GaWC Research Bulletin 391

(Z)

How Manufacturing Industries Connect Cities across the World: Extending Research on ‘Multiple Globalizations’

S. Krätke^*

Abstract

This article concentrates on a macro-level analysis of inter-urban linkages in a ‘world city network’. Empirical research on the formation of a world city network has mostly concentrated on global service providers. Yet glo­bally operating manufacturing firms as well choose distinct urban re­gions all across the world as locational anchoring points. This contribution presents the first glo­bal-scale analysis of how manufacturing firms connect cities across the world (in 2010), utilizing the methodical approach of ‘social network analysis’. We claim that the network structure of distinct industrial subsectors within the global urban system needs to be analysed in order to detect the differing ’sectoral profiles’ and nodal centralities of cities functioning as geographic hubs of transnational production networks. The analysis covers 120 top global firms from three manufacturing subsectors. Subsequently, we compare the nodal centralities of cities included in these subsectors’ global networks with the GaWC research on the producer services sector that has been at the centre of previous analyses of the world city network. The comparison reveals the cities’ differing positioning within ‘multiple globalizations’. In this way, the article aims at an extension of world city network research.

Key words: world city network, globalizing cities, global production networks, global value chains, network analysis

Introduction

Today, the cities’ economic development and internal structure is increasingly affected by global influences, particularly the cities’ external relations. Hence it is increasingly difficult to analyse the ‘inner workings’ of the city endogenously, from what is happening inside its conventionally defined administrative boundaries. We are facing the formation of new transnational urban spaces in cities all across the world. With regard to the worldwide urban system, globa­lization doesn’t simply lead to the formation of a small group of outstanding ‘global cities’ that are functioning as command and control centres of the world economy (Sassen 1991). Rather, we are facing a continued extension of transnational economic networks that include more and more cities both in the global North and South in the complex fabric of a ‘world city network’ (Taylor 2004). Thus, I prefer to speak of ‘globalizing cities’, whose economic development prospects are shaped by their specific positioning within global economic network relations and capital flows. As John Friedmann (1986) has emphasized, in the contemporary era of globalization the form and extent of a city’s integration with the world economy and the functions assigned to the city in the global spatial division of labour will be decisive for its development level and internal structural changes. These local restructuring pro­cesses include (a) the selective formation of transnational urban spaces within the city (in terms of urban zones that are shaped by the activities of transnational actors), (b) a changing sectoral composition of urban economies that is combined with new functions and industrial upgrading processes, and (c) the often two-sided impact of economic restructuring in globalizing cities on the cities’ social and spatial fabric (cf Sassen 2000, Marcuse and van Kempen 2000).

This contribution starts from a short discussion of basic approaches to ‘global urban ana­lyses’. Subsequently, the article concentrates on a macro-level analysis of the global urban system by investigating inter-urban linkages in a ‘world city network’ that comprises a multitude of cities both in the global North and South. This article presents selected findings of recent research on the manufacturing industries’ global organizational networks that link ci­ties across the world. Interestingly, the ‘ranking’ of globally connected urban centres of manufacturing activity differs considerably from the well-known rankings of global service centers. Thus we have to be aware of ‘multiple globalizations’ in the world city network.

Approaches to research on cities in globalization and the world city network

With ‘globalization’ having become a key term in the debate on the economic and social changes taking place in contemporary societies, the interrelations between globalization and urban development has become a central theme of urban research. As a result, ‘new’ rankings of cities on a global scale have been established (Short et al. 1996, Hall 2001). By investigating particular cities’ global functions and their impact on the city’s economic and socio-spatial restructuring, Sassen (1991) detected New York, London and Tokyo as the prime centres of a global urban hierarchy. These global cities are functioning as major command and control centres in a transnationally organized economy on the basis of their strong local concentration of global firms of the ‘advanced producer services’ and the socalled FIRE sector (finance, insurance, and real estate). Yet Sassen’s approach focused on comparative case studies of selected cities and thus couldn’t detect the connectivity of cities on the level of a truly ‘global scale’ urban analysis. Today, many urban researchers consider that globa­lization proceeds primarily from urban centres of activity and that more and more cities are actively or passively integrated into processes of globalization.

A second line of global city research concentrates on the inter-urban linkages in a global urban system which includes a vast number of cities beyond the small group of ‘leading’ global cities (such as those mentioned before). Representative for this line of research is the work of Peter Taylor (2004) and the GaWC (‘Globalization and World Cities Study Group’). Starting point is the thesis that connectivity in the world city network is being created through the multi-local office networks of global service providers. A city where a large number of global service firms’ head offices and branch offices are located, becomes a central node in the world city network (Beaverstock, Smith and Taylor 1999). The GaWC approach resulted in the identification and classification of corporate service centres of the global eco­nomy.

The latest version of the GaWC analysis co­vers the office networks of 175 global service firms across 526 cities worldwide in 2008 (Taylor et al. 2010). The selection of firms includes 75 financial sector firms and 100 producer service firms (of the subsectors accounting, advertising, law and management consultancy). The results of this analysis indicate that the globalization of advanced producer services has further continued, with the expansion of offices in many cities and the extension of office networks into new ones. This might be interpreted as an expansion and increasing integration of the world city network. The analysis confirms the premier position of London and New York at the top level of global service centres and the rapid advance of cities in emerging markets such as Shanghai, Beijing, and Mumbai. Overall, the world city network analysis demonstrates the inclusion of many cities within contemporary globalization processes. It highlights the geographically uneven spread of globalization as well as the rise of new urban nodes of globally connected business activity.

