Skip to content
Sign inGet started free

The record · 21 September 2026

In Jordan, two employees in five are paid below two thirds of the national median.In Portugal, one in twenty-five.

Fifteen economies, one share each: the employees paid below two thirds of their own country’s median hourly earnings, with the same share for women and for men. Not poverty, not a minimum wage, and one year per country. Everything below is checkable against the live market.

· 7 min read

40.8%of Jordan’s employees were paid below two thirds of the national median in 2024, the highest share of 103 economies
4.0%of Portugal’s employees were, in 2025, the lowest of the 103; the median economy is at 21.8%
87of the 103 economies have a larger share of women than of men in low pay. The median gap is 5.0 points

Two in five, and one in twenty-five

In their latest reported year, 40.8% of Jordan’s employees were paid below two thirds of Jordan’s own median hourly earnings, and 4.0% of Portugal’s. Those are the highest and the lowest of the 103 economies that report all three figures since 2019: first and hundred and third. The table sets them among the other economies this desk publishes in, plus Uganda.

Between the two ends the spread is wide and the middle is crowded. The median economy is at 21.8%, and in 34 of the 103 at least a quarter of all employees are low paid; in 9 the share is under one in ten, and in only 2, Portugal and New Zealand, under one in twenty. Among the fifteen in the table, Indonesia is at 28.0%, Türkiye at 23.0% and India at 22.5%, while France is at 10.5% and Spain at 9.5%.

The ranking does not follow the map of rich and poor. Jordan, the highest share of the 103, is a middle-income economy, and it stands above Uganda, at 35.3%, whose income per person is less than a fifth of Jordan’s. Nine economies report a quarter of their employees low paid while sitting in the richer half of the set.

Provenance: Source
Low-pay rate: employees paid below two thirds of the median hourly earnings of their own country, as a share of all employees, latest year reported, for fifteen economies, with the same share for women and for men. Sorted by the overall share, highest first. The rank is out of the 103 economies reporting since 2019.
EconomyYearLow paid, % of employeesWomenMenRank of 103
Jordan202440.8%42.2%40.5%1
Uganda202135.3%51.1%28.4%10
Indonesia202328.0%40.5%21.9%31
Türkiye202523.0%27.4%20.8%42
India202522.5%37.9%16.9%45
Philippines202417.3%22.1%14.2%67
United Kingdom202516.3%18.8%13.9%73
Korea, Rep.202515.8%23.1%11.0%76
Germany202415.1%20.0%12.8%78
Viet Nam202411.5%14.3%9.3%90
Japan202510.7%17.4%6.6%93
France202410.5%11.8%9.3%94
Thailand20259.6%9.4%9.9%97
Spain20249.5%12.6%7.3%98
Portugal20254.0%4.2%3.6%103

A line drawn at home, and it still moves with income

Because the threshold is national, this measure has no arithmetic reason to fall as countries get richer. A country where every employee earns a fortune and a country where every employee earns very little would both report a low-pay rate near zero, if pay were equally spread. Across the 103 the share falls with income all the same: it correlates at minus 0.72 with log GDP per person.

Sorted into income thirds, the median low-pay rate is 32.0% in the poorest third of economies, 20.4% in the middle and 15.8% in the richest. The cuts fall at about 11,700 and 37,800 dollars a head. What moves is not the level of pay, which this measure never sees, but the shape of the bottom half of each country’s own distribution: the poorer the economy, the further its low earners sit below its own middle.

The pair the headline names sits inside that pattern without being explained by it. Portugal’s income per person is 4.2 times Jordan’s, which is a large gap but not a tenfold one, and the low-pay rates differ by a factor of ten. Income sorts these economies; it does not decide them one by one.

Provenance: SourceWhat the share counts
  1. THE LINE IS DRAWN AT HOME. An employee is low paid when their hourly earnings fall below two thirds of the median hourly earnings in their own country. The threshold is recomputed for every country and every year, so it is not a currency amount, not a legal minimum wage, and it supports no comparison of pay levels across borders.
  2. THE DENOMINATOR IS EMPLOYEES. Only people who work for an employer are counted. The self-employed, employers and contributing family workers are outside both halves of the fraction, which matters most where own-account work is the norm: there, this share describes a minority of everyone in work.
  3. LOW PAY IS NOT POVERTY. This is a property of one person’s hourly pay, not of a household’s income. A low-paid employee can live in a comfortable household, and a poor household can contain no employee at all; the two measures are built from different units and can point in opposite directions.

So Jordan’s 40.8% does not say that two in five Jordanians are poor, nor that a Jordanian employee earns less than a Portuguese one: it says that 40.8 in a hundred employees in Jordan are paid less than two thirds of what the middle Jordanian employee earns in an hour. Portugal’s 4.0% is the same count, measured against Portugal’s own middle.

One share, one threshold, and a reading that loses where the line is drawn states something the source does not.

What income does not touch

In 87 of the 103 economies a larger share of women than of men is low paid, and the median gap is 5.0 points. Men are the more exposed sex in 16, and among the fifteen in the table only Thailand is in that group, by half a point.

That gap is where income stops working. It correlates at minus 0.07 with log income, which is no relationship at all, and its median across the three income thirds runs 6.5 points, 4.9 points and 4.5 points. Women are above men in 24 of the 34 poorest economies and in 33 of the 35 richest: the gap is not smaller at the top, it is more uniform.

