An income percentile calculator answers a question people ask constantly and almost never get a straight answer to: out of everyone in this population, what share earns less than this? The arithmetic is simple. The difficulty is that a percentile is only meaningful relative to a specific population, measured on a specific basis, in a specific year and country — and most tools that claim to answer it quietly hide all four of those choices.
This page makes them explicit. Arb Digital publishes it in the free tool library at arbsbuy.com with no distribution data baked in: you supply the published median for the population you want to be compared against, and a dispersion assumption, and the calculator models the rest. That is a different job from the live percentile calculator, which computes percentiles for any dataset you paste in. This one is specifically about placing an income against national earnings statistics.
What This Income Percentile Calculator Does
Enter an income, the published median for the comparison population, and a dispersion ratio — the published 90th percentile divided by the published median. The calculator fits a lognormal distribution to those two anchors and reports where the income falls in it, along with the share of the population earning more.
The supporting figures make the position concrete rather than abstract. The ratio to the median is often more informative than the percentile itself. The next decile threshold shows what income the next round-number milestone requires. The target percentile field converts any goal — the 90th, the 75th, the top 1 percent — into the income figure that corresponds to it under the same assumptions.
Nothing here is a measurement. It is a model fitted to two numbers you supply, and it will only be as good as they are. That is stated plainly rather than hidden, because a percentile presented without its population, basis, year and country is close to meaningless.
How to Use It
- Decide the population first. Individual earnings, household income, full-time employees only, working-age adults including those with no earnings — these give wildly different answers for the same person.
- Take the median from an official source for that exact population, and note the year and country. Median household income and median full-time employee earnings are not interchangeable.
- Build the dispersion ratio from the same table. Divide the published 90th percentile by the published median. Do not carry a ratio across countries or across populations.
- Match the income basis. Gross against gross, pre-tax against pre-tax. Comparing your net pay with a gross median moves the answer by a whole decile or more.
- Read the ratio to the median alongside the percentile — it is the figure that stays comparable when you change population or country.
The Formula / How It's Calculated
Earnings distributions are right-skewed: a long thin upper tail pulls the mean well above the median. A lognormal distribution captures that shape with two parameters, and two published anchors are enough to fix both.
The median of a lognormal is exp(μ), so μ = ln(median). The 90th percentile sits 1.2816 standard deviations above the mean of the underlying normal, so σ = ln(P90 ÷ median) ÷ 1.2816. An income x then sits at percentile Φ((ln x − μ) ÷ σ), where Φ is the standard normal cumulative function. Running the calculation backwards, the income at percentile p is median × exp(z(p) × σ).
Work the defaults. The median is 39,863 and the dispersion ratio is 2.2, so σ = ln(2.2) ÷ 1.2816 = 0.7885 ÷ 1.2816 = 0.6152. For an income of 52,000, ln(52,000) − ln(39,863) = 10.8590 − 10.5932 = 0.2658, and dividing by σ gives z = 0.4319. Φ(0.4319) = 0.6671, so the income sits at about the 66.7th percentile, with roughly 33.3 percent of the population above it.
The next decile up is the 70th, whose z-value is 0.5244. That gives 39,863 × exp(0.5244 × 0.6152) = 39,863 × 1.3808 = about 55,041, so roughly 3,041 more income would reach it. The ratio to the median is 52,000 ÷ 39,863 = 1.30. At the default target of the 90th percentile, the model returns 39,863 × 2.2 = 87,699 by construction, which is a useful check that the fit is behaving.
Where to Get Real Distribution Figures
The anchors matter more than the model, and both should come from a national statistics office rather than a secondary source. For the United States, the Census Bureau's report Income in the United States: 2024 publishes household income, earnings, income inequality and post-tax income based on the Current Population Survey Annual Social and Economic Supplement, with detailed tables including distribution measures.
For the United Kingdom, the Office for National Statistics publishes the Annual Survey of Hours and Earnings. Its bulletin on employee earnings in the UK: 2025 reported median gross weekly earnings of £766.60 for full-time employees in April 2025, a 5.3 percent nominal increase on the year and 1.1 percent in real terms after CPIH. That weekly figure is the source of the default median on this page, annualised, and it is a UK full-time-employee figure for April 2025 — not a household figure, not a US figure, and not a figure for anyone working part time.
Two cautions when reading either source. Survey-based earnings data usually excludes some groups by design — the self-employed are commonly outside employee earnings surveys entirely — so a self-employed reader comparing against employee percentiles is comparing against a population they are not in. And the top of the distribution is measured poorly by household surveys, because very high incomes are rare, sometimes top-coded, and under-reported. Percentiles above roughly the 95th should be treated as indicative in any survey-derived model, including this one.
