GaWC Research Bulletin 307

GaWC logo
  Gateways into GaWC


PAR-LON - Doing Business in Knowledge-Based Services in Paris and London: A Tale of One City?

L. Halbert* and K. Pain**


This paper engages with the growing literature on global city-regions. It considers the recently proposed "Mega-City Region" concept of the North West Europe ‘Polynet’ research by providing empirical evidence from its Paris and London case studies. The locations and practices of APS firms and professionals are examined to analyse the development of functionally interlinked urban systems spreading around major global cities. Results show the important role of ‘first’ cities in anchoring regional systems to Castells "spaces of flows" but it also reveals how major global cities like Paris and London contribute to enhanced global reach of secondary cities. Finally, two models of integration are discussed to account for the differences observed between a relatively well integrated London/South-East England Mega-City-Region and a more hierarchically organised Paris/Bassin Parisien system.

Keywords: Mega-City Region, Advanced Producer Services, Networks, Complementarities


After twenty years of urban literature, interest in the city-region has not abated, in fact, following a long period of attention on ‘world’ and ‘global’ cities, there has been a recent resurgence of focus on the city-region level. City-regions are recognised as key spatial organisations for human development in the context of the C21st information-/innovation-driven economy and intensified globalisation. The emergence of world and global cities (Friedmann, 1986, 1995; Hall, 1966; Sassen, 1991; Taylor, 2003) in the contemporary urban era has produced a new phenomenon at the city-region scale. Scott’s ‘global city-region’ (2001) constitutes a critical type of city-region since it represents a concentration of advanced producer services (APS) that are a key feature of global cities. From Reich's (1991) ‘symbols manipulators’ to Sassen's business and financial service providers, advanced knowledge-based services, APS, have come to be considered major players in the constitution and functioning of a global economy. World city literature includes numerous theoretical and empirical studies on APS geographies that can be divided into two parallel and complementary strands: First, APS as central elements in the formation of a world-city network at an inter-city-region level; second, APS and their role in the reorganisation of economic activities within city-regions, in other words at an intra-city-region level, in terms of sprawl, deconcentration and/or polycentricity for example.

This paper sets out to examine how APS location strategies and the day-to-day practices of APS firms and professionals contribute to: i) the formation of inter city-region interactions; and ii) internal geographic reorganisations within a given city-region. These theoretical considerations address two hypotheses that will be discussed on the basis of empirical evidence gathered in the London and Paris city-regions. The first hypothesis tries to disentangle the relation between the well-acknowledged intensified competition between firms in the context of globalisation, including in the APS sector, and the would-be increased competition between city-regions themselves. Even if at the same time inter-city-region competition exists, APS firms’ practices and locational strategies stimulate complementarities and interdependences that can bring positive economic impacts to city-regions due to their urban network constitution. The second hypothesis does recognise that to some extent the development of APS sectors is important for all world city-regions, this does not imply that their location and day-to-day practices automatically induce a similar spatial organisation. Several patterns can co-exist, as in the case of London and Paris where a ‘mega-city region’ development process has been identified for South East England, versus a multipolar-monocentric organising structure for the Bassin Parisien.

The discussion is based on empirical evidence gathered during the 2003-2006 Polynet research programme, which aimed to test the hypothesis of the formation of polycentric mega-city regions in North-West Europe (Hall and Pain, 2006). Bearing in mind similar and dissimilar results found in smaller city-regions - the Randstad, RhineRuhr, Rhine-Main, Brussels, Zurich and Dublin - in the present paper we focus on the specific findings for the Paris and London cases drawing on quantitative and qualitative data. APS location strategies and day-to-day practices are used to cast light on the still poorly understood flows and interdependencies that exist between and within the world city-regions.

The paper starts by reviewing the literature that demonstrates growing interest in APS activities as key elements of world city network formation. Particular attention is paid to the blurring of limits between the inter- and intra-city-region levels, showing how processes operating between city-regions are increasingly affecting their internal organisation. A key focus is the role of APS professionals as day-to-day actors in the life of contemporary city-region formation, highlighting the way in which the Polynet research program and the following study of ‘PAR-LON’ contributes to this type of analysis. The second part reviews some key characteristics of the Paris and London city-regions and their relevance in testing the afore-mentioned hypotheses. The results from the quantitative analysis based on world city network study using the ‘GAWC’ methodology are also developed here. In part three, additional evidence supported by qualitative analyses of in-depth semi-structured interviews with APS professionals is considered. In part four we draw together and reflect on the evidence on interdependencies between world city-regions, and on the development of differentiated patterns of organisation at a regional level in Paris and London. Finally we turn to consider the potentials to test the city-region concept as discussed in the paper with reference to other sectors and actors in order to deepen understanding of the complex mix of flows and places contemporary city-regions have become.

World City Network Theory and New Intra-Metropolitan Geographies

The world city-region literature has developed two series of parallel analyses based on the complementarities between two geographical scales of interaction.

