To link to this article:

Transcription

1 This article was downloaded by: [K.U.Leuven - Tijdschriften] On: 08 November 2011, At: 01:53 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK International Journal of the Economics of Business Publication details, including instructions for authors and subscription information: Inter-European Regional Dispersion of Corporate Research Activity in Information and Communications Technology: The Case of German, Italian and UK Regions Grazia D. Santangelo Available online: 21 Jul 2010 To cite this article: Grazia D. Santangelo (2000): Inter-European Regional Dispersion of Corporate Research Activity in Information and Communications Technology: The Case of German, Italian and UK Regions, International Journal of the Economics of Business, 7:3, To link to this article: PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: terms-and-conditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or

2 up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

3 International Journal of the Economics of Business, Vol. 7, No. 3, 2000, pp. 275± 295 Inter-European Regional Dispersion of Corporate Research Activity in Information and Communications Technology: The Case of German, Italian and UK Regions GRAZIA D. SANTANGELO ABSTR ACT The increased geographical dispersion of corporate research seems to emphasize the significance of the local dimension over the pace of technological development. However, although a large number of studies have focused on intra-border analysis in the US, the European regional dimension has been explored only recently due to the lack of data availability. This paper investigates the regional dispersion of Europeanowned ICT research activity across German, Italian and UK regions. The empirical results suggest that locally embedded value added and the subsequent localised knowledge spillovers stemming from it are promoting spatial agglomeration economies, which generate competitive corporate advantage based on untraded externalities. Key words: ICT; Corporate research and development; European regions. JEL classification: O3, F23, R1. 1. Introduction In the new socio-techno-economic paradigm, although time and space constraints have been reduced or (in some cases) completely eliminated by the pace of technological development, paradoxically geography seems to matter more than ever before. On the one hand, the shift towards a knowledge-based economy involves a shift in organisation away from top-down hierarchical infrastructures to flatter structures based on intra-firm networks of semi-autonomous corporate subsidiaries (Amin and Thirft, 1995; Archibugi and Michie, 1995; Cantwell, 1995; Zander, 1998; Stiglitz, 1999). On the other hand, the increasing globalisation of economic activity has emphasized the importance of clusters of innovative local centres over time accumulating new knowledge embodied in best practises (Amin and Goddard, 1986; Knox and Agnew, 1989; S Èolvell and Birkinshaw, 2000). The author is thankful to Professor John Cantwell for having provided the data that this paper is based on. Needless to say, the end product is the author s own responsibility. Grazia D. Santangelo, Facolta di Giurisprudenza, Universita degli Studi di Catania, Via Gallo 24, Catania, Italy; gsantangelo@lex.unicit.it. ISSN print/issn online/00/ Taylor & Francis Ltd

4 276 G. D. Santangelo Therefore, the higher the level of corporate internationalisation, the stronger the links that the firm establishes with the local system. If this is a general trend across industries, it applies particularly to science-based industries (e.g. information and communications technology ± ICT), where links between corporate competitors, and firm± university and user± producer relationships are crucial in the development and adoption of new flexible technologies (Malecki, 1985; Oakey and Cooper, 1989; Oakey, 1985). In the case of science-based and related industries (e.g. ICT), local networks and infrastructures may promote the potential for knowledge creation and learning. The point is illustrated in the case of the New Jersey telecommunications industry (Wymbs, 1998), Silicon Valley computer industry and Route 128 (Saxenian, 1994; McCann, 1995; Storper, 1997; Scott, 1988). In this context, the reason for a spatial analysis at the regional level lies in the fact that nation-state investigations are likely to neglect intra-border inequalities between regional economic units. If, at country-level, empirical studies have shown high levels of heterogeneity in the accumulation and diffusion of technology (Cantwell, 1987; Archibugi and Pianta, 1992; Padoan, 1997), at regional level the situation is even more heterogeneous in terms of both GDP growth (Champion et al., 1996; Fagerberg, 1996; Iammarino et al., 1998; Iammarino and Santangelo, 1998; Magrini, 1998) and technology development (Cowan and Foray, 1996; Cantwell and Iammarino, 1998, 1999). Although great emphasis has been placed on the role of geography in explaining local and corporate innovation, research has mainly focused on the US regions (e.g. Jaffe et al., 1993; Audretsch and Feldman, 1994; Feldman, 1994). Conversely, the focus on the European regional dimension in the development of innovation is rather recent. In an analysis of the spatial dimension of innovative activity across European regions, the European integration process should be taken into account. QuÂevit (1996) argues that the completion of the Single European Market (SEM) together with global economic and technological changes impact on the performance of European regions. The first empirical results seem to point out that income disparities may owe their existence to the process of knowledge specialisation between `knowledge-creating and `knowledge-applying regions and that they have been exacerbated by European Union (EU) integration (e.g. Magrini, 1998). The location of large MNCs seems to reinforce the top of the national urban hierarchies. A reason for this trend has been found in the European failure to take advantage of the opportunities offered by the new techno-socio economic paradigm (Begg et al., 1999). The pervasive character of ICT is provoking a transformation in urban and regional systems by generating a new network model involving core regions and neglecting the periphery (Hepworth, 1986; Malecki, 1994; Antonelli, 1995; Castells, 1985). Therefore, as technological change affects not only the long-term growth of economies but also the spatial distribution of economic activity, a concentration phenomenon has emerged (Feller, 1975). The crucial role of innovation in boosting local economies has been recognised in the EU regional policy. In order to account for this new techno-socio-economic condition, the EU has implemented its own regional policy, which, articulated through the Structural Funds action, 1 aims at economic and social cohesion across the Union (as officially stated in the Treaty on European Union). In this context, the aim of this paper is to analyse the dispersion/concentration of ICT corporate innovative activity within the European ICT industry across German, Italian and UK regions. The starting-point of the analysis is the result obtained in a previous study (Santangelo, 1998), where clusters of European