From a critical point of view, however, it is evident that in both approaches mentioned before, global city research has tended to concentrate on the respective cities’ function as financial centres and centres for the provision of specialised corpo­rate services. Nonetheless, we should be aware that the prominent examples of global finance and service centers such as London and New York represent a quite unique type of urban economic structure. They represent the world centres of today’s financial business that dominates the socalled ‘advanced’ service sector. The current failure of neoliberal finance-dominated capitalism, which has led to a global economic crisis that does not only affect the countries of the desaster’s origin, but also economic and social development in the global South, demonstrates that there is urgent need to cut back and re-regulate the ‘dealer economy’ of the financial sector (Krätke 2011) and to redirect economic development trajectories on real sector manufacturing activity, industrial upgrading and technological change, as well as social innovation (particularly with regard to employment standards and urban living environments). Taking into account the structural diversity of the worldwide urban system, we can empirically detect that many cities in the global North and South are linked to transnationally extended production networks and continue to play an important role in the development of manufacturing industries (Henderson et al. 2002; Dicken 2007). Transnationally networked cities are functioning as major locations in the global value chains of manufactu­ring industries (Derudder and Witlox 2010; Krätke, Wildner and Lanz 2012). The transnational connectivity of global manufacturing firms’ locations should thus be considered as a relevant aspect of ‘globalizing cities’ (see also Scott 2001).

The concept of global value chains (Gereffi and Korzeniewicz 1994; Sturgeon 2000; Gereffi 2006) interprets globalized production as a series of cross-border transactions between the different production stages of commodities as well as between different corporate establishments. The chain links are geographically distributed over a series of locations and entail  cross-border flows between the respective spatial nodes of activity. Although the literatures on globalizing cities and on global value chains / production networks are sharing a similar general conceptualization of economic space in terms of the discontinuous territoriality of global network relations, they have not yet been constructively integrated (first attempts at an integration have been collected in Derudder and Witlox 2010). Yet the concepts of global value chains and global production networks offer an approach to meet the demand for a multi-sectoral perspective on the economic geography of ‘globalizing cities’.

We claim that all cities included in the world city network are characterized by specific profiles of globally connected economic functions. The world city net­work includes global cities focusing on advanced producer services and the financial sector in particular, as well as many other cities with differing profiles of their globally connected activities (cf Krätke 2003; Krätke and Taylor 2004).

An alternative approach to the analysis of the world city network, which makes use of a different methodical concept of network analysis and refers to a broader range of economic subsectors as compared to the GaWC research, has been presented by Alderson and Beckfield (2004) and, most recently, by Wall and v.d. Knaap (2011). These contributions claim that the multinational firm – regardless of its particular activity branch or subsector – stands at the heart of network relations connecting the global urban system (see also the debate between Taylor, 2006, and Alderson & Beckfield, 2006). Consequently, Wall and v.d. Knaap (2011) include in their analysis the top 100 multinational corporations which stem from various industrial sectors. Wall and v.d. Knaap advance our understanding of the global urban system by analysing the hierarchical differentiation of the included multinationals’ organizational and geographical structure. The distinction of several hierachy-levels within the corporate structure demonstrates that a prominent position in the global firms’ locational network is not only occupied by the top level headquarter city, but also by cities that are situated at the intersection of different hierachy-levels: these cities take on a significant role in the direction of capital flows, since they are hosting subsidiaries which themselves control further subsidiaries that are located in other urban regions. In total, the analysis reveals the nodal centralities within the ‘all industries sector’ network and the ‘producer services sector’ network. Despite some differences concerning the positioning of individual cities, Wall and v.d. Knaap (2011) also show that a strong correlation exists between both sectoral networks (particularly at the level of top rank urban nodes of global connectivity).

However, the contribution of Wall and v.d. Knaap has a weak point concerning the delimitation of economic subsectors: The analysis is based on a selection of global firms from both the service and manufacturing sectors, and thus does not reveal the role of particular manufacturing industries in the formation of the global urban network. The ‘Fortune Global 100’ multinationals (2005) selected by Wall and v.d. Knaap are composed of 43 finance and service sector firms (also including a small number of global retail groups), 19 firms of the utilities sector (particularly oil and gaz multinationals), and only 38 manufacturing industry firms of medium high technology subsectors (such as the automotive industry) and high technology subsectors (such as the information technology industry). Consequently, the results are to a large degree shaped by the major influence of the included finance and service sector firms. Thus the detected strong correlation between the ‘all industries sector’ network and the ‘producer services sector’ network could be expected.

By contrast, we claim that the network structure of distinct industrial subsectors within the global urban system needs to be analysed in order to detect the differing ’sectoral profiles’ and nodal centralities of cities functioning as geographic hubs of transnational production networks. In this article, we try to advance ‘world city network’ research by presenting the results of a network analysis covering 120 top global manufacturing firms from three distinct industrial subsectors (see below). This research started from the thesis that the inter-urban network formed by global manufacturing firms will differ from the network formed by global firms of the advanced producer services. The question is: which particular cities are functioning as outstanding centres of global connectivity in the manufacturing industries, and which major differences concerning the sectoral profiles of globalizing cities can be detected? The analysis might e.g. reveal cities which are functioning as major hubs of distinct manufacturing industries’ global networks, but do not qualify to be included in the group of prominent urban centres of global services. The recent contribution of Wall and v.d. Knaap (2011) falls short of detecting these nodes due to the aggregation of manufacturing and service firms in an ‘all industries sector’. The article’s second section deals with the methodical approach and presents findings of the network analysis for two subsectors. In the third section, we compare the nodal centralities of cities included in these subsectors’ global production networks with the GaWC research findings concerning the finance and producer services sector. The comparison reveals differing sectoral profiles of globalizing cities and thus might advance our understanding of ‘multiple globalizations’ in the world city network.