World City Network Theory

Beaverstock et al’s brief review of world city network literature (2000) stresses two shifts. The first shift begins with the broad encompassing definition of world cities of Peter Hall which drew on the earlier work of Geddes (1915), to the economic-based approach of John Friedmann (1982). Hall depicted world cities as a concentration of international level functions such as corporate headquarters but also international political organisations, cutting-edge scientific communities, high-level cultural activities and so on. Following the ‘economic turn’ of the 1970s urban literature, Friedmann narrowed his approach to the power and control a few world cities were gaining over the global corporate economy. To measure this, Friedmann developed an attribute-based methodology that selected the number of corporate headquarters or air passengers for cities, for example to describe a world city hierarchy.

A second shift in the world city network literature was the combined result of two influential works, namely Sassen's global city hypothesis and Castells' network society. According to Sassen, some cities, at first London, Tokyo and New York, later enlarged to a set of 20 cities (Sassen, 2002) play a critical role in the development of a global economy. It is within Sassen’s global cities, as opposed to their city-regions, that the global corporate economy is purported to find the key resources needed to control and organise its structure of world-wide decentralised organisations. These resources are not only infrastructures – information and communication technologies (ICTs) and international airports - but also, and most importantly, skills and human capital, especially in the form of APS professionals. The communities and activities of these high-skilled, high-paid workers cluster within global cities which constitute the neo-marshallian nodes from which they manage the global economy (Amin and Thrift, 1992). In Scott’s framework of analysis, they are the key agents responsible for the day-to-day functioning of the world city-region networks.

With similar consequences for an understanding of contemporary economic and social changes, Castells (1989) conceptualises cities as nodes processing the flows of a global information- and innovation-driven economy. This second shift in the literature introduces an important theoretical advance which has significance for city-regions, moving attention from the ‘space of places’ to the ‘space of flows’. As a consequence of Castells’ insights, the academic community was able to engage with the city-region not so much in terms of its attributes as in previous work, but in terms of the interactions which characterise a networked urban system. Indeed, city-regions were no longer seen as static places defined by a set of resources. They were described by Veltz (1996) amongst others, as dynamic arrangements facilitating innovation, in other words as incubators, and processing flows of ideas, people, capital, and goods, in other words as commutators. The two-fold dimensionality of world city-regions gives them a substantial advantage in the global flow-based and knowledge-driven economy.

Following Castells’ analysis, the literature has progressed to describe the new global spatial order as an archipelago economy of interlinked city-regions. According to this view, city-regions have more interaction with their ‘hinterworld’ than with their ‘hinterland’ (Taylor, 2001). In the last ten years, there have been numerous attempts to describe the interlinked world city network empirically. A critical question has been which markers of interactions offer the most appropriate insights into the reality of the spaces of flows the scientific community is seeking to describe (see Derruder, 2006) for further discussion). However the dearth of reliable data, referred to as the "dirty little secret of world city" analysis by Short et al. (1996), continues to constrain understanding of the world city network and thus the interactions between and within world city-regions.

New Intrametropolitan Geographies

Increasing concerns about the challenges for policy presented by geographical changes occurring within city-regions have emerged alongside the world city network literature. Although interest has not focused entirely on world city-regions, they have played a critical part as a laboratory to test a number of hypotheses in scientific research. London, New York, Los Angeles, Montreal, Sydney or Paris, to name but a few examples have been thoroughly scrutinised as examples of intrametropolitan reorganisations of economic activities. As early as the 1980s, one important area of investigation has focused on the 'flight' or the 'exodus' of corporate complexes (headquarters and APS) from traditional central business districts (CBDs) to the peripheries of city-regions. This process, which was once seen as adding to urban sprawl, was depicted as a third wave of deconcentration, following the residential and industrial deconcentration of the fordist city-region (for a detailed review on this literature, see Halbert, 2004:chapter 1).

Several debates have arisen around at least three sets of questions: Firstly, to what extent is the corporate complex relocating its activities, or some of them, in peripheral parts of the city-region? Secondly, is deconcentration leading to multipolar city-regions (Garreau, 1991) or on the contrary to ‘edgeless cities’ (Lang, 2003) ? Is deconcentration signalling the death of cities and the birth of new urban spaces? Or is there evidence of the ongoing strength of the central area? Twenty years of empirical studies and academic controversies would tend to conclude that there is not a single model of deconcentration but that national policies and local/regional factors are important in shaping differentiated city-regional situations, hence, although deconcentration exists, it cannot be understood simply as the death of central cities as such but must be seen as part of their adaptation to a changing economy (see for example Cook et al., 2007; Taylor, 2003). Resurgence of cities theory therefore coexists with a view of the city-region as a product of both multipolarisation and diffusion trends in which secondary poles of economic activity frequently emerge in the peripheries of city-regions, including in sectors traditionally located in central business districts (CBDs). In this broad scientific framework, APS have played a critical role as indicators of the deconcentration of the corporate complex from the centre to the peripheries of city-regions as numerous articles illustrate (Bourdeau-Lepage and Huriot, 2005; Coffey et al., 1996; Halbert, 2008). Among other works, the Polynet1 research programme distinctively attempted to integrate Castells' network society theory into the scientific analysis of the city-region. Its objective was to expand understanding of the new spatial organisation of economic activities within city-regions by specifically examining the spaces of flows that interlink city-regions in the world city network. This has led to a methodological change in the research stream that needs to be clarified here.

