GaWC Research Bulletin 298

This Research Bulletin has been published in Global Networks, 10 (1), (2010), 92-113.

doi:10.1111/j.1471-0374.2010.00276.x

Please refer to the published version when quoting the paper.

(Z)

Integrating World Cities into Production Networks: The Case of Port Cities

W. Jacobs^*, C. Ducruet^** and P.W. de Langen^***

Abstract

This paper sets out a conceptual and methodological framework to study the role of advanced producer services in the international commodity chains that move through seaports. Since the early 1990s, scholars from different backgrounds started to conceptualize and empirically analyse the internationalization of the global economy. In economic geography, two successful approaches emerged. The World City Network school focuses on the role and location of advanced producer services to analyse patterns of hierarchy in internationally integrated urban systems. The Global Production Network / Commodity Chain school focuses on how power and value are created, captured and distributed between producers and suppliers that link distant local clusters of manufacturing with each other. Since the port is a logistical node and a site of production in global commodity chains, and the port-city potentially is a center for maritime and port related advanced producer services, the analysis of port-cities provides a good opportunity to integrate both approaches.

Keywords: world city, global production network, port city, methodology

INTRODUCTION

Since the mid 1990s a large body of research emerged that deals with processes of economic-geographical globalization and the formation of global production networks, global value and commodity chains (cf. Gereffi & Korzenwiecz, 1994; Gereffi et al. 2005 Dicken et al 2001; Henderson et al 2002, Ernst & Kim, 2002; Coe et al 2004; Kaplinsky 2004). This body of research aims to conceptualize and understand the processes of industrial restructuring and deregulation of national markets that has taken place since the late 1970s and resulted in a new spatial division of labor. This process is powerfully summarized by Feenstra (1998) as “the integration of world trade and disintegration of production”.

During the same period a different strand of research emerged that focused on the formation of world or global cities (Sassen 1991; Knox & Taylor 1995; Beaverstock et al. 2000; Taylor 2004). It analyses how city hierarchies and urban systems are linked worldwide through corporate networks of advanced producer services (e.g. banks, insurance companies, law firms, consultants) that control flows of capital and information. Its focus is on how the dynamics of urban hierarchies is related to the overall restructuring of the world economy (Friedmann 1986). Both schools (named hereafter as GCC-GVC-GPN and WCN) therefore provide a somewhat different - although potentially complementary - analysis of the globalization of the economy.

Few studies have attempted to conceptually and empirically integrate both bodies of research. Such integration is relevant as it allows us to provide a more complete picture on the structure and evolution of the global economy. In this contribution, we argue that port cities are an excellent analytical focus to empirically investigate the interaction between global commodity chains and advanced producer services. Manufacturers, logistics service providers and retailers control the global flow of goods and commodities that supply the world’s markets. These firms purchase advanced producer services, e.g. to finance their plants, to insure their transport and to be legally represented. Port cities are thus important nodes in commodity chains and may be attractive locations for port & maritime related advanced producer services.

The central research question of this paper is: to what extend are specific maritime and port related advanced producer services concentrated in port cities and to what extend in world cities in general? This question is theoretically and practically relevant. Theoretically, the embeddedness of port cities in global networks needs further understanding, notably in terms of local interactions, not only among port-related activities, but also between the port cluster and other types of activities. Practically, port cities do not compete only for capturing traffic but also for attracting port related firms. The attraction of APS may be a relevant strategy for port cities in upgrading their economic profiles and strengthening the position of ports in a particular commodity chain;

Despite their role in the worldwide distribution of goods (almost 90% of world trade volumes are transported by ship), seaports have not received much attention in both bodies of research. This is surprising for three reasons. First, ports are important transport nodes in vast, global value chains (cf. Wang & Olivier 2006; Jacobs & Hall 2007; Jacobs 2007; Robinson 2002). Many global commodity chains move through seaports: liquid bulk (e.g. oil and gas), dry bulk (e.g. ore and coal), food (grain, corn, soya, fruits) and all kinds of consumer goods shipped in containers¹. Second, many ports have developed into a geographical concentration of highly specialized economic activity, with a demand for specialized labor (cf. De Langen 2003). Many port complexes host manufacturing facilities, such as petro-chemical refineries, because transport costs to/from ports are relatively low. This industrial activity in its turn attracts numerous complementary activities, such as specialized suppliers of e.g. machine tools, spare parts, maintenance etc. Therefore, ports are a good example of what Humphrey & Schmitz (2002: 1018) call “export-oriented clusters that are inserted into global value chains”.

Third, WCN generally neglects the sector-specificity of advanced producer services. By focusing on port cities we can overcome this problem as we are primarily concerned with those advanced producer services that have become specialized in the maritime and transport related industries. Fourth, recent contributions in WCN have looked at the connectedness of world cities by looking at telecommunication, corporate networks (Rimmer 1999), airport traffic statistics (Derudder & Witlox 2005), and Internet connections (Choi et al 2006). These contributions have looked at the importance of crucial physical infrastructure in the global transfer of both people and information as a proxy for world city connectedness. It is indeed argued (Smith & Timberlake 1995; Derudder & Taylor, 2005) that inter-city linkages can be studied along different flows: human, information and commodity. The commodity flows- and their centers of coordination- have been up till date completely neglected within WCN-research, contrary to the works on GCC-GVC-GPN.