Inter-urban linkages of manufacturing firms in the contemporary world city network

We start from the thesis that global networks of the manu­facturing sectors connect cities across the world. The expansion and diffusion of ‘industrial urbanism’ on a global scale which is led by the formation of global production networks with local anchoring points in urban regions all over the world is a most distinctive feature of the current phase of globalization (cf Soja 2000). My research is based on the assumption that globalizing city regions are exerting distinct functions in global value chains (such as e.g. distinct manufacturing functions located in the urban region’s industrial areas, and the supply of specific producer services for the ma­nagement of these global value chains). Second, many relevant nodes or chain links of global value chains are located in ‘globalizing’ cities of the global South (Krätke, Wildner and Lanz 2012). The urban regions of Bangkok and Munich, for example, are interconnected by their functioning as spatial anchoring points of the BMW firm’s transnational production network. In similar ways, the global production network of the car manufacturer VW includes the global city regions of Sao Paulo, Mexico City, Shanghai etc.

Methodical Approach

In order to present an empirically based account of the worldwide connectivity of cities in the sphere of manufacturing industries, a network analysis was performed. According to Taylor (2004), the world city network can be characterized as an ‘interlocking network’ that allows relations between cities to be measured through data collected on firms. An interlocking network denotes a specific type of network (cf. Knoke and Kuklinski 1982) that consists of a nodal level - the cities - and the sub-nodal level of firms. The cities are connected through actors on the sub-nodal level (the global firms’ establishments). They are embedded in networks of corporate relationships that enclose transnational intra-firm transactions such as knowledge flows and capital flows. The global urban network analysis, however, cannot reveal actual flows of information, knowledge, and capital on the nodal and sub-nodal level – it rather detects the ‘channel system’ and the nodal intersections of potential flows. In this article, ‘capital flows’ refer to capital investments within a particular manufacturing sector, such as FDI between corporate establishments in different countries, whereas other types of flows such as those in the realm of the financial sector are not dealt with.

The network analysis contains three subsectors of manufacturing industries: the ‘automotive industry’, the ‘technology hardware & equipment’ sector (particularly including information & communication technology and semiconductors) - taking its label from the ‘Forbes 2000’ database, and the ‘pharmaceutical and biotechno­logy industry’. These three subsectors reflect different periods of economic development viewed from a long wave perspective. The automobile network is in a mature industry, whereas the high tech equipment and the IT sector reflects a rapidly developing sector moving towards maturity. This subsector is also characterized by a high degree of knowledge intensity and innovation. This is even more so the case in the biotech sector, which is in the early parts of its life cycle and highly knowledge intensive. These differences in economic structure may have implica­tions for the networks under consideration. First, the original centres of the respective in­du­stries might be anchored in different parts of the world, such as e.g. California and South and Southeast Asia in the case of the IT sector, Germany and the North-Eastern region of the USA (particularly the federal states of Michigan and Wisconsin) in the automotive industry. Second, the highly knowledge intensive manufacturing industries might share some locational priorities with the knowledge intensive branches of the service sector and thus tend to concentrate in metropolises with a highly developed and diversified knowledge basis.

For each subsector, the 40 largest global firms were identified according to the ‘Forbes 2000’ listing, which contains in total the world’s 2000 largest firms of all sectors in 2010 (see the listing of included firms in the Appendix).In the next step, the prominent firm data base ‘corporate affiliations’ (covering more than 1 million corporations) was utilized in order to detect the selected firms’ national and international organizational network and its linkages within the global urban system. The firm data base offers information on the multi-level corporate hierarchy of the firms included - such as the parent company, divisions and regional headquarters, subsidiaries, affiliates and joint ventures - and on the location of these firm units. The multi-level hierarchy means that e.g. a parent company located in city A can have a division in city B, which directs a subsidiary firm in city C etc. In this way, e.g. the automotive industry’s 40 largest firms contain in total more than 1000 enterprise units, thereof 881 subsidiaries. Taking the three subsectors together, the analysis includes 4512 firm units that constitute the organizational network of 120 parent companies. In the procedure of registering individual firm units, the financial subsidiaries of manufacturing firms were not included in the analysis, since the focus of this research is geared towards capital-based interconnections in the realm of manufacturing industries’ value chains.

The locations of the registered enterprise units are distributed across the world in a total of 544 cities (urban regions) in 104 different countries. The locational addresses of included firm units were assigned to the respective urban region. The urban region of large cities and metropolises was delimited to cover a radius of 50 km around the city core. For extraordinarily extended metropolitan regions such as e.g. the cases of Tokyo, London and New York, the radius has been slightly enlarged. In many cases, the global manufacturing firms’ corporate establishments are located in the fringe area of the respective urban region, contributing to the spatial expansion of major cities’ economic territory. We have also to note that in some tabular representations below the city label of ‘San Jose’ in fact represents the whole extended urban agglomeration of the socalled ‘Silicon Valley’.
The network analysis detects the geographic destination and strength of the organizational links within the global urban system. As a result, the analysis reveals the positioning of particular cities within the global production network of the respective manufacturing industry subsector. This globally extended production network might be interpreted as an organized system of channels for capital flows between cities.  