The Daily Workings of City-Regions

In debates both at an inter city-region level and within city-regions, a shift reflecting that of world city network literature can be identified in recent research programmes. Early studies of the location of knowledge-intensive business services (KIBs), and more specifically APS, relied on employment data using national censuses. Such attribute-based approaches postulate that the location of a firm is an acceptable proxy to distinguish the contribution of a particular place to the urban system it belongs to, hence the number of jobs in knowledge economy industries are supposedly efficient markers of intra-metropolitan spatial reorganisation. The continuation of this type of limited study most probably reflects the lack of alternative, superior data. At the world city network level, the number of Fortune 500 HQs has been and is still widely used as an indirect indicator of the ranking of city-regions in a world hierarchy, replicating Friedmann’s hierarchical approach to world cities. But a second set of studies has taken on board Castells' theory of cities and city-regions as nodes in the circulation of heterogeneous flows under globalisation and it is this latter approach that was the basis of the Polynet research programme.

We have already mentioned empirical works focusing on the world city and world city network levels drawing on the work of Sassen and Castells. At the level of the city-region also, several attempts have been proposed to follow suit by considering city-regions as spaces of interactions and flows rather than places with static attributes. Commuting data have provided a classic indicator of flows but these are very limited since they depict the interdependencies between the places of work and residence rather than the effective daily working of an urban-regional productive system (see Pain, 2007b; 2008 for further discussion). Other works on the use of ICT by firms (see (Halbert, 2007b) for telephone calls exchanged between firms within and outside the Paris/Ile-de-France region) attempt to develop this line of understanding. Polynet has been conceived within this general research framework as an attempt to find empirical evidence of the contribution of APS professionals to the daily functioning of several city-regions. Moving beyond approaches based on attribute data, the research has investigated how, in their day-to-day practices, APS interlink several parts of the city-region into a network structure operating at different geographical levels, regional, national, European and global. In particular, the distinction between city-region linkages that are associated with global as opposed to regional level networks has been seen as an important marker of world city-regions which are connected with distant but similar nodes for APS through processes of globalisation.

It is with these questions in mind that the Polynet research programme set out to study the emergence of the ‘mega-city region’ (MCR), defined as a networked region developing around world cities in which a set of cities that are physically separated but functionally interconnected show a complex division of labour (Hall and Pain, 2006). Importantly, in examining city-region functional connections, the MCR concept focuses specifically on the APS knowledge-related activities which interlink the world city network, thus it was hypothesised that where these are located in several cities, and not only in the central world city, their interactions potentially bypass the world city creating a MCR complex.

Several types of indicators can be used to provide evidence of MCR formation:

  • Morphological integration: A grouping of physically separate but proximate cities that are located around a world city.
  • Functional integration: Daily interactions between the constituent cities of a city-region (core-periphery and/or multipolar) as revealed by commuter data for example.
  • Economic integration: A division of labour between proximate cities with potential differentiation between a city-region ‘core’ and ‘periphery’, in particular APS in the world city and other types of activities in secondary cities.
  • Organisational (world city network) integration: City-region complementarities where APS networks are not simply represented in a world city, but interlink a number of proximate cities (the Polynet MCR hypothesis).
  • Relational integration: Intense flows of information, ideas, people, capital between different cities within an MCR (including tangible as well as intangible flows resulting from APS daily activities).
  • Political integration: The existence of strategies, policies and even institutions at the MCR level. This last dimension is not considered a constitutive element in the definition of the MCR according to the Polynet hypothesis, however its presence, or absence, is a subject of study as the need for political integration may flow from other MCR formation characteristics.

Although the principal distinguishing feature of the study was its focus on APS network analysis (quantitative measurement and interviews with APS agents), traditional attribute (population, employment, etc.) and flow data (commuting, air travel and ICT) are introduced in order to contextualize relational network data with reference to traditional space of places analysis.

PAR-LON: Networked firms, networked city-regions (Quantitative analyses)

PAR-LON: Models, Mirrors and Competitors

Since Hall's world cities analysis, Paris and London have been considered the two most dominant world cities in Europe. Parallels include:

  • the same demographic size,
  • the same historical concentration of powers (industrial, political, economic, cultural, scientific),
  • the same economic profile: overall a diversified economy and some specialisations in knowledge economy-related activities with some slight differences, London emphasis on international financial services versus Paris on sciences and innovation and high-tech industries,
  • the two largest goods consumer markets in Europe,
  • the same extraordinary physical and virtual accessibility (see Rutherford, 2004 for ICT),
  • the depth and width of labour markets.

Their proximity and comparable challenges have led to reciprocal sources of inspiration, as exemplified by the British new towns which inspired the Paris/Ile-de-France villes nouvelles, and, reciprocally, La Défense inspiring the Canary Wharf development project (Fainstein, 2001). Recent investments have also increased their functional and infrastructure complementarities. The EuroStar shuttle is an infrastructure that physically links France and UK but in practice it is a transportation service that brings together Paris and London.

But the cities “need to position themselves as key nodes in an increasingly global nexus” (Budd, 1998:663), thus they are also in fierce political competition, for example for foreign direct investments (FDIs) such as exemplified by the ‘excellence pole’ in financial services recently created in Ile-de-France to explicitly challenge New York and London's Financial Services clusters. In the fight to play host to the 2012 Olympic Games, the cities were part of the global “bandwagon of urban entrepreneurialism” (Peck and Tickell, 2002:292), but were also in direct and ardent competition with each other.

