One survey, two populations
In the second quarter of 2026 the unemployment rate for everybody aged 15 and over living in Saudi Arabia was 3.0%. That is the figure the statistics office leads with, and its own bulletin introduces it as the rate for Saudis and non-Saudis together.
For Saudi nationals the rate was 6.5%, and for non-Saudi residents 1.3%. The Saudi rate is 4.9 times the other and 5.2 percentage points above it, and both figures come from the same release, in which every table is cut by nationality.
The headline rate is a blend of the two and is therefore neither one’s rate. It is not the rate for Saudi nationals, which is more than twice as high, and not the rate for non-Saudi residents, which is less than half as high. Nothing here adds or averages the two to reach the third, which the publisher computes on the whole population.
- A SAUDI NATIONAL and a NON-SAUDI RESIDENT are counted separately in every table of this release, and the publisher writes the difference into its own bulletin: the 3.0% is introduced as the rate "for Saudis and non-Saudis", and the 6.5% as the rate "among Saudis". The headline figure is a blend of two populations and is therefore neither one’s rate.
- THE UNEMPLOYMENT RATE IS A SHARE OF THE LABOUR FORCE. To be counted unemployed a person has to be without work, available for it and looking for it. Somebody who is not looking is not in the labour force and is not in the denominator either, so a population in which few people look can hold a low rate.
- THE PARTICIPATION RATE AND THE EMPLOYMENT TO POPULATION RATIO ARE SHARES OF THE WORKING AGE POPULATION, everybody aged 15 and over. That denominator cannot be left: a person who stops looking for work leaves the labour force and stays in it. This is why the two are printed beside the rate rather than instead of it.
- BEING OUTSIDE THE LABOUR FORCE IS A THIRD STATE, neither employment nor unemployment. 34.0% of women of working age in the country are in the labour force, and the 66.0% who are not are outside it, not unemployed. The release prints that share as its own row.
- THE THREE RATES IN A ROW TIE TOGETHER. Participation minus the employment ratio, divided by participation, reproduces the published unemployment rate, and it does so in all 198 cells of this release to within 0.05 of a point. That control is what fixes the meaning of each column without a definitions page: the one cell that lands 0.05 out is the single quarter the release prints to one decimal.
- AND A DIFFERENCE IN PARTICIPATION IS NOT A DIFFERENCE IN FINDING WORK. The survey records whether a person is employed, unemployed or outside the labour force. It does not record residency status, visas, how anybody came to be in the country, or how long any search lasted, so the gap between the two populations here is a gap in who is in the labour force and nothing more.
So these rates say what share of each population was working, looking or doing neither in one quarter, and nothing about why any share is what it is, what any job pays, or how anybody came to hold it.
Two populations, three rates and two denominators, all named by the release itself. A reading that treats the headline rate as the rate for Saudis, or that reads participation outside 100 as unemployment, says what the data do not.
Three rates, three populations
48.7% of Saudi nationals of working age are in the labour force, against 82.2% of non-Saudi residents. The employment to population ratio follows the same order, 45.5% against 81.1%, and the gap in both is more than 33 percentage points.
That is what produces the gap in the rates. The unemployment rate is a share of the labour force, so a population in which 82.2% of working age people are in that force and nearly all of them are working holds a low rate, while a population in which 48.7% are in it holds a higher one. The two statements describe different denominators and do not contradict each other.
The three rates in each row reproduce one another. Participation minus the employment ratio, divided by participation, returns the published unemployment rate in every row, and in all 198 cells the release publishes it does so to within 0.05 of a point. The single cell that misses by that much is in the one quarter the release prints to a single decimal.
| Population | Unemployment rate | Participation rate | Employment to population ratio |
|---|---|---|---|
| Saudi nationals | 6.5% | 48.7% | 45.5% |
| Non-Saudi residents | 1.3% | 82.2% | 81.1% |
| Everybody living in the country | 3.0% | 67.2% | 65.2% |
The gap is among the men
94.6% of non-Saudi men of working age are in the labour force, against 63.9% of Saudi men: a gap of 30.7 percentage points. Only 5.4% of non-Saudi men are outside the labour force at all, against 36.1% of Saudi men.
Among women the two populations are 1.0 point apart. 34.7% of non-Saudi women of working age are in the labour force and 33.7% of Saudi women, so 65.3% and 66.3% respectively are outside it. Almost the whole of the nationality gap in participation sits among the men.
On unemployment the women are the furthest apart of the four groups. Saudi women hold the highest rate printed here, 9.6%, and non-Saudi women 3.8%, while Saudi men are at 4.8% and non-Saudi men at 1.1%. Being outside the labour force is neither employment nor unemployment, so the first and last columns of a row are not two parts of one whole.
| Nationality and sex | Unemployment rate | Participation rate | Outside the labour force |
|---|---|---|---|
| Non-Saudi men | 1.1% | 94.6% | 5.4% |
| Saudi men | 4.8% | 63.9% | 36.1% |
| Non-Saudi women | 3.8% | 34.7% | 65.3% |
| Saudi women | 9.6% | 33.7% | 66.3% |
One year, two years, three years
The unemployment rate among Saudi nationals fell over one year, two years and three years alike, by 0.3 points from the second quarter of 2025, by 0.6 from 2024 and by 2.0 from 2023. Every window ends in the second quarter of 2026 and every base is a second quarter.
The share of working age Saudis in work fell over all three windows too, by 0.3, 1.6 and 0.9 points respectively, from 46.4% three years ago to 45.5% now. So none of the three windows supports a reading in which the rate improved because a larger share of Saudis was working, and this edition does not offer one.
For the population as a whole the same measure went the other way, rising 0.2, 1.2 and 1.8 points over the three windows to 65.2%. And the Saudi rate rose from the quarter immediately before, from 6.4% to 6.5%, a rise the statistics office reports in its own bulletin.
| Window ending second quarter 2026 | Saudi unemployment rate | Saudi employment to population ratio | Employment to population ratio, all residents |
|---|---|---|---|
| One year, from the second quarter of 2025 | -0.3 | -0.3 | +0.2 |
| Two years, from the second quarter of 2024 | -0.6 | -1.6 | +1.2 |
| Three years, from the second quarter of 2023 | -2.0 | -0.9 | +1.8 |
What these figures do not say
They do not say that the Saudi rate is at a low. Of the fourteen quarters from the first quarter of 2023 that this edition uses, 6.5% is the third lowest: the first quarter of 2025 holds 6.3% and the first quarter of 2026 holds 6.4%, and the latest reading is above both. The rate also rose from that immediately preceding quarter, which the publisher states plainly, so nothing here is a record of any kind.
They do not say that more Saudis are in work. The share of working age Saudi nationals in employment fell over one, two and three years, so the fall in the rate happened alongside a smaller share of Saudis working rather than a larger one. A rate can fall because fewer people are looking, and these three rates are printed together precisely so a reader can see which it was.
They do not say why any of this is so, and they do not say that non-Saudi residents find work faster. The survey records whether a person is employed, unemployed or outside the labour force, never the reason. It does not record residency status, visas or how anybody came to be in the country, so this edition names no cause for any figure in it: not residency rules, not nationalisation policy, not oil, not wages, not education. A reader supplying one is supplying it from outside these data.
And they are rates rather than counts, for one quarter, in one country. No level was read, so nothing here is a number of people, jobs or vacancies; nobody is followed between quarters, so a change in a rate is the difference between two photographs of different people; nothing is broken out by region, sector, occupation or education; no other country appears; and no comparison reaches behind the first quarter of 2023, because that is where this edition’s two independent reads of the same survey begin to agree. Nothing in it is a forecast.