5 Regional Dispersion of Corporate Research Activity 277 electrical companies specialising in the same ICT technologies were identified over the period 1969± On these grounds, the present study aims to test the following hypothesis. Hypothesis: in each of the three sub-periods under analysis (1969± 77, 1978± 86 and 1987± 95) the identified technological clusters of companies grouped together as specialising in the same ICT technology (cospecialisation) also locate the related research and development (R&D) in the same region (co-location). In other words, it is tested whether cospecialised European electrical companies co-locate investments in the ICT field of technological co-specialisation in the same European region. The paper is organised in six main sections. The following section exposes the complementarity of the global and local dimension in corporate activity. Section 3 describes the data used. Section 4 discusses the general trends in the location of ICT corporate research activity in the European regions under analysis. The empirical evidence on the hypothesis of co-location and co-specialisation is presented and discussed in Section 5. Conclusions are drawn in Section `Global versus `Local : A Complementary Relationship The growing economic globalisation seems to have emphasized the importance of the local dimension in both corporate production and research activity. The understanding of globalisation proposed in this paper embraces a wider dimension than the one evoked by prefixes to the word national such as inter-, multi- or trans-. The term globalisation is here adopted to indicate the increasing geographical dispersion of intra-firm networks across a large number of locations and the consequent interdependency between the units which constitute the MNCs. The establishment of geographically dispersed networks allows the transfer of technology, skills and assets across national borders between the parent company and the subsidiaries. The sustainable competitive advantage built on this transfer lies in the two-way interaction between parent and subsidiaries. The transfer goes from the parent to subsidiaries, but also from the subsidiaries to the parent company. Local laboratories play a new role within the whole corporate structure by sourcing new knowledge from the local environment rather than carrying out mere supplyoriented activities (Papanastassiou and Pearce, 1994; Bartlett and Ghoshal, 1995). Therefore, corporate activity is more and more sensitive to localised cumulative processes as multinationals aim to outsource knowledge by tapping into local expertise. Following the `new economic geography view (e.g. Krugman, 1991a, b), 2 increasing returns are essentially a regional and local phenomenon arising from regional economic agglomeration and specialisation. In this sense, the globalisation phenomenon has emphasized the importance of sub-national local entities (e.g. regions), which shape corporate spatial organisation according to the different types of indigenous agglomeration forces. On the one hand, the localisation of new corporate units is driven by general external economies and spillover effects attracting all kinds of economic activities in certain regions. This process promotes the spatial concentration of different economic activities as well as the emergence of regional cores. On the other hand, the localisation of the new corporate units is

6 278 G. D. Santangelo driven by sector-specific economies and spillover effects attracting certain economic activities according to the indigenous sectoral specialisation. The interaction between general external economies and localisation economies generates a geographical hierarchy of regional centres as empirically shown in recent studies (Cantwell and Janne, 1999; Cantwell and Iammarino, 1998, 1999; Iammarino et al., 1998; Iammarino and Santangelo, 1999). This hierarchy appears to be stable over time as a result of cumulative causation mechanisms promoting vicious and virtuous circles and reinforcing geographical inequalities. A strong path-dependent character is found in the local ability to develop an entrepreneurial environment through expertise accumulation. The success of the `learning region relies on the capability of mobilising technical resources, knowledge and other inputs essential to innovation (Amin et al., 1994; Feldman and Florida, 1994). Localised higher value added rather than reduction of transaction costs and cheap labour seem to be the variable explaining economic agglomeration and performance. Manufacturing, human, physical and communications, and industrial governance infrastructures (Florida, 1995, 1996; S Èolvell and Bengtsson, 2000) are key factors defining the `knowledge-based or `learning region, able to maintain its sustainable advantage over time. Since, in the global arena, host locations are competing in order to attract MNCs quality investments, Howells (1999) remarks that regional systems of innovations do not undermine the globalisation process in terms of production and diffusion of technology but rather they reinforce it. Therefore, `global and `local seem to be complementary dimensions in the context of economic globalisation. 3. The Data In this paper, corporate geographical distribution of ICT technological development across Europe over the period 1969± 1995 is investigated by adopting US patent data drawn from the University of Reading database. The use of patenting in a common third country ± i.e. the US ± allows a more reliable international comparison on a similar basis. Furthermore, foreign patents (e.g. European) are expected to be of a higher quality than domestic patents (i.e. US) as it is reasonable to assume that only inventions and innovations with highest expected profits will be patented abroad due to the time and cost involved in so doing. Advantages and disadvantages of patent data are well known in the literature (see Pavitt, 1985; Griliches, 1990) and will not be discussed further here. However, it should be pointed out that in an analysis of ICT, a major problem in using patent data might be that software innovations have started to be patented since the mid± 1990s. Nonetheless, for the purpose of the present study, this is not a major issue considering the weak European performance in software technology by comparison with US competitors (Malerba et al., 1997). For the purpose of the present paper, it is worth noting that the patent document records the address of both inventor and owner of the invention. In other words, the location where the R&D activity was carried out and the location of the headquarters of the company to which the patent is assigned is provided by the patent records. Therefore, it is possible to identify geographically the location where the R&D activity underlying the invention was carried out. Moreover, the patent document information allows one to consolidate patents into corporate groups on the basis of the structure of ownership. 3 In the Reading database, each patent is classified by the type of technological activity with which it is primarily associated. The 399 original classes in this way

