only themselves. Also note that foreign workers were highly present in the
sector across countries: see Table 2. If immigrants into this sector also self-
employed themselves, the number of service providers would increase in the
sector, intensifying competition in the output, rather than input, market.
We will check on this line of reasoning later.
Second, we find that, in the hospitality sector (category h), the proba-
bility of preferring immigration restriction is higher for employers than the
others among those who anticipated immigration to reduce the cost of labor.
This is counter intuitive, given the standard argument exemplified by Lord
Layard’s letter to the FT in the introduction. Note that the estimated effect
of employ is also significantly positive in the construction, wholesale/retail
and repairing service sectors (categories f and g) among those who did not
think that immigration would depress wages. Categories f, g, h and p gen-
erally employed more foreign workers than in the other categories, as Table
2 shows. It might be that employers had experienced dissatisfaction with
the quality of migrant labor they employed. Another possibility is that em-
ployers anticipated intensified competition in their output, rather than input,
markets due to immigration.
Third, there is only one sectoral category that gives evidence consistent
with our expectation, i.e., category o that includes sewage and refuse disposal,
sanitation-related activities and recreational, cultural and sports activities.
The estimated effect of employ is significantly negative among those who
thought that immigration would reduce wages.
When discussing the positive effect of employ in the household-activity
sector above, we speculated that the size of an employer in terms of the
30