Integrating world cities and global production networks

We do not provide a complete overview of the conceptual evolution of both the GCC-GVC-GPN research and the research on World City Networks. Others have provided such contributions (see Coe et al. 2008; Hess & Yeung 2006 on GCC-GVC-GPN and Taylor 2004 on WCN). Instead, we discuss differences and similarities of both approaches and how they can complement each other in the study of the global economy.

Both approaches analyse the dynamics and geography of the global economy by starting with the premise of a new globalized division of labor and industrial organization (cf. Fröbel et al 1980), which resulted from a variety of factors such as the cost reducing effects of new communication and transport technologies and changing economic policies². Both approaches draw for example inspiration from Castells’ (1996) global transformation from a ‘space of places’ to a ‘space of flows’³. Hence, the interest of the WCN-school on the international connectedness of cities through corporate networks and the interest of the GPN-school on the ‘territorial embeddedness’ of (social) networks involved in global commodities.

Both approaches have thus ‘networks’ as the core unit of analysis, but conceptualize them fundamentally different. WCN looks at networks simply as the intra-firm links of advanced producer services between two or more different locations⁴. WCN does not take into account possible relations between these types of firms and other actors at specific points in time and space, nor does it take the historical and institutional contexts into account that underlie interactions among actors and the resulting network formation. The GPN-approach takes a more sophisticated and theoretical stand by emphasizing the relational character of ‘networks’, (Dicken et al 2001). Understanding networks as such requires the identification of the different actors in networks, their social relationships and power configurations, as well as the structural outcomes of these interactions.

As a result the methodological approach of both approaches differs. The WCN-school largely draws on large datasets on the locational patterns of the world’s leading advanced producer service providers. The GPN approach, as mentioned by Hess & Yeung (2006), suffers from a rather underdeveloped methodological foundation and is essentially a heuristic approach (Dicken 2004). As a result GPN research tends to be qualitative, with a strong preference for interviews with key actors. As Hess & Yeung put it (2006: 1201): “it falls short of delivering a rigorous analysis that can give ‘the big picture’ of GPNs on a global scale”. It are exactly these big pictures what the WCN-approach delivers, yet often at the expense of a deeper understanding of the unique historical, institutional and strategic conditions that influence certain outcomes in time and space. Nevertheless, we agree with Hess & Yeung that research on the GPNs should incorporate more explicitly quantitative data and relevant statistical tools into their analytical apparatus.

GCC-GVC-GPN is much more comprehensive in scope than the WCN. Here it is important to make the distinction between GPN on the one hand and GCC/GVC on the other. Coe et al (2008: 2) distinguish two main differences. First, GCCs/GVCs are considered as linear structures, sequential stages within the production chain through which value is added, whereas the GPN “strives to go beyond such linearity to incorporate all kinds of network configuration”. The second distinctive feature of GPN is that it is more comprehensive in the sets of actors and types of relationships it includes in the analyses compared to GCC/GVC. The latter primarily focuses on the governance of inter-firm relationships, whereas GPN encompasses both intra- and inter-firm relationships and actors such as the state, civil service organizations, and international regulatory agencies. Furthermore, GPN pays attention to specific types of relationships (such as public private partnerships) or social network formations (e.g. business associations).

Recently, the WCN school adopted a relational approach. Beaverstock et al (2002) see the formation of world city networks as the outcome of the interaction between two communities, that is territorial (city and state) and functional-economic (firm and sector). This model therefore explicitly recognizes the role and influence of sector-specific institutions as well as complex multi-scalar governance arrangements and state-enforced rules of the game on the formation of inter-urban connections. As such, they have brought world city network analyses more in line with the GPN framework, by including other actors- aside firms- into their conceptual apparatus as well as institutional and sector-specific features. More specifically, this contribution recognizes that cities are both ‘embedded’ within networks (in this case of international operating advanced producer services) and within (multi-scalar) territorial governance arrangements.

Yet few studies have investigated to what degree advanced producer services interact with activities and firms in commodity chains and global production networks. This could provide a more accurate picture of the “nexus of interconnected functions, operations and transactions through which a specific product or service is produced, distributed and consumed” (Coe et al 2008: 2)⁵. In every production network (and at different stages in the production chain) there is a need for advanced producer services, e.g. to finance production facilities or to insure goods when shipped across the ocean. Some APS’s provide services for specific firms/activities in global production networks. Do these APS locate in direct proximity of the nodes in the production chain? Or, do these APS arrange such transactions from a distance and remain concentrated in proximity of other service centers in ‘world cities’?

Such an analysis can provide insights for strategic policy. Within the GPN-approach it is argued that successful regional economic development is, to a large degree, dependent on ‘strategic coupling’ (Coe et al 2004), which refers to the capacity of local actors to match critical regional assets with extra-local actors involved in global flows. Providing certain specialized APS locally can help to align regional economies and local industries with global commodity flows. From agglomeration economics it can be argued that such proximity has indeed its advantages: it is argued that it is the industrial variety or economic diversity of sectors and clusters within a competitive arena and within the spatial concentration of (world) cities or urban regions that will lead to such positive externalities as reduced transaction costs, diffusion of new ideas, niche-market specialization (Jacobs 1969; Glaeser et al 1992; Boschma 2005). In addition, advanced producer services generally provide more value added to urban economies, since they are associated with higher income levels and more in the need for high qualified ‘white collar’ workers.