7 Regional Dispersion of Corporate Research Activity 279 identified by the US Patent and Trademark Office can be further grouped into 56 technological sectors, of which six sectors comprise the main field of ICT. 4 The data used refer to all 23 European firms classified in the broad electrical corporate industrial group in the database held at the University of Reading. For the purpose of this paper, the empirical analysis was carried out at the level of the original 399 technological patent classes by selecting those which correspond to the six ICT technological sectors. 5 The broad electrical corporate industrial group includes the electrical equipment (communications) and the office equipment (computing) industries. 6 The spatial analysis of the corporate research activity in the European ICT industry is carried out at sub-national level. The geographical distribution of European electrical corporate patenting activity in the ICT technological sectors is investigated in Germany, Italy and the UK, for which regionalised patent data are currently available in the Reading database. For each of these three countries, the sub-national entities identified correspond to territorial units as classified by the European Nomenclature of Territorial Units of Statistics (NUTS) nomenclature. In order to ensure as much comparability as possible, the NUTS 1 level is used to identify German and UK regions, while, as far as Italian regions are concerned, the NUTS 2 level is adopted. As pointed out by Eurostat (1995), despite the aim of ensuring that comparable regions appear at the same NUTS level, each level still contains great differences among the territorial units identified in terms of area, population, economic weight or administrative power. Therefore, the 16 German lèander, the 11 UK standard regions and the 20 Italian regioni seem to guarantee grounds of comparability as far as innovative activity is concerned (see Table A3 in the Appendix) General Trends in the Location of ICT Corporate Research Activity in German, Italian and UK Regions Before investigating whether the technological co-specialised groups of firms identified in the study referred to above (Santangelo, 1998) are co-located in the regions under analysis, it is worth considering the significance of German, Italian and UK regions in the overall geographical distribution of European electrical research in ICT. As reported in Table 1, in Europe almost half of the European R&D investments in ICT by European electrical companies are located in the regions under analysis. This confirms CaniÈels (1998) findings of a high geographical concentration of innovative activity in a few European regions, as well as Verspagen s (1997) argument for the existence of European `regional clubs. The heterogeneity characterising what Caracostas and Soete (1997) name as the European System of Innovation in terms of regional performance may be identified as a major factor explaining high concentration of innovative activity in a few regional locations. Inter-national and inter-regional technology gaps are the results of historically different approaches to investments in technology. On the one hand, myopic systems (either national or regional) treat investments in technological activities just like any conventional investment; on the other hand, dynamic national and/or regional systems place great emphasis on the development and accumulation of intangible assets (Patel and Pavitt, 1996). In this sense, uneven technological development and consequent economic performance is due to the cumulative and localised character of technology (Cantwell, 1994). Therefore, the emergence of

8 280 G. D. Santangelo Table 1. Distribution of US patents attributed to European electrical companies in all ICT technological sectors relative to Europe (%), by region, 1969± 95 Region No. of European electrical firms Share of home research Share of European host research Bayern (G) 20.4% % 9.0% South East (UK) 9.9% % 33.0% Baden-Wurttemberg (G) 3.2% % 61.0% Niedersachsen (G) 1.8% % 61.0% Schelswig-Holstein (G) 1.6% % 45.0% Hamburg (G) 1.4% 4 1.0% 99.0% Lombardy (IT) 1.2% % 88.0% Nordrhein-Westfalen (G) 1.2% % 63.0% Piedmont (IT) 1.2% % 5.1% Hessen (G) 1.0% % 65.0% North West (UK) 1.0% % 13.0% East Midlands (UK) 0.8% % 1.0% West Midlands (UK) 0.7% % 4.0% East Anglia (UK) 0.7% % 75.0% South West (UK) 0.6% % 8.0% Berlin (G) 0.5% % 35.0% Rheinland-Falz (G) 0.3% % 43.0% Scotland (UK) 0.1% % 5.0% Yorkshire & Humberside (UK) 0.1% % 14.0% Bremen (G) 0.1% 7 0.0% 100.0% Wales (UK) 0.1% % 23.0% Sicily (IT) 0.1% 3 0.0% 100.0% Emilia Romagna (IT) 0.1% % 40.0% Thuringen (G) 0.1% % 13.0% Brandenburg (G) 0.1% % 72.0% North (UK) 0.1% % 12.0% Friuli-Venezia Giulia (IT) 0.1% % 0.0% Lazio (IT) 0.1% 3 0.0% 100.0% Mecklenburg-Vorpormmen (G) 0.0% % 60.0% Tuscany (IT) 0.0% % 40.0% Sachsen-Anhait (G) 0.0% % 0.0% Sachsen (G) 0.0% % 0.0% Calabria (IT) 0.0% 1 0.0% 100.0% Liguria (IT) 0.0% 1 0.0% 100.0% Campania (IT) 0.0% 1 0.0% 100.0% Total (German regions only) 31.7% % 27.4% Total (Italian regions only) 2.8% % 56.9% Total (UK regions only) 14.0% % 33.8% Total (German, Italian and UK regions) 48.5% 69.0% 31.0% Total (all other European regions) 51.5% `higher and `lower order regions relies on the local ability in developing a regional system of innovation. If at the European level European-owned ICT research is highly concentrated, within German, Italian and UK regions it appears to be even more polarised in fewer locations. Table 1 also illustrates that Bayern and South East (UK) together account for 30% of the total European-owned research in ICT sectors carried out in European regions between 1969 and This great