Presentation of Research Findings

TThose urban regions that show a particularly large number of network links can be characterized as outstanding centres of the respective industry on a global scale. This ranking refers to the nodal centralities of cities included in distinct subsectors’ global networks. The presentation of research findings starts with an account of the locational centres of the automotive industry at the macro-level of the global urban system. We choose the automotive industry as a starting point, since this particular subsector presents the most contrasting example as compared to city-rankings focusing on global producer service functions. We will also deal with the strength and direction of specific inter-city connections in two subsectors of manufacturing in order to demonstrate properties of the network structure (or ‘topology’) that are not sufficiently recognized in the quantitative accounts of nodal centralities. Subsequently, the comparative ranking of urban regions according to their control capacities  and attractive power in different subsectors of the manufacturing industries will be presented as the centerpiece of the research findings.

Figure 1: Cities in Global Networks of the Automotive Industry, 2010. Urban Regions’ Control Capacity in Terms of their Outdegree and Attractive Power in Terms of their Indegree

Moreover, due to the distinction between control capacity as measured by outgoing links (outdegrees) and attractive power as measured by incoming links (indegrees), the cartographic representation reveals that a comparatively large number of cities in the respective world regions are functioning as target destinations of capital-based interconnections, whereas a significantly smaller number of cities has the potential to command and control capital flows in the automotive industry. Urban regions that contain a strong command and control capacity in the automotive industry’s global production network show a quite uneven geographic distribution on the world-scale. In North America, we find two prominent urban regions in the eastern part of the USA. In Europe, several cities in Germany and one city in Sweden stand out, and in Asia, the cities which have a particularly strong control capacity concentrate in Japan.

A first impression of the strength and direction of specific inter-city connections (as a relevant aspect of network structure or ‘topology’) in the global automotive industry can be won by representing the links that appear on the firms’ upper hie­rar­chy level (see figure 2). We don’t get a comprehensive picture of all relations here, since a considerable number of inter-city connections have been registered on the second and third hierarchy le­vels. Yet this selection offers a ‘less overloaded’ graphical account, in which the automotive industry’s major regional centres of ‘power’ become visible: The strongest interlinks o­n hie­rarchy level 1 stem from the Japanese cities Tokyo and Nagoya. Many of these links are directed towards other cities in Asia (such as Bangkok, Beijing and New Delhi). However, there are also strong outgoing links to the USA (Los Angeles), the UK (London), and Europe (e.g. Brussels, Frankfurt-Main, Amsterdam). In the USA, Detroit appears as an outstanding centre with strong links towards urban regions in Latin America (Sao Paulo, Mexico City, Caracas) and in Asia (particularly Seoul). Other regional centres with strong outdegrees include Mumbai (India), Paris, Milan, and the German cities Stuttgart, Munich and Wolfsburg.

Figure 2: Strength of specific Inter-City Connections in the Global Automotive Industry, 2010. Representation of Relations on the Firms’ Upper Hierachy Level.

Figure 3 demonstrates the the strength of specific inter-city connections in the global automotive industry for all corporate hierarchy levels. This representation offers a more comprehensive picture of the outstanding channels of capital flows in the global automotive industry, including an additional number of particularly strong inter-city connections (such as e.g. the links between Los Angeles and New York, Milan and Turin, Gothenburg and Paris). The graphic representations of network structure reveal that the ‘space of flows’ constituted by a particular industry’s global network might be conceived as a structured fabric of channels which contains some particularly intensive links and a number of major branching points (interfaces).

Figure 3: Strength of specific Inter-City Connections in the Global Automotive Industry, 2010. Aggregated Relations on all Hierachy Levels.

Figure 4: Strength of specific Inter-City Connections in the Global ‘Technology Hardware & Equipment Industry’, 2010. Aggregated Relations on all Hierachy Levels.

However, the comprehensive graphic representations of network structures and particularly strong inter-city links on a transnational level do not enable a more accurate interpretation of individual cities’ positioning in the network of globally operating manufacturing firms. As we will show in the next section, the nodal centrality of individual cities does not necessarily cor­respond to the strongest dyadic inter-city links detected (e.g. in figures 2 and 3, the Tokyo-London link is not a ‘reliable’ indicator of London’s nodal centrality in the respective industry’s network). Thus, in the next section we proceed to a ‘ranking’ of urban regions according to their nodal centralities.

Ranking of Cities according to their Role in Different Manufacturing Sector’s Global Networks

This section presents a more detailed account of the research findings in terms of a comparative ‘ranking’ of urban regions according to their control capacity and attractive power within two different manufacturing industries’ global networks. This ranking of cities doesn’t refer to specific functions of individual corporate establishments or subsidiaries, or to the respective firms’ output and employment figures in a particular urban region. The ranking solely refers to the positioning of individual cities in the capital-based production network of globally operating manufacturing firms.

The tabular representation of the city ranking in the automotive industry subsector(see table 1) demonstrates the top 60 of a total of 339 cities included in this subsector’s network. The table distinguishes between ‚outdegrees’ (outgoing links) and ‘indegrees’ (incoming links). The measurement of degrees takes the differing strength of individual inter-city links into account - i.e. the degrees of connectivity also reflect multiple links between distinct pairs of cities. Furthermore, for each city the difference between outdegree and indegree is indicated. Urban regions with a strong ‘surplus’ of outdegrees (such as e.g. Tokyo) are primarily functioning as command and control centres of capital flows in the automotive industry. Urban regions with a strong ‘surplus’ of indegrees (such as e.g. Toronto and Los Angeles) are primarily functioning as platform- or bridgehead locations for the supply of distinct market regions, the utilization of local production capacities, or the access to specific knowledge resources and innovation capabilities. We might say that these cities possess of a significant ‘attractive power’ in the automotive industry’s global networks. Of course, there are also urban regions which show a rather ‘ba­lanced’ relation of outdegrees and indegrees (such as e.g. New York), so that no definite ‘primary’ functional designation can be assigned to these cases.