To summarize, we argue that port cities are an excellent focus to integrate both research approaches. Ports are logistical nodes through which a high variety of commodity chains move and where value is created and captured. Port cities are potential locations of advanced producer services, especially those that service port related customers. It can be expected that within certain port cities, horizontal/ non-linear network relationships will occur between actors involved in the commodity chain (e.g. shipping lines, terminal and warehouse operators) and supportive advanced producer services (e.g. insurance companies or maritime law firms). Table 1 summarizes our approach to analysing the relation between port related APS and the port related activities in commodity chains.

Table 1 :World Port-City Network Analysis

world port city networks: a framework of analysis

Although seaports play a significant role in the worldwide distribution of goods, their position as transport hubs and production centers has been ignored by the research on global commodity chains or production networks. Recent contributions (Jacobs & Hall 2007; Wang & Olivier 2006; Wang et al. 2007), however, have started to conceptually integrate ports with the work on global production networks and supply chains. These studies are still in a premature phase, primarily qualitative and tend to be empirically focused on one particular port’s integration with global supply chains, but often lack a further specification of a particular chain. Exceptions are the works of Carbone & De Martino (2003) and Hall & Robbins (2007) in which they analysed the attempts of the ports of Le Havre ( France) and Durban ( South Africa) to become integrated within the respective international automotive chains of Renault and Toyota.

The main activity of ports is the transshipment of goods. Port-related activities such as wholesale, warehousing, logistics and transport are attracted to the direct proximity of ports due to reduced transaction and transport costs. In addition, some ports have also developed into major sites of production and manufacturing because of industries’ dependency on the import of raw materials. Ports such as Rotterdam, Antwerp, Houston, and Singapore host for example extensive petro-chemical refineries and storage facilities. The concentration of industrial activity attracts utility and energy companies as well as other local suppliers of components and machinery.

The relationships between port and city have changed considerably during the second half of the twentieth century (Levinson 2006)⁶. Ports have become increasingly disconnected from cities in various ways. Spatially, the increased intensity of port-industrial activity, in combination with urban growth, lack of available land for further expansion, and environmental constraints have led to the move of port facilities away from city centers (Bird 1963; Hoyle 1989). The consequent vacant and valuable urban land was used for waterfront redevelopment projects in many port cities across the world (Malone 1996). Institutionally, relationships between ports and cities have eroded due to the devolution of local government control on the port’s management and rise of a more commercial port governance model (Brooks & Cullinane 2007; Jacobs 2007). In economic terms ports have become less and less dependent on the urban labor market due to increased automatization and operational rationalization. Cities have also become less dependent on ‘their’ ports for local economic growth, but often struggle- like many industrial centers in the developed world- to upgrade and diversify their economy. The works of Hoare (1986) and Todd (1993) about British and Taiwanese ports show the weakening locational dependence between ports and adjacent industries. As Boschma & Lambooy (1999: 416) put it “old industrial regions are closely orientated towards established industries, due to strong commitments of capital goods, management, R&D and labor to traditional technologies. This makes them less fit to diversify into new activities”. Recent studies confirm that industrial port regions perform less in terms of traffic growth than port regions (Ducruet 2007) where the service sector is strong. Furthermore, De Langen (2007) shows that the growth of cargo volumes in U.S. ports does not automatically lead to a good regional economic performance (see also Grobar 2008).

Yet, the degree to which port activities and advanced producer services (APS) are interdependent is not well-known. While most evolutionary models depict ineluctable stages of spatial and functional separation between port and urban activities locally (Ducruet & Lee 2006), there is a recognition that port-city relationships still exist through more sophisticated - albeit less visible - forms within the tertiary sector (Amato 1999). This echoes earlier theoretical developments about the importance of gateway functions in the urban economy of coastal settlements (Bird 1977). The location pattern of higher-order activities such as APS tends to follow the urban hierarchy rather than the hierarchy of port throughputs, as seen in the cases of Canada (Slack 1989) and Australia (O’Connor 1989). In particular, O’Connor (1989) distinguishes three main types of locations of which port cities, which provide basic services dealing with daily physical operations, maritime industrial cities, which manage long-term contracts and host more diverse functions (e.g. banking), and international cities that control the long-term organization of global maritime shipping.

The aforementioned outcomes, however, remain limited in assessing how different types of firms actually embed within port cities beyond the sole needs of physical operations. The majority of port economic impact studies stemming from port authorities or scholars rarely indicate in detail which activities among APS are influenced in their location by port functions. Indeed, there is no reason to believe that ports, or firms within the port, do not make use of certain type of services that typically can be found in adjacent cities. Vessels, storage facilities, refineries and terminals: all need to be insured and financed. Given the technical complexity of the business and the numerous transactions and form of cooperation, the potential of conflict is high. Therefore port users need rather specific legal services. It remains the question however whether port users make use of local advanced producer services or whether they purchase the services from firms located elsewhere. New information communication technology might have led that certain port users can derive APS from distant locations. The question is thus whether spatial proximity still matters?

Our proposition is that APS-firms will co-locate nearby firms active in port-industrial complexes as physical proximity will foster the exchange of ideas and the building of trust. For example, an insurance company that specializes in the insurance of marine terminals, vessels or storage facilities will locate nearby ports as it will allow for more easy transactions as well as the opportunity to monitor closely market demands. In such a case, spill-over effects might occur in the form of a combination of different ideas originating from different professions, e.g. firms in specialized insurance products for maritime transport, such as Protection & Indemnity clubs (P&I) and Charterers Liability. However, new communication technology makes physical proximity and co-location less obvious as insurance products can be easily offered to port users from distant locations, with Face-to-Face contacts limited to an occasional meeting in which representatives are flown in. It might also be the case that large insurance companies have their headquarters at a distance, but they have opened local and specialized subsidiaries nearby their customers.