9 Regional Dispersion of Corporate Research Activity 281 polarisation identifies these two regions as `higher order locations in the development of innovative activity. To explore further the terms under which this technological polarisation occurs, Table 1 ranks German, Italian and UK regions on the grounds of the percentage of European ICT technological activity located in each of them relative to Europe over the period 1969± The table also reports the relative share of ICT research activity carried out by home and European host electrical multinationals. As expected, Bayern and South East (UK) emerged as the centres of excellence in the technological development of ICT. German regions are on average at the top of the hierarchy. Baden-W Èurttemberg, Niedersachsen Schleswig-Holstein, for instance, are all top-ranked. As far as UK regions are concerned, South East (UK) stands out among all others. Similarly, Lombardy is highly placed among the Italian regions. Surprisingly, Sicily is the third Italian region after Lombardy and Piedmont, although accounting for only small percentages of the European total. This may suggest that regional policy has been more successful than in other southern Italian regions in attracting significant foreign investments in R&D. A more detailed picture of the geographical distribution of ICT research across the European regions under analysis is drawn in Table 2. For each ICT technological sector, Table 2 ranks German, Italian and UK regions on the grounds of the European-owned research developed. Two main findings emerge. First, `higher order regions are confirmed across ICT sectors. This is the case of Bayern and South East (UK), which are top-ranked in any of the ICT technological sectors. Second, in each of these sectors, differences in ranking seem to suggest a distinction between general external economies (e.g. Bayern, Baden-W Èurttemberg, Niedersachsen and Schleswig-Holstein in Germany, Lombardy in Italy, and the South East (UK) and East Anglia in the UK) and localisation economies (e.g. Nordrhein-Westfalen, Hessen and Rheinland-Pfalz in Germany; North West, Piedmont in Italy; and the West Midlands and South West the UK), when also taking into account the whole regional economic activity. 8 In the case of general external regional economies, indigenous and foreign economic activity is spread over a large number of sectors. Conversely, in localisation economies, foreign companies target the sectors of indigenous specialisation. This seems to be confirmed in Table 2, where, while the former always appear to host some percentage of R&D in any of the ICT fields, the latter may be well ranked in some fields and bottom ranked in others according to their sector-specific specialisation. To illustrate the case, European electrical corporations locate some R&D in Baden-W Èurttemberg, Lombardy and East Anglia (although the percentages may vary) in any of the ICT sectors. Conversely, European electrical firms are not attracted at all, for instance, in South West in `telecommunications and in Rheinland-Falz and Piedmont in `special radio systems. Therefore, as far as the European ICT industry is concerned the existence of `knowledge-creating and `knowledge-applying regions are confirmed, over the period under analysis. In this sense, it is crucial to distinguish between `agglomeration economies, based on inter-sectoral co-location, and `location economies, built upon intra-sectoral corporate co-location (McCann, 1995). 5. The Regional Distribution of ICT Technological Development: Does Corporate Technological Co-specialisation Imply Co-location? In order to analyse whether the technological groups identified in the study referred to above (Santangelo, 1998) in each of the three sub-periods ± 1969± 77, 1978± 86

10 282 G. D. Santangelo Table 2. Distribution of US patents attributed to European electrical companies in each of the ICT technological sectors relative to Europe (%), by region and ICT sectors, 1969± 95 Telecommunications Other electrical communications systems Special radio systems Bayern (G) 26.9% Bayern (G) 22.1% South East (UK) 23.9% South East (UK) 8.1% South East (UK) 14.8% Bayern (G) 9.1% Baden-Wurttemberg (G) 3.1% Baden-Wurttemberg (G) 2.8% Baden-Wurttemberg (G) 1.8% Niedersachsen (G) 2.1% Niedersachsen (G) 2.6% East Midlands (UK) 1.3% Nordrhein-Westfalen (G) 1.2% North West (UK) 2.1% South West (UK) 1.3% East Anglia (UK) 1.1% Piedmont (IT) 1.9% Schelswig-Holstein (G) 0.6% West Midlands (UK) 1.0% Hamburg (G) 1.4% East Anglia (UK) 0.6% East Midlands (UK) 0.9% Schelswig-Holstein (G) 1.3% Nordrhein-Westfalen (G) 0.4% Hamburg (G) 0.8% West Midlands (UK) 1.0% Niedersachsen (G) 0.3% Piedmont (IT) 0.8% Nordrhein-Westfalen (G) 0.9% North West (UK) 0.2% Hessen (G) 0.7% East Midlands (UK) 0.8% West Midlands (UK) 0.2% North West (UK) 0.7% South West (UK) 0.8% Wales (UK) 0.2% Lombardy (IT) 0.6% Hessen (G) 0.7% Scotland (UK) 0.2% Rheinland-Falz (G) 0.5% Berlin (G) 0.5% Brandenburg (G) 0.1% Schelswig-Holstein (G) 0.5% East Anglia (UK) 0.5% Bremen (G) 0.1% Berlin (G) 0.3% Scotland (UK) 0.5% Hessen (G) 0.1% Wales (UK) 0.2% Bremen(G) 0.4% Thuringen (G) 0.1% Yorkshire & Humberside (UK) 0.2% Yorkshire & Humberside (UK) 0.3% Yorkshire & Humberside (UK) 0.1% Bremen (G) 0.1% Lombardy (IT) 0.3% Liguria (IT) 0.1% North (UK) 0.1% Rheinland-Falz (G) 0.3% Lombardy (IT) 0.1% Scotland (UK) 0.1% Thuringen (G) 0.3% Berlin (G) Lazio (IT) 0.1% Brandenburg (G) 0.1% Hamburg (G) Sachsen-Anhait (G) 0.1% Wales (UK) 0.1% Mecklenburg-Vorpormmen (G) Brandenburg (G) 0.0% Friuli-Venezia Giulia (IT) 0.1% Rheinland-Falz (G) Mecklenburg-Vorpormmen (G) 0.0% Emilia Romagna (IT) 0.1% Sachsen (G) Emilia Romagna (IT) 0.0% Calabria (IT) 0.1% Sachsen-Anhait (G) Tuscany (IT) 0.0% Mecklenburg-Vorpormmen (G) North (UK) Sachsen (G) Sachsen (G) Piedmont (IT) Thuringen (G) Sachsen-Anhait (G) Friuli-Venezia Giulia (IT) South West (UK) North (UK) Emilia Romagna (IT) Liguria (IT) Liguria (IT) Tuscany (IT) Friuli-Venezia Giulia (IT) Tuscany (IT) Lazio (IT) Campania (IT) Lazio (IT) Campania (IT) Calabria (IT) Campania (IT) Calabria (IT) Sicily (IT) Sicily (IT) Sicily (IT) Total 50.1% 56.8% 41.0%

