GaWC Research Bulletin 223

This Research Bulletin has been published in R. De Groof (ed) (2008) Brussels and Europe Brussels: Academic and Scientific Publishers, pp. 61-72.

Please refer to the published version when quoting the paper.

(Z)

Brussels in World City Networks

P.J. Taylor

Introduction: more than the ‘Capital of Europe’?

There is no doubt that Brussels is one of the major cities in the world and that it owes its position to the international political institutions it houses, notable NATO and, above all, a majority of EU activities. But political processes alone cannot make a city as Canberra, Ottawa, Brasilia, and until recently, Washington DC (Abbott1999) have exemplified. But like the latter city, Brussels appears to be transcending its political initial advantage to become more than simply the ‘capital of Europe’. This is the message of Camilla Elmhorn (2001) who couches her explanation of Brussels contemporary urban success in world city terms – she plots the rise of Brussels as a ‘political world city’. Her emphasis on the world city credentials of Brussels includes identification of an ‘advanced service economy’ and a ‘European agglomeration economy’ within the city. In this brief research report ‘ Brussels as a world city’ is considered from a slightly different perspective: I use results from GaWC (www.lboro.ac.uk/gawc) work on inter-city relations. Thus the emphasis moves from internal relations to external relations to show how the city is located in world city networks.

The case of Brussels as a world city illustrates a basic question in world city studies. The leading cities have an ‘all-round’ functional prowess from finance to culture and following Sassen (1991/2001) are often referred to as ‘global cities’ (Taylor 2005). But the world-economy cannot accommodate many ‘little New Yorks’ or ‘mini-Londons’ as the planning debacle of post-unified Berlin has illustrated so clearly (Krätke 2000). Rather the myriad cities in globalization can prosper by finding their specific niche within world city networks. And herein lies a conundrum. How can we distinguish the process of urban specialization that creates a relatively simple local economy with few external links that is extremely vulnerable from development of an urban niche that leads to a complex local economy strongly positioned in external networks and is therefore highly resilient? Elmhorn’s (2001) categorization of Brussels as a political world city implies that it has converted dangerous specialization into constructive niche. Below I present evidence for this in terms of the city’s positioning in a variety of world city networks

Key positions on inter-city relations

Although this report is primarily empirical, before I present results it is important to understand key elements of the theoretical framework underpinning the research. Following Jacobs (1969) and Castells (1996) I treat cities as process, and the latter is expressed as intra-city networks (or clusters) and inter-city networks. Focus on the latter has generated three closely inter-related dualities:

1. Competition - cooperation
2. Spaces of places - spaces of flows
3. Hierarchies – networks

Let me briefly indicate how I steer the research through these dualities.

Competition – Cooperation

Once you look beyond a single city the question of inter-city relations arises: what form are these relations? There appear to be just two possibilities: Competitive relations that treats other cities as ‘rivals’, and cooperative relations that treat other cities as ‘partners’. But it is not as simple as this contrast suggests: in his long-term study of economic change Arrighi (1994) shows clearly that competition and cooperative between cities varies with economic cycles. This pattern indicates that competition and cooperation are entwined in inter-city relations. The position taken here is that cities are inherently cooperative but with contingent competitive tendencies.

Spaces of Places - Spaces of Flows

This duality comes from Castells’ (1996) theory of space as social construction that distinguishes ‘simultaneity with contiguity’ creating spaces of places from contemporary ‘simultaneity without contiguity’ creating spaces of flows. He treats the latter as superseding former to become the spatial formation in his informational society/network society thesis. In contrast, I treat Castells’ spatial forms as historical, as alternative organizational forms: generically spaces of places are typically political (territories) while spaces of flows are typically economic (networks). As such, the position taken here is a materialist one that prioritises spaces of flows but also recognises that the two spaces occur together in variable contingent relations.