The extremes show the same thing from both ends. In Uganda 51.1% of women employees are low paid against 28.4% of men, the widest gap of the 103; Afghanistan is the widest the other way, at 11.7% of women against 26.2% of men, in an economy where few women are employees at all. Two rich economies in the table carry gaps as wide as some poor ones: Korea 23.1% against 11.0%, and Japan 17.4% against 6.6%.

What the share does not say

It does not say what anyone earns. The threshold is two thirds of a national median, so a low-paid employee in one country may take home more per hour than a well-paid employee in another. Nothing here supports a sentence comparing pay levels between two economies.

It does not say who is poor. Low pay is measured on an individual’s hourly earnings; poverty is measured on a household’s income after transfers and taxes, over a different unit and a different period. The two can disagree in both directions.

And it names no cause, no trend and no remedy. Minimum wages, collective agreements, hours, sectors, the size of the informal economy and who is counted as an employee all sit behind these figures, and none of them is measured here. Each economy appears once, in its own latest year, so nothing here says that any share rose or fell.

What the data establishes

  • In their latest reported year, 40.8% of Jordan’s employees were paid below two thirds of Jordan’s own median hourly earnings (2024), the highest share of 103 economies, and 4.0% of Portugal’s (2025), the lowest.
  • Across the 103 the median share is 21.8%; 34 economies are at 25% or above and 9 under 10%.
  • The share correlates at minus 0.72 with log income per person, with medians of 32.0%, 20.4% and 15.8% across the poorest, middle and richest thirds.
  • A larger share of women than of men is low paid in 87 of the 103, the median gap is 5.0 points, and that gap correlates at minus 0.07 with income: medians of 6.5, 4.9 and 4.5 points across the three thirds, with women above men in 33 of the 35 richest economies.

What it does not

  • That a low-paid employee is poor. Low pay is an individual’s hourly earnings against a national median; poverty is a household’s income after transfers and taxes.
  • That the threshold is a minimum wage. It is two thirds of the median hourly earnings, recomputed per country and year, and no legal floor is measured here.
  • Any comparison of pay levels between countries. Each share is measured against its own country’s middle, so a low-paid employee in one economy may earn more per hour than a well-paid employee in another.
  • That any share rose or fell. One year per economy, and different years and instruments across them: nothing here is a trend.
  • Any cause or remedy, including minimum wages, collective bargaining, hours, sector or the size of the informal economy. None of them is measured here, and nothing is said about China, which reports no year in this window, or about Finland, Iceland, Poland and Sweden, whose zeros are excluded.

Method

  1. One measure. ILOSTAT dataflow DF_EAR_XTLP_SEX_RT, low-pay rate by sex, read through the ILO SDMX API on 20 September 2026. An employee is low paid when their hourly earnings fall below two thirds of the median hourly earnings in the same country and year. For each economy the latest year from 2019 to 2025 is used, and the total, the women’s and the men’s figures are taken from the SAME year AND the same reported source, so a gap is never two different surveys held side by side.
  2. AGGREGATES ARE NOT COUNTRIES, AND FOUR ZEROS ARE NOT SHARES. Regions and income groups were removed by keeping only codes in the World Bank’s country list. Finland, Iceland, Poland and Sweden report an exact 0.0% for every sex, which no earnings distribution produces; they are excluded and named here rather than printed as the lowest in the world. That leaves 103 economies: 36 whose latest year is 2025, 37 with 2024, 9 with 2023, 6 with 2022, 4 with 2021, 3 with 2020 and 8 with 2019. Every rank and count is out of those 103. China reports no year in this window, which is why the table holds fourteen of this desk’s sixteen markets and Uganda.
  3. THE INSTRUMENTS DIFFER. 64 of the 103 figures come from a labour force survey, 13 from a household income and expenditure survey, 12 from an establishment or earnings survey, 9 from another household survey, and 5 from administrative records, an economic census, a population census or another source. The piece compares them as the ILO publishes them and says so here.
  4. The median is the 52nd of the 103 values, 21.802%, shown as 21.8%. 34 economies are at 25% or above, 9 under 10% and 2 under 5%. Jordan’s figure is 40.838%, Portugal’s 3.954%, and the ratio between them is 10.3.
  5. The correlation is a Pearson coefficient across the 103 between the share and the natural logarithm of GDP per capita at PPP (World Bank NY.GDP.PCAP.PP.CD, latest year). The income thirds cut at about 11,700 and 37,800 dollars a head, giving 34, 34 and 35 economies; each third’s figure is the median of its members. They describe how the shares sit together across economies, not how any economy moved. Jordan’s income per person is 12,632 dollars and Portugal’s 53,146.
  6. Rounding is done once, at display, from figures the source serves at three decimals: Uganda 35.304%, Indonesia 28.005%, Türkiye 23.009%, India 22.518%, Korea 15.828%, Japan 10.674%, Thailand 9.645%. The women’s and men’s figures are the source’s own for the same year; the gap quoted in the prose is their difference, computed at full precision and rounded once.

BabZituna scores every job against your profile on six real dimensions, and publishes what it can and cannot measure. See the live market.

The markets in this piece, live

Earlier editions