11 Regional Dispersion of Corporate Research Activity 283 Table 2. Continued Image and sound equipment Semiconductors Office equipment and data processing systems Bayern (G) 12.3% Bayern (G) 27.1% Bayern (G) 17.6% South East (UK) 9.7% South East (UK) 8.7% South East (UK) 7.3% Schelswig-Holstein (G) 6.1% Baden-Wurttemberg (G) 3.6% Baden-Wurttemberg (G) 3.7% Hamburg (G) 3.2% Lombardy (IT) 3.0% Piedmont (IT) 2.3% Baden-Wurttemberg (G) 3.0% Hamburg (G) 1.3% Lombardy (IT) 1.9% Niedersachsen (G) 3.0% Nordrhein-Westfalen (G) 1.2% Niedersachsen (G) 1.7% Nordrhein-Westfalen (G) 1.3% East Midlands (UK) 1.1% Hessen (G) 1.6% East Anglia (UK) 1.2% North West (UK) 1.0% Nordrhein-Westfalen (G) 1.5% Hessen (G) 1.0% Schelswig-Holstein (G) 0.9% Hamburg (G) 1.3% Piedmont (IT) 0.8% Hessen (G) 0.8% North West (UK) 1.3% Rheinland-Falz (G) 0.3% West Midlands (UK) 0.7% Berlin (G) 1.1% East Midlands (UK) 0.3% Sicily (IT) 0.6% Schelswig-Holstein (G) 1.1% South West (UK) 0.3% Rheinland-Falz (G) 0.5% South West (UK) 0.9% Lombardy (IT) 0.3% East Anglia (UK) 0.5% East Midlands (UK) 0.7% Mecklenburg-Vorpormmen (G) 0.2% South West (UK) 0.5% West Midlands (UK) 0.7% North West (UK) 0.2% Niedersachsen (G) 0.4% East Anglia (UK) 0.3% Berlin (G) 0.2% Berlin (G) 0.3% Scotland (UK) 0.2% Brandenburg (G) 0.2% Friuli-Venezia Giulia (IT) 0.2% Emilia Romagna (IT) 0.1% Bremen (G) 0.2% Piedmont (IT) 0.1% Rheinland-Falz (G) 0.1% West Midlands (UK) 0.2% Emilia Romagna (IT) 0.1% Thuringen (G) 0.1% Wales (UK) 0.2% Brandenburg (G) 0.1% Yorkshire & Humberside (UK) 0.1% Friuli-Venezia Giulia (IT) 0.2% Sachsen (G) 0.1% Friuli-Venezia Giulia (IT) 0.1% Yorkshire & Humberside (UK) 0.1% North (UK) 0.1% Sachsen-Anhait (G) 0.1% Lazio (IT) 0.1% Campania (IT) 0.1% North (UK) 0.1% Sachsen (G) 0.1% Calabria (IT) 0.1% Calabria (IT) 0.1% Thuringen (G) 0.1% Bremen (G) Bremen (G) 0.0% Scotland (UK) 0.1% Mecklenburg-Vorpormmen (G) Sachsen (G) 0.0% Tuscany (IT) 0.1% Sachsen-Anhait (G) Campania (IT) 0.0% Sachsen-Anhait (G) Thuringen (G) Brandenburg (G) North (UK) Yorkshire & Humberside (UK) Mecklenburg-Vorpormmen (G) Liguria (IT) Wales (UK) Wales (UK) Emilia Romagna (IT) Scotland (UK) Liguria (IT) Campania (IT) Liguria (IT) Tuscany (IT) Calabria (IT) Tuscany (IT) Lazio (IT) Sicily (IT) Lazio (IT) Sicily (IT) Total 44.9% 52.7% 45.9%

12 284 G. D. Santangelo and 1987± 95 ± are co-located in the regions under analysis, for each of the firms in the clusters an intra-firm cross-region RTA index was calculated at the level of the ICT patent classes 9 as a proxy for the geographical division of labour within the firm. For each European electrical firm (i), the index is defined as the share of US patents granted in region (r) in a patent class (c) relative to the firm s total number of US patents granted in the same patent class in all European regions divided by the share of patents granted to the same firm (i) in the same region (r) in all ICT patent classes considered relative to all patents granted to the same firm (i) in all European regions in all ICT classes considered. Thus, the index can be mathematically formalised as: RTA irc = (P irc /S r P irc )/ (S c P irc /S r S c P irc ) (1) where P irc is the total number of patents granted to firm (i) in region (r) in a patent class (c). It is worth emphasising that r is defined in terms of German, Italian and UK regions only in the numerators and in terms of all other European regions in the denominators. As the index is a comparative measure, high (low) values of RTA irc indicate corporate advantage (disadvantage) in locating research activity in a specific patent class in the region in question. Therefore, the index enables one to evaluate for each European electrical firm the significance of the regional location in a patent class in Europe relative to the significance of the same region in all ICT patent classes considered in Europe. The analysis moves further by testing whether co-specialised firms co-locate their research activity in the technological fields of co-specialisation on the grounds of the intra-firm cross-region RTA index. The criterion adopted to identify eventual co-locations of corporate co-specialised research in ICT requires that, in each of the three sub-periods, at least 50% of the firms in each technological cluster shows the highest RTA irc value in the patent class, upon which the cluster is built, in the same regional location, and that this regional location is common to all other firms in the cluster. Figure 1 illustrates the ICT European-owned corporate groups which were found specialised in the same technological classes (see Santangelo, 1998). In the present study, technological clusters of firms found co-located are reported in a grey background. Before going onto the discussion of the results, it is worth highlighting that in the sub-period 1969± 77, 10 out of 15 firms in the clusters locate R&D in the ICT patent classes of co-specialisation in the regions under analysis; 14 out of 17, in the sub-period 1978± 86; and 16 out of 18 in the sub-period 1987± 95. In the first subperiod, one cluster out of six is co-located, one out of seven in the second, and two out of nine in the last. In all sub-periods, clusters C, C1 and C2 are located in the South East (UK). The reason may be that the patent class/es upon which the clusters are built can be aggregated in the technological sector `special radio systems, in which the UK attracts a greater percentage of foreign European companies by comparison with Germany and Italy as illustrated in Figure 2. This may suggest a strong indigenous technological expertise in the sector in question, which creates a comparative advantage for the UK relative to German and Italian locations. The choice of the South East (UK) as regional host location of corporate technological clusters C, C1 and C2, may well be attributed to the agglomeration of economic activities based on general urbanisation economies. In the last sub-

13 Regional Dispersion of Corporate Research Activity 285 Figure 1. European electrical firms specialised in the same patent class/es and located in the same region (reported in grey background), by sub-period, 1969± 95.