Hierarchies – Networks

Inter-city relations are generally assumed to be hierarchical. This is a product of the ‘national urban systems’ school of urban geography and regional science that flourished from the late 1950s into the 1980s (Bourne and Simmons 1978). This spatial assumption was initially transferred into the world city literature by Friedmann (1986) in his ‘world city hierarchy hypothesis’ (i.e. a simple up-scaling argument: national urban hierarchy becomes world urban hierarchy). In subsequent empirical studies hierarchy (which is a relation) has been conflated with simple ranking (which is a statistical ordering) ( Taylor 1997) so that any variable measuring cities can be used to show a hierarchy (i.e the latter becomes an inevitable outcome). But historically (outside state-centric social science discourses), inter-city relations have been interpreted as networks. With contemporary trans-national inter-city relations, we need to think in network terms in the 21st century. The position taken here is that there are world city networks but with contingent hierarchical tendencies.

Empirics, Theory and Policy

Negotiating a way through these three dualities is not mere semantics. Two features arise. First, there is the identification of one side in each duality as contingent: competition, spaces of places, hierarchy. Thus to understand these features of inter-city relations requires empirical study of specific situations. Second, the other side of the dualities are considered generic to inter-city relations; the latter are inherently cooperative in a space of flows that are networks. This constitutes the theoretical basis of the study. Finally, the dualities have profound policy implications. Hierarchy is a feature of spaces of place and the product of competitive processes: hierarchies are there to be climbed. Network is an unbounded space of flows and a product of cooperative processes: networks are necessarily based upon mutualities. Very different policies emanate from these contrasting positions. The position taken here is that city competition and the need for associated ‘boosterism’ have to be empirically shown; the starting point is always city mutualities.

Connectivity audits

The results on mutualies I report here are measures of the connectivity of Brussels within specific world city networks. These connectivity audits fall into two groups: the first is based upon air passenger flows to provide a general picture of Brussels’ inter-city relations; the second uses the interlocking network model to estimate the position of Brussels in a variety of city networks.

Air Travel Connectivities

Most studies of inter-city relations using airline data have information on flights between airports but not on passenger origin and destinations. Therefore results are distorted by the hub policies of the airlines (e.g. Detroit appears more important than it is because North West Airlines use its airport as its main non-coastal US hub – a large proportion of passengers recorded going to Detroit are merely passing through its airport). In contrast, the results reported here are derived from a unique set of data that identifies the origins and destinations of all trips. By focussing on just the beginning and end of trips the ‘hub problem’ is eliminated. The data are for the first 9 months of 2001 and are described in detail in papers by Frank Witlox and his colleagues at Ghent (Derudder and Witlox 2005).

Two maps are produced that summarise the data on all passengers that started or finished in Brussels. Figure 1 shows totals of Brussels passengers to 47 cities with flows above 35,000 in the study period. This shows two interweaving patterns of flows. First, flows to EU capital cities feature prominently from Stockholm to Rome, from Lisbon to Vienna, and with London recording the largest flow. Note, also in this context the limitations of the data for short flights: the two unnamed cities with relatively small air passenger flows abutting Brussels on the diagram are Paris and Amsterdam. These are not, of course, exceptions to the high capital city flows but those flows are carried more by land. However Berlin most certainly is an exception since it has far fewer flows than either Frankfurt or Munich. Similarly flows to Barcelona and Milan approximate the level of flows to their respective cities. Even in the UK, Manchester shows appreciable flows. These ‘non-capital’ flows suggest economic linkages with Brussels are as important as economic linkages. Beyond Europe this is confirmed by the USA city links especially the prominence of New York.

Figure 2 shows a very different pattern. Whereas Figure 1 illustrates absolute flows, here we have relative flows. It shows where Brussels generates a high percentage of flows. This features 48 cities that I have grouped into 6 regions. Clearly this map illustrates particular flows such as tourism (Southern Europe) and recent migration (North West Africa: Tangier replaces London as the dominant city). This pattern reminds us that the contemporary space of flows, whether by airline or not, is a complex mix of movements based on a vast array of motives. These can only be inferred from this data in the way I have done above. Thus I employ a different method to try and portray more specific connections to Brussels.

