National insurance calculator — methodology

The four charges, and who pays each

• Employee Class 1 (primary) — the deduction on a payslip. Charged on gross pay above the primary threshold; the rate falls above the upper earnings limit.
• Employer Class 1 (secondary) — a charge on the employer, paid on top of pay and never deducted from it. One rate, a lower threshold, and no upper limit.
• Class 4 — the self-employed equivalent of employee Class 1, charged on trading profits, with the same falling shape at lower rates.
• Class 2 — no longer a compulsory charge. Above the small profits threshold the qualifying year is credited at no cost; below it, contributions may be made voluntarily.

employee NI = mainRate  × (min(pay, upper earnings limit) − primary threshold)
            + upperRate × max(0, pay − upper earnings limit)

employer NI = rate      × max(0, pay − secondary threshold)          (no upper limit)

class 4     = mainRate  × (min(profit, upper profits limit) − lower profits limit)
            + upperRate × max(0, profit − upper profits limit)

cost of the employment = pay + employer NI          (an addition, not a deduction)

Those expressions are written out to show the shape of each charge, not because the page computes them: employeeNI, employerNI, class4NI and class2Status do, and the calculator renders the bands they return. The last line is the only arithmetic the page itself performs on them, and its direction is the point.

Why the employee rate falls, and why it matters more than it sounds

Employee Class 1 runs at 8.00% between £12,570 and £50,270, and at 2.00% above that. It is one of very few rates in the UK system that decreases with income. The consequence is that a pay rise costs less National Insurance per pound to somebody already well paid — on £60,000 the next pound costs the employee 2.00% and the employer 15.00%.

Both of those figures are measured, by asking each engine what one more pound costs, rather than by comparing pay to a threshold and reading a rate off a table. The two methods disagree at exactly the interesting point: at pay of precisely the upper earnings limit, no pound has yet been charged at the upper rate, so a threshold comparison reports the main rate while the next pound genuinely costs the upper one. A forward difference has no such branch and the answer still comes out of the engine, twice.

The employer charge is a cost, not a deduction

On £60,000 of pay the employee contributes £3,210.60 and the employer contributes £8,250.00 — more than twice as much — so HMRC collects £11,460.60 on the job and the employment costs £68,250.00. None of the employer’s share is deducted from the employee, appears on their payslip or reduces their take-home pay, and no arithmetic on this site subtracts it from anything.

It is worth being precise about what is and is not being claimed. The legal incidence is settled: the employer owes the secondary contribution. The economic incidence is not — a substantial body of economic work holds that employer payroll taxes are borne largely by employees through lower wages over time. That is a claim about wage bargaining rather than about the tax code, this page marks it as contested where it states it, and it changes nothing about who pays the bill.

A director’s annual earnings period, and the threshold this software does not have

A company director pays the same Class 1 as any other employee — same thresholds, same rates, no director rate. What differs is the earnings period. A director’s contributions are worked out on cumulative earnings for the whole tax year against the annual thresholds, with whatever has already been paid deducted; everybody else’s are worked out on each pay period on its own.

That is modelled here in full rather than mentioned, and it needs no new engine and no new figure, because employeeNI is already a charge over the annual thresholds — which is exactly what the annual earnings period applies. A director’s deduction in month m is the engine called twice:

deduction in month m = employeeNI(pay to date including m)
                     − employeeNI(pay to date before m)

On £60,000 paid in 12 equal months, that gives a director 2 months with no deduction at all, then the main rate until cumulative pay passes the upper earnings limit, then the upper rate for the rest of the year. The twelve deductions add to £3,210.60 — the same annual total any other employee on that pay would reach.

The ordinary employee’s month-by-month profile is deliberately not shown beside it. Producing it needs HMRC’s monthly primary threshold and upper earnings limit, and tax-core carries neither. They are separately published figures rather than the annual ones divided by 12: the annual primary threshold is not the weekly one multiplied by 52 either, and the same disagreement runs in both directions. A monthly column obtained by dividing would be a figure this site invented printed next to figures it verified, which is precisely the failure the verification log exists to prevent. So the director’s side is exact, the comparison is stated qualitatively, and the reason is on the page rather than only in this file.

