Labour & wages

Mean versus median earnings

Average and median wages answer different questions about the pay distribution. What the ILOSTAT PPP series shows for the US, UK, Australia and New Zealand.

The average wage and the median wage are both published, both real, and different in what they tell you: the mean can be pulled upward by the very top earners, while the median sits at the exact middle of the earnings distribution. On the ILOSTAT series Younivi republishes, the gap is visible in every economy listed — in the United States for 2025, the average monthly figure of about 6,273 international dollars is well above the median of about 4,481.

The series here is ILOSTAT average and median monthly earnings, expressed in 2021 international (PPP) dollars so that countries can be compared after currency differences are taken out. Younivi republishes both series side by side on the Wages desk, with the ILO as the named producer and each figure's own year shown.

What each statistic measures

The mean adds up all earnings and divides by the number of earners. It answers: "if the wage bill were shared equally, what would each worker get?" The median is the value of the middle worker when everyone is lined up by earnings — half earn more, half less. It answers: "what does the person in the middle of the distribution earn?"

The difference is a property of averages under skew. Earnings distributions almost always have a long right tail — a minority of very high earners — and that tail lifts the sum, and therefore the mean, without moving the middle. The mean is thus optimally used as a total-income measure (it is exactly what the wage bill per worker is), while the median is optimally used as a "typical worker" measure. Neither is wrong; they describe the same distribution from two angles.

The figures in the republished series

New Zealand appears in the average series with a figure of about 4,064 for 2022 — the most recent year in the republished series — and the median series does not include the same year for New Zealand, which is why the table pairs it carefully. The point of the table is not ranking but structure: in every economy where both series exist, the mean exceeds the median, and the distance between them measures the skewness of the earnings distribution.

Why the distance between them matters

The mean–median gap is an index of concentration at the top. When most earnings gains accrue to the upper tail, the mean rises faster than the median: a country can look "richer" on the average while the middle worker's experience changes less. When distribution widens in the other direction — gains concentrated among low earners — the two series converge. That is why labour statisticians publish both: the mean tells you about the total, the median tells you about the middle, and the difference between their growth rates tells you about the distribution.

The gap should also be read with data-vintage side by side: the years in the published series differ (Australia's are 2021, New Zealand's 2022, the US and UK 2025), so a strict cross-country comparison must match years, not merely grab the latest row.

Why PPP dollars, and what they cover

PPP conversion uses price-level differences to express what a given amount of money can buy in each country, rather than what it exchanges for on the currency market. On the published series this matters: a figure in local currency alone is not comparable across countries, while a PPP figure answers "relative purchasing power of average earnings" — which is usually the real question. Its limits are that PPP baskets are consumption baskets, not native purchases specific to each country, and that the conversion is a statistical average rather than a personal exchange rate.

Coverage is a second limit. ILOSTAT earnings series measure employees' earnings from paid work; they do not include the self-employed's mixed incomes or capital income. And the definition of "earnings" can differ by country in scope (hours paid, bonuses, kinds of pay captured). The desk's figures are hence comparable within the ILO's harmonised definitions, and their producers say so on each series page.

Using the two series together

Three practical habits when reading the desk:

• Match the year before comparing countries, since economies update on different schedules.
• Compare the gap, not just the level: a pair of economies with similar means but different median gaps tells you more than two mean values alone.
• Look at the direction over time on the chart on the desk pages: whether the median is catching up to the mean is the question a trend can answer that a snapshot cannot.

Where this fits in the wider picture

A small number of very high earners is enough to move an average, and this is the single fact that makes "average pay" headlines so easy to overstate. The mean adds every earner's pay into one total and divides by the number of earners, so the top tail contributes in proportion to how much it earns. Because earnings distributions are skewed upward - a few people earn very large amounts, while the bulk cluster below - the mean sits above the middle of the distribution in almost every economy where both measures are published.

The median is immune to that pull. It reports the pay of the worker in the middle, so a very high earner entering the sample barely moves it, while the mean can shift noticeably. That is why medians sit lower than means almost everywhere, and why the size of the gap is itself informative: a wide gap signals a long and heavy upper tail, a narrow one a more compressed distribution.

For readers of pay statistics, the practical rule follows directly. A claim that "average pay" rose, or that "the average worker" earns some figure, is usually a mean-based claim about a total, not a description of the typical worker. The median answers the typical-worker question, and the two should be quoted together, because the difference between them is the story about who the growth reaches.

How to read this on Younivi

The average earnings page and the median earnings page on the Wages desk carry both ILOSTAT series with charts, tables, and the ILO attribution. Companion notes cover GDP per capita (which measures output, not pay) and unemployment series on the Labour desks. Figures are republished official data, may lag the ILO, and this note is informational only — not advice.