But before we leave airline passenger data analysis, we can begin to address the question of whether Brussels’ pattern of flows is similar to those of London and Paris. The latter two cities are chosen for comparison because they are close neighbours of Brussels that are reasonably ‘all-round’ global cities (Taylor 2005). Table 1 shows how Brussels shapes up in terms of passenger flows to New York as archetypal global city and to the Pacific Rim as archetypal globalization concept. Brussels’ absolute flows appear quite large but are put into perspective by comparison with London where Brussels’ flows are generally less than 10% of London’s flows. This contrasts with Paris where flows are generally closer to London’s levels and actually exceeds it in one case. It is noteworthy that both Brussels and Paris do better relative to London with the two Chinese cities. We can conclude from Table 1 that although Brussels is not as important as either London or Paris in generating air passengers, it most certainly has a pattern of connections in these key globalization sites that marks it out as globally positioned within air travel networks.

The Interlocking Network Model

A major positive feature of the interlocking network model is that it can be used to estimate very specific flows between cities. The model is based upon the assumption that material city networks are created by agents carrying out their everyday business involving connections with other cities. These city network makers ‘interlock’ cities through their activities. For instance, the advanced producer service firms that feature prominently in Sassen’s ‘global city’ concept commonly have multiple offices across world cities. Intra-firm flows of information, knowledge and instruction between these offices constitute such network-making. The interlocking network model is a means for estimating such flows and their aggregation defines the world city network (Taylor 2001, 2004). This particular network is deemed to be the most important world city network because, drawing on Sassen’s (1991/2001) arguments, it constitutes the enabling structure for contemporary economic globalization (Taylor 2004). But there are many other users of cities that generate office networks in the course of their work, many of which are not economic (e.g. NGOs) (Taylor 2005). Both economic and non-economic city network makers are considered below.

Connectivities from Advanced Producer Services Networks

According to Sassen (1991/2001) these very high value professional, creative, and financial services are one of the cutting edges of global city formation; I treat them as key agents of world city network formation (Taylor 2004). Traditionally single city-based, they became multi-city based as their clients globalized and today many large service firms have numerous offices across the world. We collected data on 100 such firms across 315 cities in 2000 and 2004. I use two sets of results from an interlocking network analysis of this data featuring Brussels in 2000.

First, I consider aggregate Brussels’ links to the three main globalization arenas again using London and Paris for comparison. Using a simple method for summing links to cities in regions originally employed in Taylor and Lang (2005), the orientations of a city’s hinterworld can be computed. The results in Table 2 are the ratio of positive scores (relative over-linkage) to negative scores (relative under-linkage) derived from the diagrams in the ‘Hinterworld Atlas’ on the GaWC website (www.lboro.ac.uk/gawc). Thus the ratio of 3 for London in respect of Northern America shows that the city’s positive links to Northern American cities are three times higher than the negative links. The key finding in Table 2 is that Brussels has a positively bias in its hinterworld to all three of these leading regions and is particularly strongly oriented to Pacific Asia. Note that European links do not dominate. This is encouraging evidence for Brussels economically transcending its European political importance at a global economic scale. Comparisons with London and Paris orientations confirm the city’s globality and complement similar previous analysis comparing Brussels with Antwerp (Derudder and Taylor 2003).

Second, if links to all other 314 cities in the data are aggregated global network connectivity scores are produced. These are illustrated in Table 3 in which Brussels connectivities for six advanced producer services are recorded as well as the overall connectivity. The connectivities are computed as proportions of the highest scoring city to facilitate comparison across sectors. London and Paris are used for city comparisons again and ranking (globally and for the European Economic Area) are given to simplify interpretation. Looking at the connectivities Brussels has the highest value for law with results in rankings above Paris. This result shows an economic effect of the political activities (‘capital city functions’ attract law firms). Although Brussels falls below London and Paris in all other rankings, the city does show a good all-round pattern of connections, globally in the top 20, and in the European top ten cities, in all sectors except accountancy.

The connectivity audit in this section further supports Brussels supra-European status, on this occasion in the critically important global space of flows that is the world city network.