14 286 G. D. Santangelo Figure 2. Share of research activity carried out in the ICT sectors by European foreign electrical companies and located in the three countries under analysis relative to Europe as a whole, 1969± 95. period (1987± 95), the South East (UK) seems also to host the cluster B2.2, built upon patent classes that can be aggregated in the technological sectors `telecommunications and `special radio systems. Although the co-location pattern appears to be driven by the UK attractiveness for `special radio systems, it may well be the case that the South East (UK) emerged as an appealing location for R&D in `telecommunications in the late 1980s± early 1990s (as shown by the recent investments in the region). Therefore, the average trend is that electrical firms specialise in the same patent class/es, although their R&D in the class/es in question is carried out in different regional locations. However, R&D activity in the ICT field as a whole, as well as in each of the six ICT technological sectors (Tables 1 and 2, respectively), seems to be concentrated in a few regional centres of expertise (e.g. Bayern and the South East (UK)), where regional institutions and culture, social division of labour and internal firm organisation are the main competitive advantages. This may suggest that European electrical companies concentrate their research in specific regional locations to benefit from untraded externalities, which may well go beyond intraindustry knowledge spillovers as in the Marshall± Arrow± Romer (MAR) model, involving inter-industry and university knowledge spillovers (Feldman, 1993; Audretsch and Feldman, 1994; Audretsch and Stephan, 1994; Anselin et al., 1997; Audretsch, 2000). As Feldman and Audretsch (1999) argue, diversity across complementary industries showing a common base is a source of greater innovation potential. Therefore, if inter-sectoral geographical links are important within the same industry (Swann and Prevezer, 1996; Baptista and Swann, 1998), they appear to be crucial between industries (Jaffe et al., 1993; McCann, 1995; Audretsch and Feldman, 1996; Blomstr Èom and Kokko, 1996; Malmberg, 1996; Bopp and

15 Regional Dispersion of Corporate Research Activity 287 Gordon, 1997; Storper, 1997). Knowledge externalities in industries based on new economic knowledge are crucial. In this sense, additional increases in concentration of production of innovative activity seem to lead to greater dispersion of innovative activity as new ideas need new space (Audretsch and Feldman, 1995). Following Camagni (1988), this seems to be all the more true in the current techno-socioeconomic paradigm where the creation of spatial synergies is amplified by the new complex technologies adopted. In this context, MNCs link together the localised processes of knowledge by ranking different centres when selecting their locations (Malmberg et al., 1996). However, if the impact of corporate activities is recognised, there is still room for policy in infrastructure and education in order to facilitate the process of local growth through FDI strategies. On the host region perspective, Vence-Deza (1996) places great emphasis on the importance of diversity and complementarity in boosting local expertise in lagging regions. The magnitude of spillover effects largely depends on local capability and competition which generate local higher value added (e.g. concentration of labour skills and linkage advantage). This case is clearly illustrated in Silicon Valley and Route 128, where urban industrial clusters generate inter- and intra-industry externalities, which are enhanced by proximity to early users. However, knowledge still flows more smoothly within national borders as knowledge spillovers occur more easily between regions with similar or complementary specialisation. In this sense, the European System of Innovation is characterised by several centres (which are not contiguous), where innovative activity is polarised especially in high-tech sectors (CaniÈels, 1998). 6. Conclusions This paper investigates whether the clusters of European electrical companies identified in a previous study (Santangelo, 1998) develop technological innovation in the patent classes of common specialisation in the same regional location. It also provides some preliminary empirical evidence on the spatial development of ICT research activity in the European electrical industry at regional level in the locations under analysis. In analysing whether co-specialisation occurs simultaneously in the European corporate development of ICT technologies, it was found that European electrical companies do not locate research activity in the fields of co-specialisation in the same region. However, they seem to concentrate geographically their R&D in ICT technologies. The findings seem to suggest that companies co-locate their research activity in order to enjoy untraded externalities and complementary diversity. The location of R&D laboratories in a dynamic regional environment enables the whole MNC to source a broad knowledge complementary to its technological path. Therefore, MAR spillovers do not seem to be a decisive factor influencing corporate location strategies. Conversely, European ICT companies enhance their competitive advantage in their core areas of technological expertise in heterogeneous dynamic environments through geographically dispersed intra-firm networks. Therefore, diversity across complementary technological sectors seems to drive corporate location strategy. Inter-sectoral co-location provides a source of greater innovation potential. In this sense, the paper argues that the increased concentration of production of innovative activity in `higher order regional locations promotes a greater dispersion of corporate R&D in specific kinds of technologies. In other words, if European corporate R&D in ICT as a whole appears to be concentrated