From City-Regions to an Enlarged Urban Region

Moving on to consider the quantitative results from the Polynet study, we consider the existence of an enlarged world city-region with new relational geographies associated with APS network agents in order to test different elements of the MCR definition, namely morphological, functional, economic, organisational and relational integrations.

An Enlarged Urban Region

Sprawl and urbanisation: mapping population densities (morphology) and travel to work data (functional areas) in PAR-LON shows a similar pattern of joining up of contiguous urban areas to form enlarged urban regions comprising 20 to 50 cities around the world cities Paris and London. Whereas the extent of medieval and modern cities was in the order of 10 kms, and for industrial cities 30 to 40 kms, the scale of the PAR-LON world city-region is now over 100 to 150 kms. In both cases the regions are dominated in a morphological sense by the London and Paris agglomerations. In London's case, strict post World War II planning policies, introduced in the Greater London Plan 1944 have prevented the progressive outward development to form continuous sprawl. In the Paris region, the effect of the 1965 Schéma Directeur d'Aménagement et d'Urbanisme de la Région de Paris and the later Master Plans had more mitigated effects. New towns did contend urban sprawl but at a shorter geographical radius than in London. However, in both cases, development (sprawl) has taken place beyond the zone of planning restriction leading to the formation of world city-regions of multiple proximate urban centres.

Figure 1: The Paris / Bassin parisien "Mega City-Region"


Figure 2: The London / South East "Mega City-Region"

Analysis of population, employment and commuting data also reveals similar contiguity of functional urban areas (FUAs) (in Polynet, functional urban regions (FURs) comprising a core, based on employment size/density, and a ring showing daily commuting to the core) and cross-commuting between FUAs in the PAR-LON enlarged urban regions. The data shows the city-regions as extensive spaces of places that are internally functionally connected by spaces of flows (daily travel to employment).

An economic division of labour is also well established in both city-regions. Until the 1960s, the division of tasks was of the fordist type with command, control and innovation functions in Paris and London versus manufacturing in the rest of the city-region. But since then employment structures in the enlarged Paris/Bassin Parisien and London/South East England city-regions have moved progressively away from manufacturing activity towards services, especially advanced financial and professional services (Thiard, 1998). T his is now shifting to a complex structure with not only the regionalisation of retail services to households but also of APS (services to businesses) in secondary cities.

Figure 3: Commuting patterns in the Paris / Bassin parisien Mega-City-Region


Figure 4: Commuting patterns in the London / South East England Mega-City-Region

Polycentricity and commuting: Maps of commuter data show the importance of centre-periphery movement but also the development of cross-cutting flows in both city-regions. Radial flows into and out of Paris and London dominate but other centres, some at a much greater distance, have complex interlinkages with other centres. Using daily commuter journeys as a measure of city-region polycentricity, Paris/Bassin Parisien and London/South East England are shown to be much more polycentric in terms of their functional structures than is apparent from their morphologies (see Gilli, 2005 for Paris/Bassin Parisien).

→ On the basis of morphological, functional and economic integration markers, there is thus evidence of the emergence of broadly similar patterns of PAR-LON city-region enlargement.

APS Geography

But, what about the geography of APS organisations? Where are they located? How and where do they interconnect cities within these two enlarged urban regions through their network relations?

To examine the MCR phenomenon as informed by Castells thinking, we must m ap APS network flows. In Polynet this is done through the application of the GAWC methodology. Here d ata on the links between cities associated with APS organisational networks are used in preference to data on nodal size (employment and firms) in order to show to what extent cities are interconnected across the space of places by the presence of service network activity. Because data is collected in Paris/Bassin Parisien and London/South East England for many APS networks operating at different geographical scales, it is possible to assess world city-region organisational complementarities and integration that specifically relate to Castells’ space of flows (the MCR hypothesis).

Eight service categories are included in the Polynet study: banking and financial services, insurance, law, management consulting/information technology, accountancy, advertising, advanced logistics and design consulting. These categories are defined broadly to allow for increasingly dynamic APS service market adjustment (specialisation, differentiation, inter- and multidisciplinarity), using an adapted version of the NACE international ‘Hoppenstedt’ business classification system.

The degree of polycentricity of the city-regions is assessed by measuring the relative connectivity conferred on their urban centres by the APS networks operating there. Data on office networks at four scopes - regional, national, European and global – are extracted from the World-Wide Web and from business directories to form a regional service activity matrix. This data has been computed by GaWC to show the connectivity of each urban centre, the linkages between them, and thus the degree of polycentricity for each region.

An important limiting factor must be mentioned before we turn to consider the results. The assessment of polycentricity is necessarily comparative between regions. The connectivity values for each urban centre are calculated as a proportion of the first city, hence differences in the world city network connectivity of first cities are not visible in this analysis. This proves a particularly important limitation in considering the cases of London and Paris which have the highest world city network connectivity of the European cities studied.

