Mean vs. Median, Explained: How Your Income Percentile Rank Is Actually Calculated

Why the mean and the median tell different stories

Income distributions in most countries are right-skewed: the large majority cluster somewhere in the middle, while a small number of very high earners stretch a long tail far to the right. The mean — total income divided by the number of people — gets pulled toward that tail. The median doesn't; it's simply the income of the person exactly in the middle when everyone is ranked from lowest to highest.

Working from South Korea's full National Tax Service percentile file (109 bands, all 21,078,535 wage earners for tax year 2024), the mean wage is ₩44,749,680 and the median is ₩34,167,991 — a 31% gap. The top 0.1% alone average roughly ₩999 million a year, which is exactly the kind of outlier that pulls a mean upward without moving the median at all.

So how is your 'top X%' rank actually calculated?

The calculator's rank isn't derived from the mean or the median at all. It finds exactly where your converted income falls within each country's published income-distribution (percentile) data. That's why toggling 'Compare against' between Median and Mean doesn't change your rank — that's expected behavior, not a bug, because the rank comes from real distribution data rather than a mean/median assumption.

What the toggle does change is a secondary figure: the ratio to the median or mean shown alongside your rank. $60,000 in the U.S., for example, is 1.39× the median but only 0.94× the mean — while the rank itself (top 35.2%) stays the same in both cases. Where a country hasn't officially published a mean for that income type (parts of Germany, Singapore and Hong Kong's data), selecting 'Mean' shows 'not available' rather than a fabricated number.

Turning banded data into a percentile: linear interpolation

Not every country publishes a fine-grained percentile table. Japan's National Tax Agency wage survey, for instance, only reports how many people fall into brackets like ¥1–2 million, ¥2–3 million, and so on. For data like this, the calculator computes the cumulative share of people at each bracket boundary, assigns that boundary a percentile, and assumes income between two known boundaries falls on a straight line connecting them — linear interpolation.

In Japan's data, ¥4,000,000 sits at percentile 48.02 and ¥5,000,000 at percentile 63.34. An income of ¥4,500,000, exactly halfway between them, is estimated at percentile 48.02 + 0.5 × (63.34 − 48.02) ≈ 55.68 — a top-44.3% rank. South Korea's data is far more granular (109 bands, in increments as fine as 0.1% at the top), so its interpolation gaps are much smaller.

Where there's simply no published data point to interpolate between — Singapore's individual income statistics, for example, are only published at the 20th percentile and the median — the calculator doesn't guess. It shows the ratio to the median instead of inventing a rank.

Where the 8 countries' data actually comes from

Every figure comes from a national statistical agency's own published data — not a news report or a third-party estimate. Base years, survey methods and income definitions vary by country, and that's noted throughout.

• South Korea (individual): National Tax Service wage-income percentile file, tax year 2024, all 109 bands
• South Korea (household): Survey of Household Finances and Living Conditions (Statistics Korea, Bank of Korea, FSS), income year 2023, combined household income
• Japan: National Tax Agency Private Salary Survey (2024); Ministry of Health, Labour and Welfare Comprehensive Survey of Living Conditions (2022 income)
• United States: Census Bureau, Income in the United States: 2024; SSA Wage Statistics (2023)
• United Kingdom: ONS Annual Survey of Hours and Earnings; ONS household income inequality statistics
• Germany: Destatis (Federal Statistical Office) earnings structure survey
• Hong Kong: Census and Statistics Department earnings survey; 2021 Population Census
• Singapore: Ministry of Manpower; SingStat
• Malaysia: Department of Statistics Malaysia (household and salary surveys)

The exchange-rate conversion doesn't account for cost of living

Cross-country comparisons use the Korea International Trade Association's market exchange rate as of August 14, 2026 — a straightforward currency conversion, nothing more. It shows where the same nominal amount of money would sit in each country's income distribution, not what that money can actually buy there (purchasing-power parity is not applied). Exchange rates move, so results are a snapshot, not a permanent figure.

Common questions