[Table 2 about here]
Not surprisingly, in all 14 countries, the hotel and restaurant sector em-
ployed many foreign workers (category h). Another sector with a relatively
high proportion of foreigners is of household activities (category p), e.g.,
housemaids. In the other sectors, however, we see variations across the
countries. We also notice that the required skills in sectors with a high
proportion of foreign workers vary considerably. For instance, while the
construction sector (category f ) hired many foreign workers in more than
half the countries, they also seem to have been highly present in the educa-
tion sector (category m) in Finland and the health and social service sector
(category n) in the United Kingdom.
ESS collected a two-digit NACE Rev.1 code for each respondent, and
hence we know to which NACE Rev.1 major group he or she belonged.15
We assign the relevant figure in Table 2 to each ESS respondent.16 We call
this variable isb. We removed observations that belonged to extraterrito-
rial organizations (category q). The sector should naturally employ many
foreigners, and Table 2 confirms that in 5 countries. The exclusion of the ob-
servations should remove outliers, for the proportion of those who belonged
to such an organization in each country is very low: the highest is .013 in
France.
15Note that the coding was based on each respondent’s answer to ESS Question F24:
What does or did the firm or organization you work or worked for mainly do or make?
Since a respondent could give an answer based on the past work, the respondents without
a NACE Rev.1 code are not identical with the unemployed.
16This variable is similar to what Mayda (2006: Table 3A, Specifications 11 and 12)
constructed at the occupation level. We use sector rather than occupation because our
focus is on employers, not employees.
12