Results show that multinationals are more likely to shut down operations, that job creation in foreign-
owned plants tend to be more persistent, while there is no significant difference between persistence of
job destruction in foreign or domestic plants. A similar result concerning the likelihood of plant
shutdowns is found in Fabbri, Haskel and Slaughter (2002) in their work on the UK. Plants belonging
to MNEs are more likely to be shut down than domestic plants, conditional on a set of plant and
industry characteristics.
3. The empirical analysis
3.1. Description of the data
The aim of this section is to compare labour demands in MNEs’ affiliates and in national firms
(henceforth NEs) in a sample that allows to analyse separately eleven European countries (Belgium,
Denmark, Finland, France, Germany, Italy, The Netherlands, Spain, Sweden and the United
Kingdom).9 The source of the data is the Amadeus database, a commercial database collected by
Bureau Van Dijk that reports balance sheet and other information for about 5 million firms in 31
countries between 1993 and 2000. This is the only data set for which it is possible to carry out cross-
country analyses comparing NEs and MNEs.10 We restrict our analysis to manufacturing firms in the
eleven sample countries that reported ownership status. From these we exclude firms with less than 50
employees and/or with less than three observations for the employment variable.11 The final total
number of firms used is 14,900. Their distribution by country is reported in Table 1.
9. These are the countries for which a sufficiently high number of firms reporting ownership status was available.
10 According to Amadeus dataset, a foreign firm is such when its ‘ultimate owner’ is non-resident in the country analysed.
An ultimate owner can be a company, a public authority, a state, a mutual fund, a pension fund, a nominee, a trust or a
trustee that owns at least 50% of the company. This information on the ownership status is not as accurate as we would
have liked. First, it is time invariant, as the database does not report this information by year. As a consequence, for all the
period, we classify as MNEs the firms which are foreign-owned in 2000. Second, information on the outward activities of
firms is often missing, so that it is not possible to distinguish between national firms with or without foreign subsidiaries:
Third, there is no information on firms’ closure, as the data set only reports information on firms which were active in 2000.
11 Data for firms with less than 50 employees are quite often not reliable, when not missing altogether.