Here again, the results first show striking similarities between the MCR processes identified for PAR-LON. Compared to other city-regions studied in Polynet, urban centres outside Paris and London are better connected to the world city network. A t the international, European and global scales of network connectivity, polycentricity becomes apparent, reducing the primacy of these world city nodes at the MCR scale (Taylor et al., 2008). Significantly, the upgrading of MCR polycentricity at the global level distinguishes PAR-LON from the other North West European city-regions studied. The reasons are as yet undetermined but the high world city network connectivity of first cities Paris and London seems to be the most likely causal factor. Resultant mul ti-nodality and integration within Castells’ network space of flows, replicates PAR-LON functional integration and polycentricity in the space of places, previously noted.

The inconclusiveness of the analysis as to the reasons for this phenomenon illustrate a further limitation of the GAWC quantitative methodology. Its focus is the space of flows, yet flows of information between places remain theoretical, a postulate of organisational networks of firms. According to the GAWC quantitative methodology, actual flows can only be inferred from the network analysis. Hierarchical and command/control structures that centralise network flows of information are only partially reflected in connectivity weightings and other differences in office roles and functions which can shape inter-city interactions (relational complementarities) are not apparent.

To evaluate relational integration in Paris/Bassin Parisien and London/South East England, it is necessary to access data on city-region circulation of heterogeneous information flows under globalisation and this is not possible on the basis of quantitative data alone. It is semi-structured face-to-face interviews with senior business actors that provide in-depth insights into the effective interactions taking place (or not) in APS office networks and their implications for relational polycentricity.

→ On the basis of network integration markers, as observed through the GAWC quantitative methodology, there is evidence of the emergence of broadly similar patterns of PAR-LON city-region enlargement. However, the qualitative data on the day-to-day practices of APS agents which is discussed next leads to two models that denote an important distinction between effective MCR processes in these two world city-regions. South East England exhibits relational polycentricity versus Bassin Parisien which seems to have a centre-periphery relational model at best, in other words, it is lacking in cross-cutting integration conferred by active agent practices.

Qualitative analysis: APS firms in the city

The focus of the qualitative analysis is the organisational practices of city-based office networks. The methods employed and the potential limitations of qualitative analysis are discussed in detail by Pain and Hall (2008). Data from interviews of 45 minutes to 1.5 hours duration with 529 firms and 106 business, professional and government institutions across the study provides context to the PAR-LON results. In London/South East England, data from a total of 148 taped interviews has been analysed: 120 APS firms and 28 business, professional and economic development organisations. 81 firms located in the seven APS nodes outside London profiled in the GaWC quantitative analysis: Southampton, Reading, Cambridge, Crawley/Gatwick, Swindon, Milton Keynes, Bournemouth/Poole. In Paris/Bassin Parisien, the analysis is based on 60 interviews with firms conducted during Polynet and around 40 achieved more recently. Outside the Paris region urban area, selected nodes were Rouen, Le Mans, Beauvais, Orléans, and Reims. Within the Paris agglomeration, interviews were conducted in key central business districts such as Paris, La Défense and nearby municipalities but also in "secondary" urban nodes such as Saint-Quentin-en-Yvelines, Evry or Roissy.

The major results follow.

First Agglomerations First

The first result is the prevalence of the first city (APS node) for most if not all firms interviewed (in Paris and London but also in Frankfurt, Brussels, Amsterdam, Dublin, Zurich, Dusseldorf). According to static attribute-based approaches, first cities are the places where APS firms cluster, but they are also the nodes in the networked organisations of these firms which makes them fluid spaces or containers of the active relations and interactions between APS agents and associated information flows and knowledge transfer. Based on APS quantitative analysis, along with Frankfurt and Dublin, Paris and London are relatively primate nodes within their regions but PAR-LON in particular are global APS flow spaces, or centralities, that facilitate network connectivity and, as seen, this is the likely reason behind PAR-LON MCR polycentricity at the global level. It is in these central places that the corporate complex finds heterogeneous material and immaterial resources (transport and ICT infrastructures, office spaces and labour) and where APS firms access multinational corporate (MNC) clients, including APS competitors, in addition to local markets. A symbiotic relation exists between the HQ for MNCs in general and APS firms (Halbert and Petit, 2007).

Access to rare and specific human resources is shown to be of even greater importance to APS firms than access to their clients in London and in Paris (for London see also Cook et al., 2007; Taylor, 2003). The depth and breadth of human skills necessary to process the PAR-LON information flows requires the availability, and close co-presence, of many agents with special talents and knowledge. Tacit or coded information exchanges, formal and informal information, club effects and the importance of social networks are quoted in the interviews. These characteristics form "informational districts" (Amin and Thrift, 1992) or rather small and dense spaces of places that have the function of spaces for flows.

The key notions describing these spaces are probably the ones of milieus and of localised productive interactions based on reputation, skills and professional and interpersonal interactions. However, it is important to avoid the suggestion of an opposition between the local and the global in the case of PAR-LON first cities. Although APS professionals are locally embedded in milieus, these can be construed as being transnational in their constitution, APS firms are part of networked organisations that extend to the global level (Global Production Networks which follow MNC location strategies and seek access to worldwide markets) and it is this global level of interaction which distinguishes PAR-LON, as well all other first cities in the Polynet study. As a consequence, this reinforces the interactions between first cities, although in a complex manner, whereby market competition between APS leads to urban complementarities between first cities but, as already discussed, also at the intra-MCR level. PAR-LON first city global connectivity seems the likely cause of a process that interlinks the enlarged MCR to the world city network.