Outdegree and indegree represent relevant measures of centrality in the network analysis context. However, the network analysis method offers a variety of more complex measures of centrality, such as e.g. the socalled ‘betweenness-centrality’ (Jansen 2003), which is also reported in the ta­bular representation for each city included. The measure of ‘betweenness-centrality’ indicates the extent to which an urban region is positioned as an intersectional node within the connecting links of all other cities included in the network, and thus has a ‘mediating position’ in the capital flows that run through the overall network structure. The highest betweenness-centrality is recorded for the urban regions of Tokyo and Detroit, which also corresponds to their outstanding position in terms of other (degree-based) measures of centra­lity. Comparatively high values of betweenness-centrality are also indicated for the urban regions of Go­thenburg, Milwaukee, Frankfurt-Main, Stuttgart and Hannover, which are all positioned on the upper-level ranks of outdegrees. However, the presentation of research findings will focus on the degree-based measures of centrality, which can be interpreted more easily.

Table 1: Urban Nodes of Global Networks in the Automotive Industry (based on the organizational Networks of the Sector’s 40 largest Corporations, 2010)

The tabular respresentation of cities (table 1) is sorted according to outdegrees (outgoing links) which are interpreted as a measure of a city’s ‚control capacity’ within worldwide capital flows. With regard to the automotive industry, the urban regions of Tokyo and Detroit take on the first and second rank. On the subsequent ranks we find the urban regions of Go­thenburg, Stuttgart, Nagoya, Milwaukee, Wolfsburg, Frankfurt-Main, Paris, Munich and Hannover. Thus German cities are quite strongly represented amongst the top 60 listing. The overall picture indicates the dominance of global firms from some European countries, the USA and Japan in the automotive industry. However, it is remarkable that several urban regions of the ‘global South’ appear on comparatively high rank positions: Two of these are cities of India, in particular Mumbai (rank 12), and Pune (rank 25). In China, the urban region of Shanghai (rank 23) has achieved a high rank, and in Brazil, the urban region of Sao-Paulo (rank 25) has joined the top 60 listing. This finding clearly indicates the increasing integration of the socalled ‘emerging markets’ in global production networks. Particularly Mumbai shows a comparatively strong ‘surplus’ of control capacities, which reflects the fact that in recent times globally operating firms from India have entered the automotive industry’s global production network.

As regards cities that possess a strong attractive power (and thus are primarily functioning as ‘platform‘- or bridgehead locations within the automotive industry’s global networks), which is measured according to a strong surplus of incoming links, the urban regions of London, Toronto, Los Angeles, Madrid, Birmingham and Sao-Paulo are to be highlighted.

It is also worth noting that in the group of ‘outstanding’ global cities which have been characterized as prime centres of the global financial industry and the advanced producer services - such as London, New York, Tokyo and Los Angeles -, only Tokyo has a top position in terms of an outstanding surplus of control capacity within global networks of manufacturing, whereas the other global cities mentioned before take on comparatively low ranks (yet also the global cities Paris and Milan display a surplus of control capacity in the automotive industry’s global networks). This corresponds to our preliminary thesis that globalizing cities are characterized by differing profiles of glo­bally connected economic functions. This thesis is further supported by the contrasting example of global networks of a different manufacturing subsector (see below).

The tabular representation of city rankings in the ‘technology hardware & equipment’ subsector(see table 2) demonstrates the top 60 of all cities included. In this particular subsector of manufacturing, the urban regions of Tokyo and the prominent ‘Silicon Valley’ (here labelled as ‘San Jose’) take on the first and second rank. Thus Tokyo functions as an outstanding centre in different industrial subsectors. On subsequent ranks we find the urban regions of Osaka, Stockkolm, New York, Helsinki, Taipai, Chicago and Boston. US cities are quite strongly represented amongst the top 60 listing, and the overall picture might indicate a dominance of US and Japanese firms in this subsector. Again, it is remarkable that several urban regions of the ‘global South’ appear amongst the top 60 ranks, such as e.g. Shenzen (rank 24), Bangalore (rank 26), Hong-Kong (rank 26), Beijing (rank 28), Mexico-City (rank 29) and Sao-Paulo (rank 31). The continued integration of the socalled ‘emerging markets’ in global production networks is also apparent in the manufacturing industry’s ‘technology hardware & equipment’ sector. As regards cities that are primarily functioning as platform- or bridgehead locations within this subsector’s global networks (as mea­sured by the surplus of incoming links), the urban regions of Hong-Kong, Toronto, London, Sao-Paulo and Milan stand out. The global service centres Paris and London take on rank positions 9 and 13 in the list (with Paris showing a considerably larger surplus of outgoing links). Besides the examples of Tokyo (rank 1) and New York (rank 5), this fin­ding indicates that global firms of the ‘high technology’ manufacturing industries (such as the ‘technology hardware & equipment’ sector including information techno­logy and semiconductors) more frequently choose prominent ‘global city regions’ as locational anchoring points than global firms of the ‘medium high technology’ sectors (such as the automotive industry). Altogether, the comparative ranking detects considerable differences between the nodal centrality of distinct cities in two particular subsectors of manufacturing.

