called the effective rate — is the total tax you paid divided by your total income. The first is always the higher of the two, and they answer entirely different questions: marginal decides whether the overtime shift is worth taking, average tells you what the year actually cost.

Almost all the confusion between them traces back to a single misreading of how brackets apply, and that misreading is expensive. It is why otherwise sensible people turn down raises, refuse hours in December, or believe a bonus pushed them into a higher bracket and left them poorer. A bracket cannot do that. What follows is the mechanism, and then the places where the price of the next dollar really does turn ugly.

Brackets are segments, not categories

A tax bracket is not a box your income falls into. It is a segment your income passes through, and only the dollars sitting inside a segment are taxed at that segment's rate. Earning one dollar more than a threshold means exactly one dollar meets the higher rate. Everything underneath is untouched.

Take an illustrative ladder — round numbers chosen to keep the arithmetic legible, not real combined federal and provincial rates, which differ by province and move most years:

Someone earning $70,000 under that ladder does not pay 33% on $70,000. They pay nothing on the first $16,000, 25% on the next $40,000, and 33% on the final $14,000 — zero plus $10,000 plus $4,620, or $14,620 in total. Their marginal rate is 33%. Their average rate is $14,620 divided by $70,000, which is 20.9%. Both figures are correct, and a twelve-point gap between them is entirely ordinary.

Your real ladder stacks a federal set of thresholds on a provincial one, each with its own boundaries, so the combined rate steps up more often and in smaller increments than any single table implies. The tax bracket tool lays both ladders out for your province.

The raise that cannot cost you money

Run a raise through the same ladder. Someone earning $54,000 sits just under the third band; a $4,000 raise pushes $2,000 of income across the line.

The raise is $4,000 gross and $2,840 net. The tax on it comes to $1,160 — 25% on the first half, 33% on the second, 29% blended. The average rate moved less than a point, because only $2,000 out of $58,000 ever met the higher rate.

No arrangement of brackets makes this go the other way. Every ladder is built the same way: a rate applies above a threshold, never retroactively below it. When a raise genuinely does leave a household worse off, brackets are never the cause — a benefit phasing out with income is, and that is a different mechanism, covered further down.

Marginal is the price of the next dollar

The marginal rate is a price, and it works in both directions: every additional dollar of income costs it, and every dollar of deduction refunds it. That makes it the only rate worth consulting about a change at the margin.

A second job, an overtime block, a rental suite, a bonus — that income stacks on top of what you already earn, so its first dollar meets the rate that applied to your last dollar of salary, not your much gentler average. Someone with a 20.9% average and a 33% marginal rate who budgets the side income at 20.9% will overstate its take-home by more than a third, and will find the shortfall at filing time rather than at payday.

The same price tag explains what an RRSP deduction is worth. At $70,000 on the ladder above, a $5,000 contribution drops taxable income to $65,000 — still inside the 33% band — so it saves $1,650. Contribute $20,000 instead and the top $14,000 of the deduction is worth 33%, while the remaining $6,000 lands back in the 25% band: $4,620 plus $1,500, or $6,120, a blended 30.6% rather than the headline figure. The bigger the deduction, the more of it refunds at a lower rate. That is the arithmetic behind holding a deduction back for a year when income peaks instead of claiming everything at once — the RRSP calculator prices a specific contribution, and the fuller RRSP guide covers the timing.

Run your numbersTax brackets→

Average is what the year cost

The average rate looks backward, and it is the right tool for anything to do with cash flow: what share of gross pay ended up as tax, whether this year was heavier than the last, how much gross income a spending plan of $60,000 actually requires. It is also the honest way to compare two compensation packages, because it describes the whole bill rather than the last slice of it.

The two rates diverge in a characteristic shape. The marginal rate steps — flat, jump, flat again. The average rate curves up smoothly, always trailing below the marginal line and closing on it only slowly, because there is always a long tail of income taxed at every lower rate beneath you.

MarginalAverage

Marginal and average rate as income rises

	Marginal	Average
$50k	25	17.9
$100k	33	25.4
$150k	43	32.1
$200k	48	32.1
$250k	48	36.5

Illustrative ladder only: 0% to $16,000, then 25%, 33%, 43% and 48%. Real combined rates differ by province.

That gap is why the average rate is a poor guide to decisions and the marginal rate is a poor description of your year. Worth reading alongside this: what the deductions on your paycheque actually are. Withholding is calculated as though your current pay continues at that level all year, which is why a one-off bonus month is usually over-withheld and handed back months later.

Where the effective marginal rate spikes

Published brackets are not the full price of the next dollar. Income-tested benefits and credits phase out as income rises, and a lost benefit costs a household precisely what tax does. Stack a phase-out on a bracket and you get an effective marginal rate that appears in no published table.

• The Guaranteed Income Supplement is reduced by 50 cents for each dollar of other income. A retiree in a 25% bracket who is receiving it faces an effective marginal rate near 75%.
• Old Age Security is recovered at 15 cents on the dollar above a threshold, so income in that range for someone in a 43% bracket is really costing 58%.
• The Canada Child Benefit tapers against family net income, at a rate that rises with the number of children. A family with several children inside a phase-out range can face an effective marginal rate above the highest bracket published anywhere in the country.
• The age amount, provincial credits, and various rent and drug-plan benefits each carry their own taper, and they stack rather than take turns.

Two things follow. Deductions are worth more than their headline rate whenever they pull income out of a phase-out range, because the refund is the bracket rate plus the clawback rate recovered. And in retirement the shape of income matters more than its size — moving income to a lower-rate spouse can lift a household out of two tapers at once, which is the real argument for pension income splitting.

Payroll contributions run the same logic in reverse inside the year. CPP and EI stop once their annual maximums are reached, so take-home per dollar rises in the closing months for anyone who hits both ceilings — a drop in the effective marginal rate that has nothing to do with brackets and resets every January.

Which number, when

• Marginal: extra shifts, side income, a bonus, an RRSP contribution, whether to realize a capital gain this year or next — any question of the form "is the next dollar worth it".
• Average: budgeting, comparing one year against another, converting a spending target into the gross income that funds it, and judging whether a pay change actually improved anything.
• Effective marginal, clawbacks included: any decision taken while a benefit is phasing out, which for families with young children and for most retirees means nearly every decision.

Put your own figures through the income tax calculator and read both outputs rather than one. The marginal rate is the number to act on; the average rate is the number to plan around. Confusing them costs money in both directions — declining income that was worth taking, and budgeting for take-home that never arrives.