Why the Population Definition Changes Everything
The same person can be at the 40th percentile or the 80th depending only on who they are being compared with, and none of those answers is wrong.
Household versus individual is the largest single lever. Household income sums everyone in the home, so a two-earner household appears far higher in a household distribution than either earner does in an individual one. Comparing a personal salary against a household median — an extremely common error — makes almost everyone look poorer than they are.
Full-time-only versus all-workers is the next. Excluding part-time workers removes a large group with lower annual earnings and raises every threshold accordingly. Age matters as much: earnings typically rise through a career and flatten or fall later, so a 25-year-old at the 40th percentile of all workers may be well above the median for their own age group. Geography matters too, which is why the cost of living calculator is a necessary companion — an identical income buys very different lives in different cities, and a national percentile says nothing about that.
Finally, gross versus net. Published earnings statistics are almost always pre-tax. If you want to compare living standards rather than positions in a pay distribution, the after-tax income calculator and the take-home pay by state calculator get you to a comparable net figure first.
What a Percentile Does Not Tell You
This is worth stating directly, because the number invites over-interpretation. A percentile is a description of a measured population in a particular year and country. It is not a judgement about the person whose income it places, and it carries no information about whether that income is fair, sufficient or well-earned.
It is also silent on wealth. Income and net worth diverge sharply — a retired household drawing modestly on substantial assets can sit low in an income distribution and high in a wealth one, and a young professional with a large salary and large debts is the reverse. It says nothing about stability either: two people at the same percentile, one on a permanent contract and one on irregular freelance work, are in very different positions despite an identical annual figure.
And it moves without anyone's circumstances changing. Percentile thresholds are recalculated each year in nominal terms, so in an inflationary period an unchanged real income drifts downward through the distribution. Comparing a percentile computed against 2024 data with one computed against 2019 data measures the intervening inflation as much as anything else. Converting both to real terms with the inflation calculator before comparing is the only way to separate the two effects.
Arb Digital builds long-term online growth programmes for consultants, practices and small firms, so enquiries arrive steadily rather than through referral luck.
Web Growth Services Talk to Arb DigitalCommon Mistakes to Avoid
- Comparing a personal salary with a household median — the two populations are different, and the mismatch understates individual position badly.
- Mixing net and gross — published statistics are almost always pre-tax, so entering take-home pay against them shifts the answer by a decile or more.
- Carrying a dispersion ratio between countries — earnings inequality differs enormously, and the ratio has to come from the same table as the median.
- Trusting the extreme tail — survey data measures very high incomes poorly, so anything above roughly the 95th percentile should be read as indicative.
- Reading a percentile as a verdict — it describes a measured population in one year and country, and says nothing about any individual's circumstances.
Related Free Tools From Arb Digital
Pair this with the annual income calculator to get to a comparable annual figure, the hourly to salary calculator if you are paid by the hour, the after-tax income calculator for the net view, the lifetime earnings calculator for the same income projected across a career, the inflation calculator for real-terms comparison, and the percentile calculator for percentiles of any dataset. The full free online tools hub lists everything else.
Frequently Asked Questions
It is the share of a defined population whose measured income falls below a given figure. Being at the 70th percentile means about seventy percent of that population earned less and thirty percent earned more, in the year and country the data covers.
Because there is no single median. It depends on the country, the year, and whether the population is households, all workers, or full-time employees only. Supplying it yourself from an official source keeps the answer honest and never out of date.
From your national statistics office. In the United States the Census Bureau publishes annual income reports from the Current Population Survey, and in the United Kingdom the Office for National Statistics publishes the Annual Survey of Hours and Earnings.
No, and the difference is large. Household income sums every earner in the home, so a dual-income household sits far higher in a household distribution than either person does in an individual one. Compare like with like.
Because earnings are right-skewed, with a long thin upper tail that pulls the mean above the median. A lognormal reproduces that shape closely across most of the range from just two anchor points.
Less accurate than in the middle. Household surveys measure very high incomes poorly because they are rare and sometimes top-coded, so any survey-derived model, including this one, should be treated as indicative above roughly the 95th percentile.
No. It describes how measured incomes were distributed in a population, not what any household needs. Costs vary by location, family size and circumstances, none of which appear in a distribution statistic.
This calculator models a distribution from figures you supply and is provided for general information only. It is not financial advice, it contains no official statistics of its own, and the results are modelled rather than measured — take medians and percentile data directly from your national statistics office and treat any figure here as indicative.