Competition and Complementaries

This question is addressed here specifically in the context of the value chain developed by APS firms. Place competition exists in the case of the location of identical activities by an APS firm. There is a competition between places inherent in the need to access the best set of resources available in different milieus because the trend is toward cutting-down similar activities to reduce cost. But there are complementarities between places as well because of the potential synergies between APS offices (within and between networks) located in different city-regions. Access to new markets is another counter-trend. Commercial strategy leads firms to multiply offices to get closer to clients not only at a world city level, but also within enlarged world city-regions. Increased interactions are occurring between specialised offices (same and different networks) in different cities (between and within enlarged urban regions) as a result of the development of Global Production Networks in APS firms where some offices specialise in key global knowledges, markets and activities.

First cities in each MCR are found to be the most important global knowledge nodes. In all cases they have a concentration of global functions and international high skilled labour. A firm’s ‘people’ are their capital assets and are referred to in Paris interviews as “golden nuggets”! Close proximity for face-to-face contact and intense high-value, high-complexity knowledge exchange take place in first cities - e-mail has become a dominant high volume means of information transfer but in all regions are said to involve lower value exchanges. Mobility and accessibility for face-to-face contact not only remain crucial but seem to be increasing as the pace of virtual exchanges intensifies (Storper and Venables, 2004).

Functional specialisation is identified between offices in first and other cities in Paris/Bassin Parisien and London/South East England but this does not prohibit the presence of complementary functions, relational integration and polycentricity. Indeed such relations and polycentricity are identified in London/South East England (Pain, 2008). However Paris/Bassin Parisien lacks the cross-cutting integration derived from active agent practices, hence relational polycentricity is notably absent.

→ On the basis of relational integration markers, there is evidence of the emergence of a marked difference in the processes underlying PAR-LON city-region enlargement.

PAR-LON: Two Different Spatial Organizations

So an important result concerns not so much the first city but the role of secondary cities for APS firms. This directly addresses the MCR hypothesis as we look at how APS firms use secondary cities in their day-to-day activities.

To access clients, they develop new offices in secondary cities. This creates a direct dependence with central offices in charge of corporate management, research and development (R&D) activities and some specialised production activities. Market-led interdependencies and complementarities occur both in Paris/Bassin Parisien and London/South East England. Secondary cities can also be part of a tasks divisions strategy, for instance to relocate back-office activities in secondary cities. A functional division and complementarity can be found to some extent both in Paris/Bassin Parisien and London/South East England. Last, secondary cities can also be used to offer specific competencies and skills. Knowledge specialisation and complementarity exist in London/South East England but not in Paris/Bassin Parisien. This difference between the regions proves vitally important because it is the key mode for flows of specialist information and knowledge transfer which the Polynet study tries to measure.

Two models of APS strategies therefore emerge: London/South East England is relationally polycentric and MCR-like; Paris/Bassin Parisien is an enlarged urban region but still predominantly hierarchised in terms of flows of knowledge between APS (centre-periphery organisation rather than polycentric).

Elements of the 2 “models”

What possible reasons can explain this crucial distinction between two similar world city-regions? Are differences in the global connectivity of first cities Paris and London an explanation? Are policy and regulatory differences implicated?

The first factor to explain the differentiation between Paris/Bassin Parisien and London/South East England is linked to the location of the clients of APS firms and the demand for their services. All in all, the corporate complex is mostly located within the central agglomeration of Paris and surrounding municipalities (Halbert, 2004, 2007a) whereas this is not the case in London/South East England. Paris has not seen the development which has occurred closer to London in South East England (American MNCs located in the Thames Valley corridor near Heathrow for example). This development includes APS firms, hence in South East England multi-sector clusters are forming which create relational as well as functional complementarities. Because global networks exist outside the first city, albeit international wholesale functions remain concentrated in London, much smaller urban centres gain from global connections, knowledge and specialisms. The Paris/Bassin Parisien does not reflect the same trend.

Second, an element of explanation can be found in the geography of office spaces. Although a large part of the neo-classic analyses postulate that tenant geography is predominant in the location strategies of office spaces, another strand demonstrates the importance of the real estate development industry (Fainstein, 2001) and of investors (Henneberry and Roberts, 2008; Nappi-Choulet, 2006). A dearth of study on their role in reshaping the economic geography of firms working in office spaces, i.e. including APS, limits our current understanding. However, it can be argued that the real estate industry in London/South East England has been much more eager to take risks in developing office spaces in secondary cities of South-East England (Henneberry and Roberts, 2008) than in Paris/Bassin Parisien.

Third, it is clear that public policies matter as well, either because they influence the geography of office spaces or because they have encouraged and/or restricted the development of certain cities / places over others. In Paris/Bassin Parisien mixed policies favour and/or limit the development of secondary cities and this has restricted the capacity of some secondary cities to play a part in the knowledge economy. It has been hypothesised elsewhere that the limited development of secondary Bassin Parisien cities partly results from the massive public investments concentrated on the villes nouvelles (Halbert, 2006). So that deconcentration trends have been restricted to the Paris agglomeration rather than to more distanciated cities, as in the London/South East England case. The development of higher education institutions (universities and Grandes Ecoles) and of a fast transport system (RER) in the villes nouvelles are examples of that. Although green belt policy has restricted development on the outskirts of London, policies allowing secondary city economic expansion in London/South East England, for example, in Reading, new towns, such as Milton Keynes, and more recent growth corridor policy, have facilitated MCR development.


