Most international trade theory, whether classical or "new," predictst that increased globalization will be associated with increased locational concentration of particular economic activities, and hence increased specialization of national and regional economies, due to the greater freedom for industries to locate according to comparative advantage and economies of scale, and to integrate production systems based on an internationalization of intermediate goods sourcing. Relatively little empirical evidence exists on whether these predictions are correct. This paper presents the results of a statistical investigation of the trade-location relationship, using the OECD-STAN database, from 1970 to 1995. This investigation shows that in spite of rapidly rising trade, only in a very few industries has the spatial distribution changed substantially over the period studied. While intra-industry trade has risen across-the-board, locational concentration and specialization have increased little, if at all, in the European Union countries, and European economies remain much less specialized than equivalent regions of the USA. The paper then tries to speculate as to why this might be the case. Much of the intra-industry trade observed in Europe is probably not intermediate divisions of labor (production sharing), but head-to-head competition of largely national industries competing around similar products, through cross-market penetration. The question is how they manage to survive as such in an age of globalization. One hypothesis is that there are evolutionary dynamics involved: mature national firms and production clusters have capacities to adapt to changing circumstances which permit them to survive in more open markets. One major technique for adaptation is product differentiation, both horizontal (making the same products as competitors, through uptake of global state-ofthe- art knowledge) and vertical (quality differentiation, based on superior local knowledge). In this sense, the response of the European economies to globalization may reflect fundamentally different evolutionary dynamics from their American counterpart, whose regions integrated early on before they had mature industrial complexes, and where new industries tend to assume highly localized patterns, that serve as locational cores for the entire national industry. Most importantly, all of this implies that we need to develop non-deterministic theories of the relationship between trade and location, which take into account much more than the standard factors of comparative advantage and scale and integrate a dynamic evolutionary perspective.