Hypothetically speaking, however, it will be not unimaginable that the relationships between the two clusters in the urban region are rather underdeveloped: there is in other words no related variety and hence no positive externalities. This does not imply that firms within the port cluster do not engage with advanced producer services or vice versa. Transport firms, specialized manufacturers or operators embedded within the port cluster are still in the need for financial services, insurance and legal representation. In this case they acquire these services from somewhere else and outside the urban region. Yet another possibility is when the urban region in question does contain a very strong cluster of advanced producer services that maintains particular links with both the port cluster in the urban region and with port clusters elsewhere. For example, Lloyd’s from London is one the world’s leading marketplaces in maritime shipping insurances and forms the core of a highly spatially concentrated cluster in the City of London that facilitates international business relations in shipping, arbitration and insurance (Bennett 2001). Nevertheless, the centrality of world maritime centers is likely to vary significantly over time (Braudel 1985): the new continental centrality of Hamburg following the collapse of the socialist block largely explains the shifts of some important shipping lines’ headquarters from London in the early 1990s (Ducruet 2006).

One possible answer may be brought by looking at the diversity of situations that all respond to two concomitant mechanisms (Figure 1). The first mechanism is defined by a level of functional specialization and/or diversity of the local economy in relation with the port. Load centers are well positioned in GPN-GCC-GVCs, because the physical infrastructure and transport - or gateway - functions occupy an important share of their economy. In turn, such places have only a low or medium position in WCNs because agglomeration economies have had limited impact, due to geographical remoteness or to the lock-in effect of pre-existing urban centers (Fujita & Mori 1996). On the other hand, service centers reach important ranks in the hierarchy of WCNs, thanks to successful agglomeration economies although the initial advantage of water transport is no longer dominant. For such places, functions related to the physical transfer of goods are often limited due to lack of space, congestion, environmental concerns, and the development of more sophisticated activities - or central place functions.

The second mechanism is defined by the combination and concentration of both GCC-GVC-GPNs and WCNs: urban functions coexist in relative harmony and interdependence. Port cities and world port cities have in common to keep a sustainable balance through efficient planning policies despite external pressures (e.g. port competition) and internal pressures (e.g. urban growth) (Lee et al. 2008). While numerous port cities may attempt evolving towards load centers, service centers, or world port cities, some cases are successful and others are not. Of course, explanations based on spatial cycles (i.e. transport node development, economies of scale, agglomeration, diversification, congestion, relocation) are confronted to specific territorial contexts, policies, and to firm-specific factors and strategies that may modify the general trend.

Figure 1: Analytical framework and port city typology. Source: realized by authors based on Ducret & Lee (2006)

Our empirical approach is as follows. Based upon the dataset of the World Ship Register⁷ we can map the location of the world’s maritime advanced producer services. The following port-and transport related APS are discerned: insurance (P&I) and law, consultancy and maritime organizations (international business associations and regulatory agencies)⁸. The database covers around 7,000 establishments of around 650 firms. Although we recognize that this dataset is not complete, we do think that it is a valid representation of the spatial distribution of specialized APS activity. For these APS firms we have located their headquarters and counted the number of (inter) national links in the form of intra-firm networks for every port city. We can then compare these data with the conventional list of world cities as developed by the Globalization and World Cities Research Network (GaWC) (see for example Beaverstock et al. 1999; Derudder & Taylor 2005). However, we have decided not to make a ranking by assigning rather arbitrary scores (e.g. a port city with a headquarter receives more points than a port city with only a firm establishment) as has been done in a number of WCN contributions. The second step is linking these data with port throughput figures as an indication of the position within commodity chains. This information is relatively easy obtained as most ports present their performance in these ways on an annual basis. We were however unable to obtain information on the added value of the port throughputs. This data is only available in some individual cases while the definition and calculation method of added value greatly differs from one port to another. By combining both data sources we can see which and how port cities are positioned within the global flow of commodities and within the corporate networks of specialized advanced producer services. Based upon these analyses we hope to obtain points of entry for further research and formulate hypotheses.

empirical results

The Global Picture of Maritime and Port-related APS

When we look at the global picture of the location of maritime related APS, one of the clearest findings is the highly dominant position of London. In terms of maritime law and insurance (see fig.2), for example, London has twice as much establishments and headquarters than New York, the second city on the list. In addition, London acts as a class apart in terms of international links, especially with Hong Kong, Singapore, New York and Tokyo. Much of London’s central and dominant position can be explained by historical and institutional factors. Most international contracts between ship owners, insurance companies and third parties are based upon English law ever since ‘Britannia rules the waves’. This history might also explain the current internationally strong position, both in terms of establishments and connectedness, of former British crown colonies of Hong Kong and Singapore. Those are intensively connected with London but do not share significant relations with each other or with the rest of Asia.