This paper follows suit to the growing literature on global city-regions. It discusses the newly suggested "Mega-City Region" concept of the Polynet research by providing empirical evidence from the Paris and London case studies. The locations and practices of APS firms and professionals are studied to analyse the development of functionally interlinked urban systems spreading around major global cities.

The paper first offers new methodological insights by proposing an operational implementation of Castells "spaces of flows" applied at the intra-metropolitan level. The method does not hold much regard for static attributes, such as the location of activities, although most quantitative data are of this type, but for the daily practices of people, firms, and even institutions, which connect places into a geography of flows. Urban entities can be considered in this capacity as the first and foremost examples of 'flow-places' (Halbert and Rutherford, 2008; Pain, 2007a).

The first finding is that in both cases, as in all other Polynet cities, the first city, i.e. Paris and London agglomerations constitute the anchor points where globalisation flows are processed both quantitatively and qualitatively. It is in the first agglomeration of global city-regions that most interconnections linked to APS networks are located. This confirms how the dense central areas of major global city-regions do contribute to integrating city-regions in the metropolitan archipelago economy. However, Paris and London metropolises remain unique in Europe insofar that the global reach of most of their secondary cities measured by APS firms' network, is considerable. This might be the consequence of a size effect: secondary cities in large metropolitan regions such as London and Paris have a market base sufficient enough to stimulate the development of local APS offices. It can also reflect a network effect where the engagement of the first agglomeration in contemporary global circulations has beneficial impacts on secondary cities that can thus access other non-local markets.

However, the similar and specific pattern observed in both cases still allows very different geographic organisations. London and South-East England is the closest real life example of the MCR defined by Hall and Pain 2006 with intense relational integration. On the contrary, the Paris / Bassin Parisien case offers another interpretation. The spatial organisation may be polynuclear in its forms but, seen through the APS professionals marker, it remains relatively monocentric (hierarchical pattern). The regional competition, including in terms of national government investments, seems to have favoured the closely located villes nouvelles against more distant secondary cities.

A final finding is that although (APS) firms are engaged in fierce market competition, including the ongoing assessment of cities' hub characteristics against one another, it nonetheless holds true that the geography of intra-firm APS networks allows complementarities and interdependencies between global city-regions.

In conclusion, we would like to stress that the use of APS professionals' practices is one way among many others to analyse the ongoing transformations happening within and between global city-regions. There are many alternative methodological possibilities for broadening what has until now remained a rather narrow definition of global activities (to name but a few, cultural industries, regional innovation systems, but also ‘globalisation from below’ with the role of migrants as global/local mediators in world cities, or the role of international investors forging would-be ‘glocal’ governance in large cities (Torrance, 2008), etc.). However, what remains in spite of the narrowly based APS approach is the interest that attaches to the methodology we use. It has proven efficient to deepen our understanding of the transformation of city-regions both internally and in terms of their connections with other places. It is tempting – and seems partly true- to come to the conclusion that the distinction between the two case studies reflects the considerable differences in national and regional spatial planning policies in France and the UK. But an important finding is that knowledge is produced in but also flows through PAR-LON, reflecting complementary functional world city network relations.


Amin A, Thrift N, 1992, "Neo-marshallian Nodes in global Networks" International Journal of Urban and Regional Research 16 571-587

Beaverstock J V, Smith R G, Taylor P J, 2000, "World-City Network: A New Metageography?" Annals of the Association of American Geographers 90 123 - 134

Bourdeau-Lepage L, Huriot J-M, 2005, "On poles and centers: Cities in the French Style" Urban Public Economics Review 3 13-36

Budd L, 1998, "Territorial Competition and Globalisation: Scylla and Charybdis of European Cities" Urban Studies 35 663-685

Castells M, 1989 The Informational City: Information Technology, Economic Restructuring and the Urban-Regional Process. (Blackwell, Oxford)

Coffey W, Drolet R, Polese M, 1996, "The intrametropolitan location of high order services: patterns, factors, mobility" Papers in Regional Science 75 293-323

Cook G A S, Pandit N R, Beaverstock J V, Taylor P J, Pain K, 2007, "The role of location in knowledge creation and diffusion: evidence of centripetal and centrifugal forces in the City of London financial services agglomeration " Environment and Planning A 39 1325-1345

Derruder B, 2006, "On Conceptual Confusion in Empirical Analyses of a Transnational Urban Network" Urban Studies 43 2027-2046

Fainstein S, 2001 The City Builders: property development in New York and London 1980-2000 (University Press of Kansas, Lawrence)

Friedmann J, 1986, "The World City Hypothesis" Development and Change 17 69-83

Friedmann J, 1995, "Where we stand: a decade of world city research", in World Cities in a World System Eds P L Knox, P J Taylor (Cambridge University Press,, Cambridge) pp 21-47

Friedmann J, Wolff G, 1982, "World City formation: an agenda for research and action" International Journal of Urban and Regional Research 3 309-344

Garreau J, 1991 Edge City. Life on a Nex Frontier (Doubleday, New York)