16 288 G. D. Santangelo in a few regions, the analysis also reveals a pattern of geographical dispersion as far as specific ICT technological fields are concerned. The success of innovative activity appears to be increasingly more embedded in local centres of expertise. It follows that intra-european corporate development of ICT technology cannot be explained by the classical argument of corporate strategies aimed at the mere reduction of production costs. Rather, value added factors embedded in the social system determine European electrical MNCs decisions in the choice of regional location. The existence of location hierarchies in ICT research in Europe may be due to key factors (i.e. local infrastructure, culture and business capabilities) attracting foreign corporations. The regions more appealing to foreign research in ICT are also the most dynamic local economic centres in the countries under analysis. As the building of local expertise is a pathdependent process, multinationals investments widen the gap between core and peripheral regions as a result of vicious and virtuous circles. This implies that the competition for attracting foreign R&D is highly uneven as established regional systems are more likely to host cutting-edge innovative activity. Nonetheless, the process is not irreversible as regions unable to adapt their institutions to cope with new opportunities may pass into relative decline and lagging regions may successfully open up to compensate with opportunities for development (Metcalfe, 1996). The findings of the paper also confirm the importance of the regional dimension in the study of technological change. National systems-based analyses fail to provide a complete picture of intra-border phenomena in an age where social and economic cohesion across European regions is a major political target. This confirms Nelson s (1984) argument that, although national policies still impact on the development of high-tech industries, the globalisation process calls for a more integrated world. Notes 1. The Structural Funds are the European Regional Development Fund (ERDF), the European Social Fund (ESF), the European Agriculture Guidance and Guarantee Fund (EAGGF) and the Financial Instruments for Fisheries Guidance (FIFG). 2. New economic geography has emerged from the (re)discovering of geography by economists willing to persuade their colleagues that `space matters (Krugman, 1991a). Although the theoretical links claimed by new economic geographers with traditional location theory (LÈosch, 1954), recent critiques have pointed out that this `geographical turn is a quite distinct methodological and epistemological genre from tradition location theory and its further development, namely regional science and economic geography. For a critical survey see Martin (1999). 3. Mergers and acquisitions are largely recognised in the data through the practise in most groups of centralising the patent application procedure in the parent company. Accordingly, the consolidation of affiliates research activity into corporate groups does not affect the information recorded in the patent database because of the standardised structure of the patent document. In other important cases affecting the ultimate ownership of significant numbers of patents, the change in ownership structure is incorporated into the organisation of the data, which involves in some cases the creation of a new corporate group and, in others, the expanded consolidation of groups with newly acquired subsidiaries. 4. A list of the six ICT technological sectors is provided in the Appendix (Table A1). 5. A list of the ICT original technological patent classes is provided in the Appendix (Tables A1a and A1b). 6. Some of the 23 electrical corporations were dropped from the analysis on the grounds of the relatively small number of patents in the technological patent classes considered, the outcome being that the firms in the sample vary from one sub-period to another. As listed in Table A2, 20 firms were examined in the sub-period 1969± 77, whilst 21 and 19 were taken into account in the two later subperiods, respectively

17 Regional Dispersion of Corporate Research Activity These territorial levels of analysis have also been adopted by Cantwell and Iammarino (1998, 1999) in the case of Italy and the UK, respectively, and by Cantwell and Noonan (1999) in the case of Germany. 8. The distinction between these two types of agglomeration economies is obviously not clear-cut. However, the categorisation of these regions proposed here takes into account other studies (e.g. Linders, 1985; Heidenreich and Krauss, 1998; Iammarino et al., 1998; Iammarino and Santangelo, 1998) and EU publications (European Communities, 1993a, b) based on other economic indicators. 9. A selection of ICT patent classes is considered in this paper. The analysis focuses on the patent classes labelling the technological clusters in Figure 1 in each of the sub-periods (1969± 77, 1978± 86 and 1987± 95) rather than all the ICT patent classes recorded in the Reading database. 10. References Amin, A., Bradley, D., Howells, J., Tomaney, J.J. and Gentle, C., ª Regional Incentives and the Quality of Mobile Investment in the Less Favoured Regions of the EC,º Progress in Planning, 1994, 41. Amin, A. and Goddard, J.B., ª The Internationalization of Production, Technological Change, Small Firms and Regional Development: An Overview,º in A. Amin and J. Goddard, eds, ittechnological Change, Industrial Restructuring and Regional Development. London: Unwin Hyman, Amin, A. and Thirft, N., Globalisation, Institutions, and Regional Development in Europe. Oxford: Oxford University Press, Anselin, L., Varga, A. and Acs, Z., ª Entrepreneurship, Geographic Spillovers and University Research: a Spatial Econometric Analysis,º ESRC Centre for Business Research, University of Cambridge Working Papers, No. 59, Antonelli, C., ª Localized Technological Change in the Network of Networks: the Interaction between Regulation and Evolution of Technology in Telecommunication,º Industrial and Corporate Change, 1945, 4(4), pp. 737± 54. Archibugi, D. and Michie, J., ª The Globalisation of Technology: a New Taxonomy,º Cambridge Journal of Economics, 1995, 19, p Archibugi, D. and Pianta, M. The Technological Specialisation of Advanced Countries. Dordrecht: Kluwer Academic Press, Audretsch, D.B., ª Knowledge, Globalisation and Regions,º in J.H. Dunning, ed., Regions, Globalization and the Knowledge Based Economy. Oxford: Oxford University Press, Audretsch, D. and Feldman, M.P., ª Knowledge Spillovers and the Geography of Innovation and Production,º Centre for Economic Policy Research Discussion Papers, No. 953, London, Audretsch, D.B. and Feldman, M., ª R&D Spillovers and the Geography of Innovation and Productionº, American Economic Review, 1996, 86(3), pp. 630± 40. Audretsch, D.B. and Stephan, P.E., ª How Localized are Networks in Biotechnology,º Wissenschaftszentrum Berlin fèur Sozialforschung, FS IV 94± 9, Audretsch, D.B. and Stephan, P.E., ª Company ± Scientist Locational Links: the Case of Biotechnology,º The American Economic Review, 1996, 86(3), pp. 641± 52. Baptista, R. and Swann, P., ª Do Firms in Clusters Innovate More?º Research Policy, 1998, 27(5), pp. 525± 40. Bartlett, C.A. and Ghoshal, S., Transnational Management ± Text, Cases and Reading in Cross-Border Management, Chicago: IRWIN, Begg, I., Fagerberg, J. and Guerrieri, P., ª Conclusions and Policy Implication,º in J. Fagerberg, P. Guerrieri and B. Verspagen, eds. The Economic Challenge for Europe: Adapting to Innovation-Based Growth. Edward Elgar: Cheltenham, Blomstr Èom, M. and Kokko, A., ª Multinational Corporations and Spillovers,º Centre for Economic Policy and Research, No. 1365, Bopp, R. and Gordon, P., ª Agglomeration Economies and Industrial Economic Linkages: Comment,º Journal of Regional Science, 1997, 17(1), pp. 125± 27. Camagni, R., ª Functional Integration and Locational Shifts in New Technology Industry,º in P. Aydalot and D. Keeble, eds, High Technology Industry and Innovative Environments: the European Experience. London: Routledge, CaniÈels, M., ª The Geographic Distribution of Patents and Value Added Across European Region,º MERIT Working Papers, No. 2± 98± 004, 1998.

