Field entry, published 2026-11-03, updated 2026-11-03. Group: ISCO-08 in statistics and surveys

How are migrant worker statistics built with ISCO-08?

Migrant worker statistics are built with ISCO-08 by crossing an occupation group, coded from the work a person does, with a migration related variable recorded in the same source, such as country of birth, citizenship, place of previous residence, or duration of stay, depending on the definitions that source uses. ISCO-08 itself contains no migration information. Its role is to provide the occupation side of the table in a standard form, for comparison on a shared basis. The migration side of the table does most of the definitional work, and it is also where comparability most often fails.

Which migration variable is in the table?

The first question for any migrant worker table is not about occupation at all. It is about which concept the table uses, because country of birth, citizenship and recent arrival describe different populations, and a person can belong to one category but not another. Country of birth is fixed over a lifetime in most cases, citizenship can change, and duration of stay changes with time, so tables built on different variables will show different occupational patterns even when the underlying workers are similar. A table that says migrant workers without naming its variable has not yet defined its population.

Sources differ in what they can record. A census can usually support detailed breakdowns by country of birth and occupation, including for small areas, while a sample survey may support only broad groupings because migrant workers in detailed occupation groups may be few in the sample. Administrative sources may record citizenship or permit type, but may not record occupation in ISCO-08 form at all, or may record it only at hiring. Each source therefore offers a different combination of occupation detail and migration detail.

Why is occupation coding especially sensitive here?

Occupation coding for migrant workers raises practical issues that can affect the table before any comparison begins. Job titles and qualifications earned elsewhere may not translate directly, descriptions may be given in a second language, and work experience in another labour market may involve tasks that fit ISCO-08 groups less neatly. Interviewers and coders need clear guidance for such cases, including how to handle foreign titles, unrecognised qualifications and mixed duties, and offices need consistency checks that specifically include these cases. Without that attention, coding differences can masquerade as occupational differences.

Recognition of qualifications adds a further layer. A worker may hold a qualification from another country that is not recognised locally, and may therefore work in a different occupation group from the one that qualification would suggest in the country of origin. ISCO-08 coding follows the work actually done, not the qualification held, so the table will show the current occupation, not the occupation the worker trained for. That distinction matters for interpretation. A table showing concentration in particular groups does not by itself show that qualifications match those groups.

What comparisons are useful, and which ones mislead?

Useful comparisons keep the migration variable, the occupation level, the population and the period constant, and then vary one element at a time. Comparing the occupational distribution of workers born abroad with that of workers born in the country, within the same source and period, can reveal structural differences worth investigating. Comparing the same migrant group over time, on a consistent basis, can reveal whether its occupational profile is changing. Comparing the same group across places requires checking that both places used the same migration variable and the same ISCO-08 level, since a difference in either can create an apparent difference in occupation.

Misleading comparisons usually break one of those conditions quietly. A table that compares a detailed occupation breakdown for one group with a broad breakdown for another, a table that mixes country of birth in one place with citizenship in another, or a table that compares a census snapshot with a survey covering a different period, will generate differences that reflect method rather than work. Small groups deserve particular care, since a detailed occupation cell for a small migrant population can rest on very few cases, and a striking share in such a cell may be unstable.

What can these tables show, and what must they not claim?

Migrant worker tables built with ISCO-08 can show the occupational structure of a defined migrant population, how that structure differs from other workers in the same source, and how it changes over time on a consistent basis. They can support discussion of labour market integration, sectoral concentration, working time and, when linked carefully, pay and conditions, provided those additional variables come from the same source or from a defensible linkage. They cannot show, on their own, why a pattern exists, whether it reflects choice, constraint, recognition barriers, language, networks, discrimination or employer practice, or whether it is desirable.

That boundary is not a reason to avoid the tables. It is a reason to read them with their definitions attached, to state the migration variable in every commentary, to check occupation level and cell size, and to bring in further evidence before explaining a pattern. When those habits are followed, ISCO-08 gives migrant worker statistics a common occupation language. When they are skipped, the same tables can turn a definitional difference into a false story about work.

Sources checked

1. ILO: ISCO-08

Back to the ISCO-08 in statistics and surveys group or browse all occupations.