Figure 2: Global Network of P&I and maritime law offices. Source: derived from World Ship Register 2008

The position of the Asian world cities within the P&I and maritime law networks matches closely the conventional GaWC list (Taylor 2004). Singapore, Tokyo and Hong Kong are the leading cities in terms of APS in general as well as in this specific niche market. However, some cities that did not rank high in the GaWC research, such as Houston, Rotterdam, Panama City, Piraeus-Athens, Hamburg and Antwerp, clearly emerge as prime locations. This seems to prove the influence of a major seaport on the international connectedness of cities in terms of specialized advanced producer services. In the extreme cases of Panama City and Piraeus-Athens, this can be explained by the presence of crucial infrastructure or a flexible business environment for shipping (e.g. flag of convenience) in Panama and of a high concentration of ship owners (Piraeus-Athens). What our data also shows is the relatively strong position of some cities that do not have a seaport, such as Madrid, Mexico City, Moscow and Paris. This can be attributed to the dominance of the city in the national urban hierarchy ( Mexico City), and to centralist features of maritime related administration ( Madrid, Moscow) or both ( Paris).

Next figure presents the total picture of all maritime APS in our database (i.e. P&I and law, consultants, and maritime organizations). The graph represents the relative degree of connectedness of cities through the intra-firm networks, which is defined by the proportion of exclusive linkages with other cities. For example: London hosts 71 headquarters that control 479 establishments in 350 different cities. Most of these cities are only connected to a London-based headquarter, as seen in the number of smaller dots forming the tributary gravitational area of London in figure 3. With reference to the GaWC results, London can be said to be the world’s maritime cluster nucleus.

Figure 3: Graph from intra-firm links among cities connected through maritime APS (Derived from World Ship Register, based upon TULIP software).

At the same time, other regional cores such as Houston, New York, Oslo, and Rotterdam also possess a dedicated network of subsidiaries while being strongly connected with London. Thus, the hierarchy of cities stemming from the location pattern of APS’ intra-firm networks is closely related with the number of establishments under control. Compared with figure 2 that is based on absolute scores, Singapore and Hong Kong are less important due to the fact that most of their establishments are controlled from external headquarters. Smaller cities such as Haugesund ( Norway), St. Petersburg ( Russia), and Gdansk ( Poland) score better because they act as subcluster nuclei in the Scandinavian and Baltic regions respectively. Conversely, Shanghai may have many connections with other leading APS centers absolutely, in relative terms it does not serve an exclusive hinterland and more acts as a regional location for foreign based companies. The Chinese network gravitates around Beijing instead, probably due to its role as political and administrative center. Another important dimension in the graph is the closeness of some clusters that are in reality geographically distant from each other. In figure 3 for instance, Haugesund locates in the vicinity of the Houston cluster, probably due to their common commodity focus that is based on oil products.

Relation with Commodity Flows

In the next step we look at how these centers of specialized advanced producer services perform in terms of the physical flows of commodities by making use of port throughput data. Total throughput volume is measured in metric tons (including containers)⁹. In order to make it comparable with the APS establishments we standardized the data¹⁰. While figure 4 provides an overview of the different scores of port cities resulting from the crossing of the two main measures, table 2 goes deeper in the analysis of regional and local variations through the European case. Figure 4 can be easily compared with the conceptual typology presented in figure 1.

Figure 4: Port traffic volume and importance of maritime APS establishments (Sources: derived from World Ship Register, 2008; Eurostat, 2008, AAPA, 2008; port authorities)

What becomes clear from figure 4 is that there are only a few true world port cities: locations that are both leading nodes in commodity flows and centers of specialized maritime APS. The world port cities are Singapore, Rotterdam, Hong Kong, Hamburg, Houston, New York, Dubai, Tokyo, Antwerp and Shanghai. To a lesser extend we can identify cities such as New Orleans, Vancouver, Amsterdam, Bremen-Bremerhaven, Los Angeles and Ho Chi Minh City. As mentioned, London is somewhat of a special case since it dominates in APS establishments, but its port does not handle much traffic. The service centers are port cities that have strong positions within APS networks but not within the physical flows of goods. Some of these maritime centers do not even have a port at all, most notably Mexico City DF, Madrid, Kuala Lumpur, Beijing, Taipei, Seoul and Moscow. Leading service centers are then Piraeus-Athens, Panama City, Genoa, Istanbul, Limassol and Oslo. Some other service centers have a more balanced profile due to the larger throughput volumes: Southampton, Barcelona, Manila, Genoa, Gothenburg, and St. Petersburg.

Our data also clearly identifies load centers, i.e. port cities with a strong position in commodity flows but with a weak position in advanced producer services. One first group is dominantly Asian. The ‘Asian profile’ of the world’s load centers is not a surprise given the rise of manufacturing in China and Southeast Asia. Another explanation why these port cities score weak on APS is evolutionary: they have only developed during the 1990s as growth poles with a strong dependency on established urbanized areas for advanced services. Therefore we could argue that traditional port activity in these load centers use advanced producer services from service centers, as seen previously in figure 3. Load centers such as Ningbo ( China), Chiba ( Japan), and Shenzhen ( China) may use APS located in service centers or in the respective world port cities of Shanghai, Tokyo, and Hong Kong. Another typical group is composed of dedicated terminals nearby mining regions such as in Australia (e.g. Dampier, Port Hedland) or Brazil (e.g. Tubarao, Itaqui) for exporting specific commodities. A third group within the load centers are gateways to inland urban concentrations, located at the head of mass freight corridors, such as Port Klang ( Kuala Lumpur), Le Havre ( Paris), and Santos ( Sao Paulo).

