The comparison that should not be possible
In 2024, on the World Bank’s modelled ILO estimates, Nigeria’s unemployment rate was 3.0% and Germany’s was 3.4%. Viet Nam’s was 1.6%, the lowest of the nine countries in this piece. The Philippines reported 2.2% and Indonesia 3.3%. Spain, in the same year and the same source, reported 11.4%.
Nobody reading those numbers believes that it is easier to find work in Lagos than in Munich. That instinct is correct, and the reason it is correct is not that any of these countries is misreporting. Every one of these figures is produced to the same international definition. The definition is the thing worth understanding.
The rate counts a state of mind as well as a state of the market
To be counted as unemployed, a person has to fail to have work, be available to take it, and have actively looked for it in a recent period. All three at once. The third test is the one that does the damage in a comparison like this.
In an economy with unemployment insurance, a formal job market and a public employment service, looking for work is a thing you can do, and can be recorded doing, while not working. In an economy without those, not working is not an option that exists for most people. What happens instead is informal work: an hour here, a stall, a family plot, a relative’s shop. That is employment under this definition, and it removes the person from the numerator entirely.
- Without work: not in paid employment or self-employment, not even for one hour in the reference week.
- Available for work: able to start within a short reference period.
- Actively seeking work: having taken a specific step to look, within a recent reference period.
Someone who sells phone credit at the roadside for four hours a week fails the first test and is counted as employed. Someone who has stopped looking because there is nothing to look for fails the third and is counted as outside the labour force. Neither appears in the unemployment rate.
The international definition sets three tests, and a person must pass all three at once. This is why the measure behaves the way it does in an economy with no unemployment insurance: failing the third test moves you out of the numerator without anything improving.
Read the other two columns and the table inverts
Vulnerable employment is a defined category: own-account workers plus contributing family workers, as a share of everyone in work. It is a proxy for informal and precarious arrangements, not a direct measurement of them. Wage and salaried employment is its rough complement, the share of workers in an employment relationship with someone else.
Ordered by unemployment rate, the nine countries here run from 1.6% to 32.3%. Ordered by wage and salaried share they run almost the other way. Nigeria reports the third lowest unemployment rate in the table and the lowest wage share in it by a distance: 13.6%, against 91.8% in Germany. India reports 4.2% unemployment and 72.1% vulnerable employment, the highest share here.
Viet Nam and Indonesia sit in the same shape: unemployment under 3.5%, roughly half of all workers in vulnerable employment, and fewer than half in wage or salaried work. The Philippines is a step further along, with 63.5% wage employment, and Egypt further still at 72.9%.
| Country | Unemployment, % of labour force | Vulnerable employment, % of employed | Wage and salaried, % of employed |
|---|---|---|---|
| Viet Nam | 1.6% | 51.9% | 46.2% |
| Philippines | 2.2% | 33.9% | 63.5% |
| Nigeria | 3.0% | 66.7% | 13.6% |
| Indonesia | 3.3% | 50.3% | 46.3% |
| Germany | 3.4% | 4.4% | 91.8% |
| India | 4.2% | 72.1% | 24.6% |
| Egypt | 6.8% | 24.1% | 72.9% |
| Spain | 11.4% | 10.3% | 85.3% |
| South Africa | 32.3% | 11.7% | 82.7% |
South Africa is what tests the reading
If the argument above were simply "poorer countries report lower unemployment", South Africa would break it. It reports 32.3% unemployment in 2024, by far the highest of the nine, and it is not a rich country.
Look at its other two columns and the reading holds rather than breaks. South Africa reports 82.7% wage and salaried employment and 11.7% vulnerable employment, closer to Spain’s 85.3% and 10.3% than to Nigeria’s 13.6% and 66.7%. It has the employment structure of a formal economy, which means the category "unemployed" describes something real and available there in a way it does not in Lagos or Hanoi, and the measure duly finds an enormous number of people in it.
The lesson is not that one number is true and another false. It is that the unemployment rate answers a question about a formal labour market, and it can only answer that question where one exists to be asked about.
What this changes for somebody actually looking
If you are reading a national unemployment figure for Viet Nam, Indonesia, Nigeria, India or the Philippines and treating it the way you would treat one for Germany, you are reading it wrong, and so is anyone quoting it at you. A fall in the rate in those economies is as consistent with fewer formal openings as with more.
The two columns to read instead are in the table. A rising wage and salaried share is the thing that means more of the jobs on offer are the kind you can apply to, be interviewed for, and be paid a salary by. That is the market this desk covers, and in most of the world it is a minority of all work.
What the data establishes
- In the World Bank’s 2024 modelled ILO estimates, Nigeria reports a lower unemployment rate than Germany, 3.0% against 3.4%.
- Among these nine countries, the ones reporting the lowest unemployment rates report the highest shares of vulnerable employment and the lowest shares of wage and salaried work.
- South Africa reports both the highest unemployment rate here and an employment structure closer to Spain’s than to Nigeria’s.
- The international definition of unemployment requires a person to be without work, available, and actively seeking, all three at once.
What it does not
- That any country’s statistics are wrong, or that any national agency is misreporting. Every figure here is produced to the same definition.
- That informal work causes a low unemployment rate. Nine countries show an association in one year. No document cited here identifies a mechanism or rules out others.
- That Nigeria’s labour market is in better or worse condition than Germany’s. These three indicators do not measure wages, hours, security or living standards.
- That the unemployment rate is a useless statistic. Within one country over time, on a stable definition, it still tracks change. The failure is in reading it across countries with different labour market structures.
- Anything about 2025 or 2026. The latest year carrying all three indicators for all nine countries is 2024.
Method, and what was left out
- One source and one year. All three columns are World Bank Open Data indicators SL.UEM.TOTL.ZS, SL.EMP.VULN.ZS and SL.EMP.WORK.ZS, reference year 2024, read on 30 August 2026. 2024 is the latest year carrying all three indicators for all nine countries.
- All three are modelled ILO estimates, not national survey outputs. The ILO harmonises and imputes so that countries can be compared with each other. A national statistics office may publish a different figure for the same country and year on its own definitions, and neither figure is thereby wrong. Any national number quoted against these will not match, and should not be expected to.
- Vulnerable employment is own-account workers plus contributing family workers as a share of total employment. It is a proxy for informal and precarious work. It is not a measurement of the informal economy, which is defined and measured separately.
- The three columns have two different denominators. Unemployment is a share of the labour force. The other two are shares of the employed. They are not parts of one whole and do not sum.
- Nine countries were chosen to span the range and to include the counter-case, not sampled. A different nine would produce a different table. The argument rests on the direction of the relationship and on South Africa, not on any average of these rows.
- No figure in this piece comes from our own platform.
Primary sources
- World Bank Open DataUnemployment, total (% of total labor force), modeled ILO estimateSL.UEM.TOTL.ZS · August 30, 2026
- World Bank Open DataVulnerable employment, total (% of total employment), modeled ILO estimateSL.EMP.VULN.ZS · August 30, 2026
- World Bank Open DataWage and salaried workers, total (% of total employment), modeled ILO estimateSL.EMP.WORK.ZS · August 30, 2026
- International Labour OrganizationResolution concerning statistics of work, employment and labour underutilization (19th ICLS)ICLS-19 · August 30, 2026
- BabZitunaHow we test our matching for biasbias · August 30, 2026
The market this desk reads is the wage and salaried one. You can browse the live openings we hold, by city, or see how we test our own matching for bias.