18 290 G. D. Santangelo Cantwell, J.A., ª The Reorganisation of European Industries after Integration: Selected Evidence on the Role of Multinational Enterprise Activities,º Journal of Common Market Studies, 1987, XXVI (2), pp. 127± 51. Cantwell, J.A., ª Introduction: Transnational Corporations and Innovatory Activities,º in J.H. Dunning, ed., Transnational Corporations and Innovatory Activities. London and New York: Routledge, 1994, Cantwell, J.A., ª The Globalisation of Technology: What Remains of the Product Cycle Model?º Cambridge Journal of Economics, 1995, 19, pp 155± 74. Cantwell, J.A. and Iammarino, S., ª MNCs, Technological Innovation and Regional Systems in the EU; Some Evidence in the Italian Case.º International Journal of The Economics of Business, 1998, 5(3), pp. 383± 407. Cantwell, J.A. and Iammarino, S., ª Multinational Corporations and the Location of Technological Innovation in the UK Regions,º Regional Science, 1999, 34(4), pp. 317± 32. Cantwell, J.A. and Janne, O.E.M., ª Technological Globalisation and Innovative Centres: the Role of Corporate Technological Leadership and Locational Hierarchy,º Research Policy, 1999, 28, pp. 119± 44. Cantwell, J.A. and Noonan, C., ª Multinational Corporations and the Location of Technology Innovation in the German Regions.º University of Reading, mimeo, Caracostas, P. and Soete, L., ª The Building of Cross-border Institutions in Europe: Towards a European System of Innovation,º in C. Edquist, ed. Systems of Innovation. London: Pinter, Castells, M., ed. High Technology, Space and Society. Beverlyhills: Sage Publications, Champion, T., Mù nnesland, J. and Vandermotten, C., ª The New Regional Map of Europe,º Progress in Planning, 1996, 46, pp. 1± 89. Cowan, R. and Foray, D., ª The Changing Economics of Technological Learning: Implications for the Distribution of Innovative Capabilities in Europe,º in X. Vence-Deza and J.S. Metcalfe, eds, Wealth from Diversity. Dordrecht: Kluver Academic Publisher, European Communities, Portrait of the Regions, Vol. 1. ± Germany, Benelux, Denmark. Luxembourg, 1993a. European Communities, Portrait of the Regions, Vol. 2. ± France, United Kingdom Ireland. Luxembourg, 1993b. Eurostat, Nomenclature of Territorial Units Statistics. Luxembourg, Fagerberg, J., ª Heading for Divergence?º Journal of Common Market Studies, 1996, 34(3), pp. 431± 48. Feldman, M., ª An Examination of the Geography of Innovation,º Industrial and Corporate Change, 1993, 2, pp. 451± 70. Feldman, M., The Geography of Innovation. Dordrecht: Kluwer Academic Publishers, Feldman, M. and Audretsch, D.B., ª Innovation in Cities: Science-based Diversity, Specialization and Localized Competition,º European Economic Review, 1999, 43, pp. 409± 29. Feldman, M. and Florida, R., ª The Geographic Sources of Innovation: Technological Infrastructure and Product Innovation in the United States,º Annals of the Association of American Geographers, 1994, 84(2), pp. 210± 29. Feller, I., ª Invention, Diffusion and Industrial Location,º in L Collins and R. Walker, eds, Location Dynamics of Manufacturing Activity. London: John Wiley & Sons, Florida, R., ª Toward the Learning Region,º Future, 1995, 27(5), pp. 527± 36. Florida, R., ª Capitalism, Globalism and Regionalism,º paper presented at the European Union Conference, Brussels, September Griliches, Z., ª Patient Statistics as Economic Indicator.º Journal of Economic Literature, 1990, XXVIII, pp. 1661± 707. Heidenreich, M. and Krauss, G., ª The Baden W Èurttemberg Production and Innovation Regime: Past Success and New Challenges,º in H.J.J. Braczyk, P. Cooke and M. Heidenreich, eds, Regional Innovation Systems. London: UCL, Hepworth, M., ª The Geography of Technological Change in the Information Economy,º Regional Studies, 1986, 20(5), pp. 407± 24. Howells, J., ª Regional Systems of Innovation?º in D., Archibugi, J. Howells and J. Michie, eds, Innovation Policy in a Global Economy. Cambridge University Press, Iammarino, S., Prisco, M.R. and Silvani, A., ª The Geography of Production and Innovation: How Regional ª Stylesº Play in the Global Scenario,º The Regional Science Review, 1998, 18. Iammarino, S. and Santangelo, G.D., ª FDI and Regional Attractiveness in the EU Integration Process: Some Evidence for the Italian Regions,º paper presented at the Second European Urban Regional Studies Conference, University of Durham, September, 1998.