The final group simply consists of port cities that do not excel in advanced producer services nor handle enormous traffic volumes. These port cities are often second-order cities in their national urban system (e.g. Valencia, Liverpool, Trieste, and Constanta), and they also handle relatively limited cargo throughputs compared to larger load centers. In evolutionary terms, those port cities may have dropped down the hierarchy of cargo throughput and/or APS due to a variety of factors such as congestion, lack of space, and changing trade patterns, as seen in the case of Liverpool and Baltimore.

A more detailed picture may be obtained from the comparison of the rank in the APS hierarchy and the distribution of commodity shares in total port traffic by port city (table 2). Because bulk traffic is, by nature, heavier than other traffic, it dominates in a large number of ports, but the reasons differ greatly. For instance, the importance of bulk traffic in large agglomerations such as London, Oslo, Brussels¹¹, and Paris covers the need of the local market (e.g. movement of construction materials, scrap metal, etc.), while in Rotterdam and Genoa liquid bulk better relates to the existence of an internationally-embedded petro-chemical complex. Furthermore, the higher share of ro-ro traffic in some national capitals (e.g. Helsinki, Dublin, Piraeus-Athens, and Oslo) reflects the importance of passenger traffic and short-sea shipping rather than automotive chains. The specialization in container traffic shows the importance of the distribution and logistics function of some port cities such as Hamburg and Antwerp in north Europe for hinterland drainage, creating value-added locally through warehousing and other ancillary services. The importance of container traffic in the south (e.g. Piraeus-Athens and Limassol) is better explained by transshipment activities that do not create much added value locally because containers are simply shifted from one vessel to another across the terminal. The higher share of general cargo in Odessa and Istanbul possibly relates to the lesser degree of containerization in those ports. In the end, the direct relation between traffic specialization and rank in APS hierarchy is difficult to establish firmly. Ports exert various functions on different scales, responding to the need of the local, regional, and international economy. Commodity specialization illustrates only partially local industry specialization and insertion in particular value chains. Nevertheless, it is interesting to measure to what extent the activity of an entire port depends on a certain commodity, because traffic itself is a unique measure of the insertion of ports into global production networks.

Table 2: Commodity traffic at main APS concentrations in Europe

Source: realized by authors based on Eurostat (2008) and port authorities’ websites

* Traffic shares for Istanbul are calculated based on quay length due to lack of data

conclusions

In this paper we have argued that integrating the core conceptual ideas of WCN and GCC-GVC-GPN is a fruitful exercise as it can provide a more complete picture of the world as both space of flows and space of places. It helps us to identify the role and location of specialized advanced producer services in the global circuits of production. More specifically, we have demonstrated that port cities are an excellent focus to analyse such interaction. Based upon our analysis we were able to present an alternative topology of ‘world cityness’: those locations that are firmly inserted in physical commodity chains and that are at the same time centers of specialized advanced producer services. What has become clear is the global dominant position of London as center for maritime and transport related APS, despite the rather low throughput figures of its port. Port cities that score good on both variables are Singapore, Hong Kong, Rotterdam, Hamburg, New York, Houston, Antwerp, Tokyo and Dubai. Such locations are the most likely to generate intense interactions between those actors involved in global production and distribution of goods on the one hand and advanced producer services on the other.

However, we were not able to identify the specialized APS that are involved in a particular commodity chain. For example, the port cities of Rotterdam, Singapore and Houston are home to many APS while their ports have a dominant position within the commodity chains of crude oil. Would such a situation result that the APS located within these port cities have more customers within these specific chains? In addition, we were unable to identify (yet) the urban economic impact of being a center in maritime related APS. It remains the question how many jobs are generated by maritime APS within specific port cities or how many value is added by APS along the chain. A third point of attention is the time-factor. We have in this study only presented a static global picture. It would be interesting if we were able to identify how certain patterns evolve over time. For example, the contemporary strong position of Dubai in terms of APS must in some way or another be related to the emirate’s rapid urban economic growth over the last two decades.

ACKNOWLEDGEMENTS

The authors would like to thank Hans Koster for his contribution to data collection and cartography.

REFERENCES

Amato, D. (1999) Port planning and port/city relations, The Dock and Harbor Authority, July-December, 45-48.

Beaverstock, J.V. Taylor, P.J. and R.G. Smith (1999) ‘A roster of world cities’, Cities, 16(6), 445-458.

Beaverstock, J. V, Smith, R.G and Taylor, P.J. (2000) ‘World City Network: a new metageography?’, Annals of the Association of American Geographers, 90(1), 123-134.

Beaverstock, J.V., Doel, M.A., Hubbard, P.J. and P.J. Taylor (2002) ‘Attending to the world: competition, cooperation and connectivity in the World City Network’, Global Networks, 2(2), 111-132.

Bird, J. (1963) The major seaports of the United Kingdom, London: Hutchinson.

Bird, J. (1977) Centrality and cities, London and Boston: Routledge and Kegan Paul.

Bennett, P. (2001) ‘Mutual risk: P&I insurance clubs and maritime safety and environmental performance’, Marine Policy, 25(1), 13-21.

Boschma, R. A. (2005) ‘Proximity and innovation: a critical assessment’, Regional Studies, 39(1), 61-74.

Boschma, R.A.. and Lambooy, J. (1999) ‘Evolutionary economics and economic geography’, Journal of Evolutionary Economics, 9(4), 411-429.