Tab. 2: Urban Nodes of Global Networks in the ‘Technology Hardware & Equipment’ Sector (including Semiconductors), (based on the Sector’s 40 largest Corporations, 2010)

Comparison of Outstanding Network Nodes of Global Services and Global Manufacturing

In order to prove the thesis that the contemporary world city network encloses ‘multiple glo­balizations’ particularly with regard to the differing sectoral profiles of globalizing cities, we proceed to a comparative ranking of the outstanding network nodes of global services and global manufacturing. The ranking of global service centres offered by the GaWC (Taylor et al. 2010) will be compared with the ranking of globally connected manufacturing centres that has been presented in the foregoing section.

However, due to the differing calculation me­thods we cannot directly compare both analyses’ connectivity measures. In the GaWC approach, the measurement of inter-city connectivity is based on the products of each firm’s ‘service value’ in a distinct city with the same firm’s service value in all other cities included in the respective firm’s organizational network. The ‘service values’ are differentiated according to an investigation of the relative importance of each firm’s establishments (Taylor et al. 2010). The GaWC operation of multiplying the ‘service values’ of firm units assigns a particularly strong weight to global firms’ headquarters and regional headquarters or divisions. In the network analysis approach presented here, the global firm’s establishments are classified according to their position in the corporate hierarchy, and the measurement of inter-city connectivity is based on counts of outgoing and incoming links. While the GaWC method offers a more detailed registration of individual firm units’ properties, the benefit of a network analysis that includes the differing directions of inter-city links is to enable a distinction bet­ween the ‘functional’ roles of cities in the respective sector’s global network. Nonetheless, it is still possible to compare the cities’ rank positions concerning their degree of connectivity in global services and global manufacturing. In the GaWC ranking, the individual cities’ degree of connectivity is expressed as a fraction of the highest scoring city’s connectivity (i.e. London = 1.00). The same expression of relative connectivity has been calculated for the cities’ degree of connectivity in the manufacturing sector. The ranks of the cities’ connectivity in manufacturing were calculated according to the sum of outdegrees and indegrees (without taking diagonal values, i.e. the ‘internal’ connectivity between a global firm’s different establishments in a particular city, into account). Since the GaWC analysis doesn’t cover all the cities included in our network analysis of manufacturing subsectors, the subsequent comparison only refers to cities appearing in both analyses. The tabular representations (see table 3 and 4) display a selection of those cities that might best exemplify the characteristic ‘grouping’ of urban regions according to their positioning in different sectors’ global networks. 

The comparison focuses on the difference between the cities’ rank positions in global services and manufacturing activity in distinct subsectors, starting with the global automotive industry. In the tabular representation of selected cities (see table 3), the left section (cities No. 1-26) displays cities that possess a ‘surplus rank’ in manufacturing in terms of a positive rank difference. These cities are characterized by strong global connectivity in the manufacturing sector (as represented here by the automotive industry), which considerably exceeds their rank position in the sphere of global services. The upper right section of table 3 (cities No. 27-38) presents cities with a comparatively balanced relation of global connectivity in both the service and manufacturing sector. This means that the rank difference is rather small (compared to the indicated maximum of positive or negative differences), which can be interpreted as follows: Globalizing cities such as Tokyo, Toronto, Paris, Milan, Bangkok and Sao-Paulo are functioning to nearly the same extent as globally connected centres of manufacturing firms (in the automotive industry subsector) than as global service centres. The lower right section of table 3 (cities No. 39-56) contains cities with a negative rank difference, which means that these cities possess a ‘surplus rank’ in the finance and service sector. Here, we find well-known examples of leading global cities such as New York, London and Sydney as well as a number of globalizing cities in the world region of Asia, such as e.g. Singapore, Taipai, Kuala-Lumpur, and in particular the Chinese cities of Hong-Kong, Beijing and Shanghai.  Cities that appear in this section of table 3 are characterized by strong global connectivity in service sector activities, which exceeds the degree of global connectivity they have achieved in the sphere of manufacturing (as represented by the automotive industry). However, the point of delimitation between the three city groupings mentioned before is open to adjustment and could be set in slightly different ways (without changing the overall interpretation of results).

Table 3: Comparison of Cities’ Global Connectivity Rankings with regard to the Service Sector and the Automotive Industry Sector 

Table 4: Comparison of Cities’ Global Connectivity Rankings with regard to the Service Sector and the ‘Technology Hardware & Equipment’ Sector (including Semiconductors) 

In table 4, the upper left section (cities No. 1-18) again displays cities that possess a ‘surplus rank’ in manufacturing in terms of a positive rank difference. Among the urban regions whose global connectivity in the ‘technology hardware & equipment’ sector considerably exceeds their rank position in the sphere of global services are e.g. the ‘Silicon Valley’ (labelled as ‘San Jose’), Osaka, Austin, Philadelphia, San Diego, Boston and Helsinki. The lower left section of table 4 (cities No. 19-26) presents cities with a relatively balanced relation of global connectivity in both the service and the high technology manufacturing sector. Besides cities such as Chicago, Munich and Singapore, also the prominent global city regions of New York, Paris, London and Tokyo are functioning to nearly the same extent as globally connected centres of high technology manufacturing firms than as global service centres (in terms of the rank comparison). The right section of table 4 (cities No. 27-46) contains cities with a negative rank difference, indicating a ‘surplus rank’ in the finance and service sector. Cities that appear in this section of table 4 (such as e.g. Mumbai, Moscow, Buenos-Aires, Beijing, Sydney, Zurich and Madrid) are characterized by strong global connectivity in service sector activities, which exceeds the degree of global connectivity they have achieved in the sphere of high technology manufactu­ring industry.

