Volume 9 • 2022 • Number transnational corporations investment and development

that 
relocate 
(%)
Welfare 
state 
expenditure 
as % of GDP 
in 2013
Welfare 
state 
expenditure 
as % of GDP 
in 2019
Number 
of firms 
relocating
Percentage 
of firms 
that 
relocate 
(%)
Austria
26.2
25.9
528 11.7
28.3
26.9
60 11.7
Australia
15.9
16.3
40 2.5
21.3
21.8
d
22 31.8
Belgium
25.2
25.4
801 16.2
29.2
28.9
147 10.9
Canada
17.0
16.2
477 2.3
17.0
18.0
c
..
..
Czechia
17.5
17.9
242 20.2
21.1
19.2
42 7.1
Denmark
26.9
26.0
321 10.3
32.6
28.3
20 5.0
Estonia
15.2
12.7
44 27.3
15.6
17.7
..
..
Finland
27.5
23.7
316 25.9
29.5
29.1
99 19.2
France
29.3
28.8
1 494 11.0
31.7
31.0
207 9.7
Germany
26.1
24.8
1 959 7.7
27.2
25.9
346 9.8
Greece
18.0
21.5
28 3.6
25.9
24.0
..
..
Hungary
21.2
22.9
31 3.2
22.2
18.1
7 14.3
Ireland
15.8
16.6
53 5.7
21.5
13.4
6 33.3
Italy
22.9
24.8
2 644 19.9
29.1
28.2
506 7.9
Japan
14.5
18.6
422 14.7
22.7
22.7
d
545 5.7
Korea, Republic of
3.6
7.6
54 18.5
9.5
12.3
..
..
Mexico
4.0
6.7
47 10.6
7.6
7.5
..
..
Netherlands
b
21.8
21.1
–
–
24.1
16.1
7 28.6
Norway
21.6
20.0
220 16.4
23.2
25.3
44 18.2
Poland
21.8
19.4
72 5.6
20.5
21.3
11 18.2
Portugal
16.6
22.4
161 7.5
25.8
22.6
24 4.2
Slovenia
22.6
20.0
95 22.1
23.8
21.1
..
..
Slovakia
18.0
15.6
113 17.7
17.7
17.7
..
..
Spain
20.6
21.3
1 454 12.2
25.7
24.7
211 6.2
Sweden
30.2
27.0
744 20.4
27.4
25.5
47 10.6
Switzerland
18.0
17.7
1 017 10.7
26.1
27.4
c
..
..
United States
14.4
15.8
1 006 1.9
18.8
18.7
..
..
United Kingdom
18.3
20.1
523 18.0
23.3
20.6
125 11.2
Total
-
-
14 906
13.1
-
-
2 476 9.1
Source
: Authors’ compilation based on data from OECD social expenditure database (for welfare state expenditures); ORBIS and the 
Orbis crossborder investment database (for firms that are relocating).
Note
: The sample from 2013 to 2019 may not contain the same companies covered in 1997–2007 given that the latter sample is 
generated using Orbis cross-border investment database. However, we do not believe that this can be a problem because results for the 
latter period remain largely the same and they present a robustness test for the earlier period.
– Nil.
-
Not applicable.
.. Not available.
a
For welfare state expenditure, countries are included for a restricted period due to information availability. For the number of firms and 
percentage of relocation, several countries did not enter the regression sample because there is no complete data coverage for each 
year in the 2013–2017 period.
b
For the period 1997–2007, we did not observe any firms relocating from the Netherlands, which means that this country did not enter 
the regression sample for this period.
c
Latest available information in 2018.
d
Latest available information in 2017.

---

12
TRANSNATIONAL CORPORATIONS 
Volume 29, 2022, Number 2
3.2 Independent variables
We include several independent variables in our empirical estimation model. One of 
the main variables of interest is “home welfare spending”, which is defined as the 
total public social expenditure by country as a percentage of GDP for every home 
country in our sample. Firm size is measured via the natural logarithm of a firm’s total 
turnover. Labour intensity is included as the ratio of the number of employees to 
turnover. We also include a measure of productivity, namely unit labour costs which 
is the ratio of a firm’s average wage over turnover level. Host welfare spending is 
included as the total public social expenditure by country, and as a percentage of 
GDP for every host country in our sample. Industry differences are captured by 
including NACE two-digit sector dummies. Eurostat classifies high-tech industries 
in accordance with the following 2-digit NACE industry codes: 24, 29, 30, 31, 32, 
33, 34 and 35; and low-tech industries are classified by the following 2-digit NACE 
Table 2a. Descriptive statistics for key variables, 1997–2007 
Variable
All manufacturing industries
Mean
Standard