Braudel, F. (1985) Civilization and capitalism, New York: Harper and Row.

Brooks, M. and K. Cullinane (eds.) (2007) Devolution, port governance and performance, Dordrecht: Elsevier.

Carbone, V. and De Martino, M. (2003) ‘The changing role of ports in supply-chain management: an empirical analysis’, Maritime Policy and Management, 30(4), 305-320.

Castells, M. (1996) The rise of the network society. The information age: economy, society and culture, Oxford: Blackwell.

Coe, N.M., Dicken, P. and M. Hess (2008) ‘Global production networks: realizing the potential’, Journal of Economic Geography, 8(3), 271-295.

Choi, J.H., Barnett, G.A., B.S. Chon (2006) ‘Comparing world city networks: a network analysis of Internet backbone and air transport intercity linkages’, Global Networks, 6(1), 81-99.

Derudder, B. and F. Witlox (2005) ‘An appraisal of the use of airline data in assessing the World City Network: A research note on data’, Urban Studies, 42(13), 2371-2388.

Derudder, B. and P.J. Taylor (2005) ‘The cliquishness of world cities’, Global Networks, 5(1) 71-91.

Dicken, P. (2004) ‘Geographers and globalization: (yet) another missed boat?’, Transactions of the Institute of British Geographers, 29(5), 5-26.

Dicken, P., Kelly, P.F., Olds, K. and H. W-C. Yeung (2001) ‘Chains and networks, territories and scales: towards a relational framework for analysing the global economy’, Global Networks, 1(2), 89-112.

Ducruet, C. (2006) ‘Port-city relationships in Europe and Asia’, Journal of International Logistics and Trade, 4(1), 13-35.

Ducruet, C. (2007) ‘Port regions: redefinition and comparability’, Paper presented at the International Congress on Ports in Proximity, Antwerp and Rotterdam, December 5-7.

Ducruet, C. and S.W. Lee (2006) ‘Frontline soldiers of globalization: port-city evolution and regional competition’, GeoJournal, 67(2), 107-122.

Ernst, D. and L. Kim (2002) ‘Global production networks, knowledge diffusion and local capability formation’, Research Policy, 31(8/9), 1417-1429.

Feenstra, R.C. (1998) ‘Integration of trade and disintegration of production in the global economy’, Journal of Economic Perspectives, 12(4), 31-50.

Friedmann, J. (1986) ‘The world city hypothesis’, Development and Change, 17(1), 69-83.

Fröbel, F., Heinrichs, J. and O. Kreye (1980) The new international division of labour, Cambridge: Cambridge University Press.

Fujita, M. and Mori, T. (1996) ‘The role of ports in the making of major cities: self-agglomeration and hub-effect’, Journal of Development Economics, 49(1), 93-120.

Geddes, P. (1924) ‘A world league of cities’, Sociological Review, 26, 166-167.

Gereffi, G., Humphrey, J. and T. Sturgeon (2005) ‘The governance of global value chains’, Review of International Political Economy, 12(1), 78-104.

Gereffi, G. and M. Korzeniewicz (eds.) (1994) Commodity chains and global capitalism, Westport: Praeger.

Glaeser, E.L., Kallal, H.D., Scheinkman, J.A. and A. Shleifer (1992) ‘Growth in cities’, The Journal of Political Economy, 100(6), 1126-1152.

Grobar, L.M. (2008) ‘The economic status of areas surrounding major U.S. container ports: evidence and policy issues’, Growth and Change, 39(3), 497-516.

Granovetter, M (1985) ‘ Economic action and social structure: the problem of embeddedness’ American Journal of Sociology, 91(3), 481-510

Hall, P. (1966) The world cities, London: Heinemann.

Hall, P.V. and G. Robbins (2007) ‘Which link in which chain? Inserting Durban into global automotive chains’, in J.J. Wang, D. Olivier, T.E. Notteboom and B. Slack (eds) Inserting port cities in global supply chains, Aldershot: Ashgate, 221-231.

Henderson, J., Dicken, P., Coe, N. and H. Yeung (2002) ‘Global production networks and the analysis of economic development’, Review of International Political Economy, 9(3), 436-464

Hess, M. and H.W.C Yeung (2006) ‘Whither global production networks in economic geography? Past, present and future’, Guest editorial, Theme Issue on Global Production Networks, Environment and Planning A, 38(7), 1193-1204.

Hoyle, B.S. (1989) ‘The port-city interface: Trends, problems and examples’, Geoforum, 20(4), 429-435.

Hoare, A.G. (1986) ‘British ports and their export hinterlands: a rapidly changing geography’, Geografiska Annaler, 68B(1), 29-59.

Humphrey, J. and H. Schmitz (2002) ‘How does insertion in global value chains affect upgrading in industrial clusters?’, Regional Studies, 36(9), 1017-1027.

Jacobs, J. (1969) The economy of cities, New York: Vintage.

Jacobs, W. (2007) Political economy of port competition. Institutional analyses of Rotterdam, Southern California and Dubai, Nijmegen: Academic Press Europe.

Jacobs, W. and P.V. Hall (2007) ‘What conditions supply chain strategies of ports? The case of Dubai’, GeoJournal, 68(4), 327-342.

Kaplinsky, R. (2004) ‘Spreading the gains from globalization. What can be learned from value-chain analysis?’, Problems of Economic Transition, 47(2), 74-115.