Figure 5 offers a cartographic representation of the cities’ differing sectoral profiles with respect to globally connected service firms and the total of 120 included manufacturing sector firms (based on rank differences). The map demonstrates many different combinations of rank positions: Some cities in the world city network achieve a top rank of global connectivity in both service and manufacturing industries. Examples are Tokyo, Paris and New York. Some cities display a second-level rank in both sectors, such as e.g. Mexico-City, Sao-Paulo, Mumbai and Seoul. Other urban regions get a higher level rank in global connectivity of manufacturing sectors as compared to their role in global networks of the service sector, such as e.g. the ‘Silicon Valley’, Detroit, Stockholm, Berlin, Hannover, Helsinki, Nagoya. The opposite combination, i.e. a top level rank in the global connectivity of service sector firms and lower rank concerning manufacturing industries, can be registered in cities such as Beijing, Shanghai, Hong-Kong, Sydney, Madrid and Moscow.

Figure 5: Differing Sectoral Profiles of Globalizing Cities according to the Cities’ Rank Differences of Global Connectivity in Service and Manufacturing Industries

Altogether, the comparative ranking of urban network nodes of global services and global manufacturing sectors confirms the thesis that different profiles of globally connected economic activities can be detected in globalizing cities all across the world. The world city net­work includes glo­balizing cities focusing on advanced producer services and the financial sector in particular, as well as many cities with differing profiles of their globally connected activities. Thus there are possibly different pathways or sectoral trajectories of ci­ties in globalization, a perspective which might have significant implications for urban economic deve­lopment strategies.

Conclusion and Outlook

We are used to think of the global economy as a mosaic of national state territories which are the ‘containers’ of national economies. This article presented a different view, one which starts from a relational perspective and characterizes the world eco­no­my as a globally extended network of cities and metropolitan regions. According to Taylor (2004), urban deve­lopment might be regarded as a pro­cess of networking that is unfolding on different but intertwined spatial scales. Transnationally operating firms from different sectors are most important actors in the formation of these multiple relational networks.

This article focused on the presentation of new empirical research on the formation of a world city network that functions as a backbone for channeling worldwide capital flows. This research confirms the thesis that glo­bally operating manufacturing firms are connecting cities across the world and thus contribute to the emergence of ‘multiple globalizations’ in the world city network. The analysis detected the cities which are functioning as major hubs of distinct manufacturing industries’ global networks and revealed different profiles of globally connected economic activities in globalizing cities all across the world. The ‘ranking’ of globally connected urban centres of manufacturing activity differs considerably from the well-known rankings of global service centers. In conclusion, we emphasized that there are different pathways or sectoral trajectories of ci­ties in globalization. The vision of a ‘post-industrial’ city which has led many urban researchers to concentrate on service sector activities might become increasingly questionable, as global finance and service centers represent the major geo­graphic hubs of the desastrous development model of finance-dominated capitalism. Cities that are searching for a sustainable development path might be better off by extending and improving ‘real sector’ economic activities. The sectoral trajectory of ci­ties in globalization needs not to be confined to an expansion of global services. Cities might as well improve their economic position by participating in global value chains of manufactu­ring industries.

The inclusion of many cities in global value chains of the manufacturing sector also opens up a perspective on the variable positioning of cities in a globalizing economy. This concerns the question as to how the rather ‚peripheral’ cities – i.e. cities showing a comparatively weak connectivity to the established prime centres of globalized economic activities - could improve their role (and increase the economic returns from participating) in global value chains. In this respect, further research might concentrate on the prospects for ‚upgrading’ the position of ‚peripheral’ cities within global production networks. Altogether, we need more research on the developmental and dynamic aspects of the cities’ inclusion in global networks, e.g. exploring the impact on urban economic growth, labour market divisions, and the formation of transnational urban spaces within globalizing cities.

The unequal extent and different form of the cities’ integration in global economic networks also raises questions concerning the governance of urban development. In the sphere of urban governance, many cities are striving to take on global city functions in order to strengthen their reputation and position in a worldwide inter-urban competition. Very often, urban go­vernance in the globalization arena is geared towards restructuring the city’s spatial fabric and built environment according to the presumed ‘needs’ of global finance and service firms, with the consequence of foste­ring socio-spatial polarization. More effort and initiative is needed for developing strategies that are socially inclusive and might benefit urban residents beyond the upper strata of business elites in the finance and service sector. I have emphasized that there exist different pathways for cities in globalization. A city’s integration in the globa­lization process might as well be taken as a resource for upgrading its industrial structure and employment standards and as a resource of urban policies aimed at spreading the gains of glo­bally networked economic activities in favour of the whole population of an urban area.

APPENDIX: LISTING OF INCLUDED FIRMS

Automotive Industry Firms (40):
Continental, BMW, Hyundai, Fiat, Ford, Honda, Johnson Controls, VW, Denso, Daimler, Renault, Hyundai, Porsche, Nissan, Peugeot Citroen, Suzuki, SAIC (Shanghai Automotive), Michelin, General Motors, Toyota, Volvo, Bridgestone, Dongfeng Motor Corporation, Aisin Seiki, Mitsubishi, Tata Motors, Mazda, Magna International, Genuine Parts, Isuzu, Bosch, MAN, Paccar, Goodyear, Fuji Heavy Industries, Mahindra & Mahindra, Yamaha Motor Corporation, Harley-Davidson, TRW Automotive, Pirelli

