46th Congress of the European Regional Science Association: "Enlargement, Southern Europe and the Mediterranean", August 30th - September 3rd, 2006, Volos, Greece
European economic integration and its enlargement phases foster the flows of international investments. This explains why the spatial distribution of economic activity is to an increasing extent determined by the location decisions of multinational firms (MNEs). Therefore, from a regional economy perspective, the geography of multinational activity is a policy concern for regional and local authorities. Not only does foreign activity have several contributions and implications for the regional economic development but also understanding the location process of these economic activities is crucial for adequate policy responses. This paper analyses the location decisions of European multinational firms across European countries (25) and regions (NUTS2). An emphasis is placed on whether location determinants for MNEs either change, or if not, which one dominates, at different spatial levels. We expect to find that agglomeration and dispersion forces behave in a different fashion depending on whether country or regional geographical scales are taken into consideration. We seek to explain how economic integration causes some multinational activity to locate in peripheral countries meaning a dispersion process at a supranational scale. The experience of Spain and Portugal in 1986, and more recently CEE countries, confirms this. They have received huge amounts of FDI coming from European countries. However, when the location decision is considered within a country, agglomeration tendencies come to dominate. The regions more developed within the host country attract the majority of these foreign inflows. With the recently incorporation of the knowledge capital model into a New Economic Geography (NEG) setting the links between multinational production and agglomeration are theoretically considered. But NEG setting seems to be scale independent. The same framework is used for analysing regional, urban and international issues. Then, an empirical test is necessary in order to account for these geographical scale differences. Empirically, we use a range of discrete choice models for location choices of MNEs across Europe. The independent variables are grouped into four categories: market access, labour market, agglomeration and others. The first results seem to confirm our hypothesis of a different behaviour of the location determinants depending on the geographical scale taken into consideration.