Knox, P. and P.J. Taylor (eds.) (1995) World cities in a world system, Cambridge: Cambridge University Press.

Langen, P.W. de (2003) The performance of seaport clusters, Rotterdam: ERIM.

Langen, P.W. de (2007) ‘The economic performance of seaport regions’, in J.J. Wang, D. Olivier, T.E. Notteboom and B. Slack (eds) Inserting port cities in global supply chains, Aldershot: Ashgate, 187-202.

Lee, S.W., Song, D.W. and Ducruet, C. (2008) ‘A tale of Asia’s world ports: the spatial evolution in global hub port cities’, Geoforum, 39(1), 372-385.

Levinson, M. (2006) The box. How the shipping container made the world smaller and the world economy bigger, Princeton: Princeton University Press.

Malone, P. (1996) City, capital and water, London: Routledge.

O’Connor, K. (1989) ‘Australian ports, metropolitan areas and trade-related services’, The Australian Geographer, 20(2), 167-172.

Robinson, R. (2002) ‘Ports as elements in value-driven chain systems: the new paradigm’, Maritime Policy and Management, 29(3), 241-255.

Rimmer, P.J. (1999) ‘The Asia-Pacific Rim's transport and telecommunications systems: spatial structure and corporate control since the mid-1980s’, Geojournal, 48(1), 43-65.

Rozenblat, C. (ed.) (2004) Port cities in Europe: A comparative analysis, Le Havre: IRSIT (in French).

Smith, D.A. and M. Timberlake (1995) ‘Cities in global matrices: toward mapping the world-systems’ city system’, in P.L. Knox and P.J. Taylor (eds) World cities in a world-system, Cambridge: Cambridge University Press, 79-97.

Slack, B. (1989) ‘Port services, ports and the urban hierarchy’, Tijdschrift voor Economische en Sociale Geografie, 80(4), 236-243.

Taylor, P.J (1997) ‘Hierarchical tendencies amongst world cities: a global research proposal’, Cities, 14(6), 323-332.

Taylor, P.J. (2004) World city networks: A global urban analysis, London: Routledge.

Todd, D. (1993) ‘The interplay of trade, regional and technical factors in the evolution of a port system: the case of Taiwan’, Geografiska Annaler, 75B(1), 3-18.

Wang, J.J. and D. Olivier (2006) ‘Port-FEZ bundles as spaces of global articulation: the case of Tianjin, China’, Environment and Planning A, 38(8), 1487-1503.

NOTES

^* W. Jacobs (contact author), Department of Applied Economics, Erasmus University Rotterdam, The Netherlands, Email: wjacobs@few.eur.nl

^** C. Ducruet, Centre National de la Recherche Scientifique (CNRS), UMR 8504 Géographie-cités, Equipe P.A.R.I.S., Paris, France, Email: ducruet@parisgeo.cnrs.fr

^*** P.W. de Langen, Department of Applied Economics, Erasmus University Rotterdam, The Netherlands, and Port of Rotterdam Authority, Email: delangen@few.eur.nl

1. As mentioned by Coe et al (2008: 6): “In fact, with the vastly increased complexity and geographical extensiveness of production networks (…) the logistics problem is absolutely central. We need to understand it. And, yet it is virtually ignored outside the specialist technical world of supply chain management”.

2. We recognize that both schools have drawn inspiration from works that date back much further. In the case of WCN, reference is often made to Geddes (1924) and Hall (1966) that have used the term ‘world city’, albeit in a different historical context and functional meaning. In the case of GPN, much inspiration is drawn from Granovetter (1985) on the ‘social embeddedness’ of rational economic behavior. Here we primarily refer to Castells (1996) as a shared source to summarize the importance of ‘networks’, both in metaphorical and material sense, that underpin the economic geography of the global economy and its conceptualizations (see Dicken et al 2001).

3. Both approaches explicitly recognize that the world has become both (see Henderson et al 2002; Derudder & Taylor 2005).

4 As mentioned, other approaches have looked at airport and internet traffic statistics as well.

5. An important step to integrate WCN with GPN is by accepting the horizontal non-linear relations that exist at certain place and at a certain stage within ‘the production chain’ (Coe et al 2008). GPN can incorporate linkages between producers and industrial suppliers on the one hand and advanced service providers on the other. Such an approach demands that we include both material flows as well as adopt a sector-specific focus to the analyses of advanced producer services.

6. Port cities are historically commercial centers with considerable geo-economic and political power. During the 1600s, a banking system emerged in cities such as Venice, Amsterdam and London around the commodity trade through these seaports (O’Connor 1989). Over the last two centuries the location of these financial and trade related economic activities has become more dispersed.

7. Available at http://e-ships.net/. Accessed between May 2008 and September 2008.

8. WCN usually also discern other types of APS such as ICT, banking, marketing and accountancy. We have not included these types of APS in our study for two reasons. Firstly, we believe that there is limited sector-specific specialization among these types of APS. A second reason is more pragmatic in the sense that these types are not included in the World Ship Register database.

9. Data was extracted from Eurostat, the Association of American Port Authorities, and from various port authorities’ websites.

10. Standardization of the data is as follows: value x - (average) / standard deviation

11. The case of Brussels stands out due to the fact that it acts as the seat of the European Union which attracts all kinds of European-wide associations in terms of shipping and transport.