Geddes P, 1915 Cities in evolution (Williams and Norgate, London)

Gilli F, 2005, "Le Bassin parisien. Une région métropolitaine" Cybergeo, 305

Halbert L, 2004 Densité, desserrement, polycentrisme et transformation économique des aires métropolitaines. Interpréter la concentration des activités d'intermédiation dans la zone centrale de la région francilienne, Université Paris-1, Panthéon-Sorbonne, Paris

Halbert L, 2006, "The polycentric city-region that never was: the Paris Agglomeration, Bassin Parisien and Spatial Planning Strategies in France" Built Environment 32 184-193

Halbert L, 2007a, "From sectors to functions: producer services, metropolisation and agglomeration forces in the Ile-de-France region" Belgéo 1 73-94

Halbert L, 2007b, "L'horizon des entreprises franciliennes vu par les appels téléphoniques ", in Données Urbaines, Tome 5 Eds M-F Mattei, D Pumain (Economica, Paris) pp 269-282

Halbert L, 2008, "Examining the Mega-City-Region Hypothesis: Evidence from the Paris City-Region/Bassin parisien" Regional Studies 42 1147 - 1160

Halbert L, Petit T, 2007, "Des sièges sociaux aux activités tertiaires centrales des grandes entreprises en Ile-de-France: développement économique, aménagement et réorganisations spatiales de l'économie métropolitaine. ", (Institut d'Aménagement et d'Urbanisme de la Région Ile-de-France, Paris)

Halbert L, Rutherford J, 2008, "Conceptualising city-regions as flow-places", in Regional Studies International Conference (Prague)

Hall P, 1966 The World Cities (McGraw-Hill, New York)

Hall P, Pain K, 2006 The polycentric Metropolis. Learning from mega-city regions in Europe (Earthscan, London)

Henneberry J, Roberts C, 2008, "Calculated Inequality? Portfolio Benchmarking and Regional Office Property Investment in the UK" Urban Studies 45 1217-1241

Lang R E, 2003 Edgeless cotoes/ Exploring the elusive metropolis (Brookings Institution Press, Washington D.C.)

Nappi-Choulet I, 2006, "The Role and Behaviour of Commercial Property Investors and Developers in French Urban Regeneration: The Experience of the Paris Region" Urban Studies 43 1511-1535

Pain K, 2007a, "'City of London Global Village: Understanding the Square Mile in a Post-Industrial World Economy'", in The Geo-Politics of the City Ed S Barber (European Research Forum, London) pp 19-38

Pain K, 2007b, "Integrating the European Space – Flows and places in North West European city-region networks", in Cities and Networks in Europe Ed N Cattan (John Libbey Eurotext, Montrouge France, Montrouge France)

Pain K, 2008, "Examining Core-Periphery Relationships in a Global Mega-City Region - The Case of London and South East England" Regional Studies 42 1161-1172

Pain K, Hall P, 2008, "Informational Quantity versus Informational Quality: The perils of navigating the space of flows" Regional Studies 42 1065-1077

Peck J, Tickell A, 2002, "Neo-liberalizing Space" Antipode 34 380-404

Reich R B, 1991 The Work of Nations. Preparing ourselves for 21st-Century Capitalism (Vintage Books, New York)

Rutherford J, 2004 A Tale of Two Global Cities: Comparing the Territorialities of Telecommunications Developments in Paris and London (Ashgate, Aldershot)

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

Sassen S, 2002, "Locating Cities on Global Circuits", in Global Networks, Linked Cities Ed S Sassen (Routledge, New York) pp Bulletin 46, pp. 41-36

Scott A, Soja E, Agnew J, 2001 Global City-regions: Trends, Theory, Policy (Oxford University Press, Oxford)

Short J, Kim Y, Kuss M, Wells H, 1996, "The dirty little secret of world cities research " International Journal of Urban and Regional Research 20 697-717

Storper M, Venables A J, 2004, "Buzz: face-to-face contact and the urban economy" Journal of Economic Geography 4 351-370

Taylor P J, 2001, "Urban Hinterworlds: Geographies of Corporate Service Provision under Conditions of Contemporary Globalization" Geography 86 51-60

Taylor P J, 2003, "Financial Services Clustering and its Significance for London", (Corporation of London, London)

Taylor P J, Evans P, Pain K, 2008, "Application of the Inter-Locking Network Model to Mega City-Regions: Measuring Polycentricity within and beyond City-Regions" Regional Studies 42 1079-1093

Thiard P, 1998, "Les mutations du système productif et de l'activité dans le Bassin Parisien" Travaux de l'Institut de Géographie de Reims 25 31-58

Torrance M I, 2008, "Forging Glocal Governance? Urban Infrastructures as Networked Financial Products" International Journal of Urban and Regional Research 32 1-21

Veltz P, 1996 Mondialisation, villes et territoires l'économie d'archipel (Presses universitaires de France, Paris)



* Ludovic Halbert. Latts, Université Paris-Est, UMR CNRS 8134, Email:

** Kathy Pain, Department of Geography, Loughborough University

1. European Commission INTERREG IIIB North-West Europe: ‘POLYNET: Sustainable Management of European Polycentric Mega-City Regions’, 2003-06


Edited and posted on the web on 18th June 2009