A pay rise cannot move you into a bracket that leaves you worse off

It is one of the most widely held false beliefs about money: that a raise can push you into a higher tax bracket and leave you taking home less than before. People turn down overtime over this. Occasionally they turn down promotions.

Brackets cannot do that, and the arithmetic takes one paragraph. But the belief is not baseless — something in the system genuinely can leave you worse off, it is just not the thing everyone blames. Both halves are worth having.

Why brackets cannot do it

Income tax is marginal. A tax band does not apply to your income; it applies to the part of your income that falls inside it.

Take a simple system: 20% on everything up to 50,000, and 40% above that.

Earn exactly 50,000 and you pay 20% of 50,000 — that is 10,000 — and keep 40,000.

Now take a 1,000 raise. The higher rate applies to the 1,000 above the threshold, and only to that. You pay the same 10,000 as before, plus 40% of 1,000, which is 400. Total 10,400, and you keep 40,600.

You are 600 better off. Not 1,000, because the last slice was taxed harder — but better off, and unavoidably so. Every additional pound is taxed at some rate below 100%, so every additional pound leaves something behind. The take-home line only ever bends. It never goes down.

Where the belief comes from

Two places, both understandable.

The phrase. “Moving into a higher tax bracket” sounds like you move, entire, into a new category — as though crossing the line reclassifies all your income rather than just the part above it. If that were how it worked the fear would be correct, and dramatically so. It is a genuinely misleading way to describe a system that taxes slices.

Payroll. Get a large bonus in one month and the tax taken can look extraordinary, because withholding often works by assuming this month’s pay continues all year — so a one-off is briefly treated as though you had become much richer permanently. It corrects, either across the year or at filing. But the payslip in front of you is real and looks like exactly the thing people were warned about.

What actually can leave you worse off

Not brackets. Cliffs, and they are usually not in the tax system at all.

The distinction that matters is between a taper and a cliff.

A taper withdraws something gradually as income rises. It raises your effective marginal rate — sometimes startlingly — but it still leaves you ahead. The clearest example is the UK’s personal allowance, which is withdrawn by £1 for every £2 earned above £100,000. Between £100,000 and £125,140 that produces an effective marginal rate of about 60%, higher than the top headline rate. You keep 40p of each additional pound rather than the 60p you might have expected. Less good. Still positive.

A cliff removes something entirely at a threshold. Cross it by a pound and you lose the whole thing. That is the case where earning more genuinely leaves you with less, and cliffs live mostly in means-tested support: benefits that stop at a fixed income, subsidised childcare that has a hard cutoff, help with housing or health costs that is all-or-nothing. Cross the line and you can lose several thousand for the sake of one.

So the fear is real and misfiled. It has been attached to brackets, which are smooth by construction, when it belongs to benefit thresholds, which frequently are not.

The number actually worth knowing

Not your tax band. Your effective marginal rate — what you keep out of the next unit of income, once everything that varies with income is counted: tax, national insurance or payroll contributions, student loan repayments, benefit withdrawal, allowance tapers.

That number is often nothing like the headline rate, and it is not monotonic. It can be higher in the middle of the income range than at the top, which is counterintuitive enough that most people have never checked their own.

If you are weighing a raise, more hours, or a second job, that is the figure to work out — and the only one where the honest answer is sometimes “less than you would think”, and very occasionally, at a genuine cliff, “don’t”.