What National Insurance is not charged on

It is a charge on work. Dividends, savings interest, rental income and pension income bear no National Insurance at any rate, however large they are — income tax may be due on all four, and often is, but no National Insurance. That is not a simplification this page chose: the National Insurance surface of tax-core takes exactly one amount per call, pay or profit, and has no parameter anywhere for a dividend, an interest payment, a rent or a pension withdrawal. There is nowhere to enter them because nothing would happen to them.

It is also not devolved. Every figure on this page is set at Westminster and identical in Scotland, England, Wales and Northern Ireland. Income tax on earnings is devolved to Scotland and does differ, which is what makes this a common mix-up; it changes nothing here. The Scottish pages on this site say the same thing from the other direction: employeeNI takes no region argument, because there is no region for it to take.

Rates and thresholds

Every figure above is UK-wide, and every threshold is HMRC’s published annual figure rather than a weekly one multiplied by 52. The same gov.uk table publishes both and they disagree by enough to change an answer; packages/tax-core/VERIFICATION.md records the trap, and records that the one threshold which does multiply out cleanly is the lower earnings limit — which this software does not use, so checking a single row would prove nothing.

Sources

• gov.uk — National Insurance: how much you payThe taxpayer-facing statement of the employee Class 1 rates and of what National Insurance is charged on.
• gov.uk — National Insurance rates and categoriesThe category letters. This calculator prices the ordinary one and none of the reliefs that use the others.
• gov.uk — Rates and thresholds for employersHMRC’s published ANNUAL thresholds, employee and employer, which are deliberately not the weekly figures multiplied out — and the monthly thresholds this software does not carry.
• gov.uk — Self-employed National Insurance ratesClass 4 and the current status of Class 2, which is credited rather than charged above the small profits threshold.
• gov.uk — CA44: National Insurance for company directorsThe annual earnings period, which is the whole of what makes a director different and the only director-specific rule modelled here.
• gov.uk — Changes to salary sacrifice for pensions from April 2029The dated change that stops salary sacrifice being fully exempt from National Insurance. Outside every tax year offered here, and the reason this page will not state the exemption unconditionally.

Six things this calculation does not model, stated here rather than buried

• Category letters other than the ordinary one. Every employment has one, and several of them reduce or remove the employer charge entirely — employees under 21, apprentices under 25, veterans, and employees at freeports and in investment zones. The Employment Allowance, which lets many employers set a fixed amount against their secondary Class 1 bill for the year, is not modelled either. Every one of those makes the employer figure shown here too high for the employers they apply to, and none of them are in tax-core.
• Age. Employee Class 1 is not charged below 16 and is not charged after State Pension age, while the employer’s charge continues throughout. There is no age input on the page, so it prices somebody of working age. A reader over State Pension age and still working should read their own figure as nil and their employer’s as unchanged.
• More than one employment. National Insurance is charged per employment and each gets its own thresholds, so two half-sized jobs do not produce the same contributions as one whole one. The statutory annual maximum for somebody with several employments, and the rules for holding Class 1 and Class 4 in the same year, are not applied.
• Pay periods, for anybody but a director. The engines are annual and HMRC’s weekly and monthly thresholds are not in this software, for the reason set out above. So a month containing a bonus is not modelled for an ordinary employee, and the annual figure shown is what a full, level year comes to.
• Qualifying years. What is paid and what is credited are different questions. The lower earnings limit — which governs whether a year counts towards the State Pension, and below which an employee pays nothing and can still qualify — is recorded in VERIFICATION.md as a figure this package deliberately does not carry, because it affects entitlement rather than liability.
• Class 1A and 1B on benefits in kind, Class 3 voluntary contributions, income tax of any kind, student loan repayments, pension contributions and tax codes. A salary sacrifice reduces the pay both Class 1 charges are worked out on and is priced on its own page; from April 2029 only the first £2,000 a year of employee pension contributions made that way keeps the National Insurance exemption, which is outside every tax year offered here.