Connectivities from Selected Non-Economic Networks

As previously noted, there are non-economic agents that use cities in their routine work these have often ‘gone global’ in their activities. Their office networks in aggregate define new world city networks that are non-economic in nature. Table 4 has the same structure as Table 3 and reports on connectivities derived from data on office networks of selected political and cultural network makers (Taylor 2005).

The first point to note is that Brussels is weakest in terms of the cultural network makers (design and construction, architecture, media). In contrast it out ranks both London and Paris in two political network-making categories (international organizations, UN agencies). However, all Brussels rankings are relatively high and comparable to the economic connectivities in Table 3. Thus I conclude that Elmhorn’s (2001) designation of Brussels as a ‘political world city’ is a good description: Brussels has turned its political specialization into a niche within a well-rounded city complex operating through world city networks.

Conclusion: from relations to assets

This has been a simple empirical report showing Brussels’ eminence in world city networks, complementing Elmhorn’s (2001) prior research on the city’s intracity structures. Turning to policy implications of world city researches, the latter, a city’s internal relations, can be deemed its ‘place assets’ whereas the external relations presented above, its connectivities, can be deemed a city’s ‘flow assets’. Both are clearly very important for development of a vibrant city but policy nearly always focuses on the place assets for obvious practical reasons. Policy-makers are constrained by their territorial jurisdictions wherein place assets are contained; they are not politically equipped to deal with flow assets in any comprehensive manner. A consequence of policy being concerned for place assets is that the contingent half of the key dualities recognised above - competition, spaces of place, hierarchy – define the policy framework. In other words the more basic inherent nature of cities – cooperation, space of flows, network - is severely neglected.

Connectivity audits such as those presented above are a very small first step towards pointing city policy makers to think beyond their jurisdictions. Making city mutualities more of a priority does not mean political twinning, rather it is about engaging with existing network makers. Like place policy-making, network policy-making should be about identifying weaknesses and strengths in order to rectify and sustain respectively. For Brussels, the audits suggest a network policy of sustaining political links, developing economic links, and rectifying cultural links.

REFERENCES

Abbott, C (1999) Political Terrain: Washington DC, from Tidewater Town to Global Metropolis. Chapel Hill: University of North Carolina Press

Arrighi, G (1993) The Long Twentieth Century. London: Verso

Bourne, L S and Simmons J W (eds) Systems of Cities. New York: Oxford University Press

Castells, M (1996) The Rise of Network Society. Oxford: Blackwell

Derudder, B and Taylor, P J (2003) The global capacity of Belgium’s major cities: Antwerp and Brussels compared, Belgeo 4, 459-76

Derudder, B and Witlox, F (2005) An appraisal of the use of airline data in assessing the world city network: a research note on data, Urban Studies 42, 2371-88

Elmhorn, C (2001) Brussels: a Reflexive World City. Stockholm: Almqvist & Wiksell International

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

Jacobs, J (1969) The Economy of Cities. New York: Vintage

Krätke, S (2000) Berlin: the metropolis as a production space. European Planning Studies 8, 7-27

Sassen, S (1991/2001) The Global City: New York, London, Tokyo. Princeton, NJ: Princeton University Press

Taylor , P J (1997) Hierarchical tendencies amongst world cities: a global research proposal, Cities 14, 323-32

Taylor , P J (2001) Specification of the world city network, Geographical Analysis 33, 181-94

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

Taylor, P J (2005) Leading world cities: empirical evaluations of urban nodes in multiple networks, Urban Studies 42, 1593-1608

Taylor, P J and Lang R (2005) US Cities in the World City Network. Washington DC: The Brookings Institution (Metropolitan Policy Program, Survey Series)

Table 1. Brussels’ hinterworld in passenger traffic: comparison with London and Paris (2001)

Table 2. Regional orientations of hinterwords: comparison with London and Paris

Table 3. Connectivity audit for Brussels 2000: Economic linkages

Table 4. Connectivity audit for Brussels 2000: Non-economic linkages

Note: This Research Bulletin has been published in R. De Groof (ed) (2008) Brussels and Europe Brussels: Academic and Scientific Publishers, pp. 61-72