Technology Hardware & Equipment Firms (including Semiconductors) (40):
Hewlett-Packard, Apple, Cisco Systems, Nokia, Hon Hai Precision Industry, Dell, Qualcomm, Nintendo, Ericsson, EMC Corporation, Research in Motion, Corning, AU Optronics, Kyocera, Fujitsu, Quanta Computer, ASUSTek Computer, Motorola, Sharp Corporation, NEC Corporation, Alcatel Lucent, Western Digital, Seagate Technology, Acer, Zhongxing (ZTE Corporation), Tyco Electronics, Compal Electronics, Flextronics International, Olympus Corporation, Hitachi, Sony, Panasonic, Nikon, Toshiba, STMicroelectronics, Intel Corporation, Texas Instruments, Samsung, Taiwan Semiconductor, Advanced Micro Devices

Pharmaceutical & Biotechnology Firms:
Bayer, Pfizer, Sanofi-Aventis, Roche, Novartis, AstraZeneca, Eli Lilly, Johnson & Johnson, GlaxoSmithKline, Merck & Co, Abbott Laboratories, Amgen, Bristol-Myers Squibb, Teva Pharmaceutical Industries, Takeda Pharmaceutical Company, McKesson Corporation, Cardinal Health, Novo Nordisk, Astellas Pharma, Merck KgaA, Gilead Sciences, Eisai Corporation, Biogen Idec, UCB S.A., Genzyme Corporation, Allergan, Daiichi Sankyo Co., CSL Limited, Celgene Corporation, Forest Laboratories, Mylan, Shire, Warner Chilcott, Life Technologies Corporation, Cephalon, Actelion, Novozymes, Ono Pharmaceutical Co., Alexion Pharmaceuticals,
Watson Pharmaceuticals

REFERENCES

Alderson, A.S. and Beckfield, J. (2004) ‘Power and Position in the World City System’. American Journal of Sociology 109, 811-851

Alderson, A.S. and Beckfield, J. (2006) ‘Wither the parallel Paths? The Future of Scholarship on the World City System’. American Journal of Sociology 112, 895-904

Beaverstock, J.V., Smith, R.G. and Taylor, P.J. (1999) ‘A Roster of World Cities’. Cities 16 (6), 445 – 458

Derudder, B. and Witlox, F. (eds) (2010) Commodity Chains and World Cities. Oxford: Wiley-Blackwell

Dicken, P. (2007) Global Shift. Mapping the Changing Contours of the World Economy. 5th edition. London: Sage

Friedmann, J. (1986) ‘The World City Hypothesis’. Development and Change 17 (1), 69 – 84

Gereffi, G. (2006) The New Offshoring of Jobs and Global Development. Geneva: ILO

Gereffi, G. and Korzeniewicz, M. (eds) (1994) Commodity Chains and Global Capitalism. Westport: Praeger

Hall, P. (2001) ‘Global City-Regions in the twenty-first Century’. In A.J. Scott (Ed.) Global City-Regions. Trends, Theory, Policy. Oxford: Oxford University Press, 59-77

Henderson, J., Dicken, P., Hess, M., Coe, N. and Yeung, H. (2002) ‘Global Production Net­works and the Analysis of Economic Development’. Review of International Political Economy, 9 (3), 436-464

Jansen, D. (2003) Einführung in die Netzwerkanalyse. Grundlagen, Methoden, Forschungsbeispiele. Opladen: Leske & Budrich

Knoke, D. and Kuklinski, J.H. (1982) Network Analysis. Sage University Papers No. 28. London: Sage

Krätke, S. (2003) ‘Global Media Cities in a Worldwide Urban Network’. European Planning Studies 11 (6), 605-628

Krätke, S. and Taylor, P.J. (2004) ‘A World Geography of Global Media Cities’. European Planning Studies 12 (4), 459-477

Krätke, S. (2011) The Creative Capital of Cities. Interactive Knowledge Creation and the Urbanization Economies of Innovation. Oxford: Wiley-Blackwell

Krätke, S., Wildner, K. and Lanz, S. (eds) (2012) Transnationalism and Urbanism. New York, London: Rout­ledge

Marcuse, P. and van Kempen, R. (eds) (2000) Globalizing Cities. A New Spatial Order? Oxford: Blackwell

Sassen, S. (1991) The Global City: New York, London, Tokyo. Princeton: Princeton Univer­sity Press

Sassen S. (2000) Cities in a World Economy. London: Sage.

Scott, A.J. (ed) (2001) Global City-Regions. Trends, Theory, Policy. Oxford: Oxford Univer­sity Press.

Short, J.R., Kim, Y., Kuus, M. and Wells, H. (1996) ‘The dirty little Secret of World Cities Research: Data Problems in Comparative Analysis’. International Journal of Urban and Regional Research, 20 (4), 697 – 718

Soja, E. (2000) Postmetropolis. Critical Studies of Cities and Regions. Oxford: Blackwell

Sturgeon, T.J. (2000) ‘How do we define Value Chains and Production Networks?’ MIT IPC Globalization Working Paper 00-010, Cambridge, MA: Massachusetts Institute of Tech­nology

Taylor, P.J. (2004) World City Network. A Global Urban Analysis. New York, London: Rout­ledge

Taylor, P.J. (2006) ‘Parallel Paths to understanding Global Intercity Relations’. American Journal of Sociology 112, 881-894

Taylor, P.J., Ni, P., Derudder, B., Hoyler, M., Huang, J. und Witlox, F. (2010) Global Urban Analysis. A Survey of Cities in Globalization. London: Earthscan

Wall, R.S. and v.d. Knaap, G.A. (2011) ‘Sectoral Differentiation and Network Structure within contemporary Worldwide Corporate Networks’. Economic Geography 87 (3), 267-308

NOTE

^* Stefan Krätke, Economic and Social Geography, European University Viadrina Frankfurt (Oder)