Economic transformation processes, the rise of the knowledge economy and the application of new technologies have each played central roles in the changed focus on cities in both policy terms and in spatial research over the last three decades. Within the political debate cities have become increasingly important strategic elements particular in terms of the on-going discourses around territorial competitiveness and cohesion (e.g. OECD, 2006; Ministers responsible for Spatial Planning and Territorial Development of the European Union, 2011). Specifically a more economic figure of thought suggests that some specific cities such as New York, London, Singapore and Paris are playing a critical role in the global network economy. These cities are conceptualised as being central nodes in the spaces of flows since they are (apparently) equipped with the appropriate urban assets to take part to a relatively high extent in transnational flows of capital, commodities, knowledge, labour, tourists and cultural symbols than cities at a lower level of the urban hierarchy. Consequently, they are often labelled as ‘world’ or ‘global cities’ (cf. Friedmann, 1986; Sassen, 1991; and regarding the differences between the two: Derudder, 2006), since they function as international ‘hubs’ for instance in relation to the interaction of skilled labour and their ‘tacit knowledge’, as financial control centres or as the major points of origin for the generation of different kinds of innovations. In this sense they entail new centralities and spatial inequalities. Consequently, the much of the related policy discourse is targeted at the extent to which a given city can be considered as providing the competitive urban assets to sustain (or even improve) its socio-economic performance in a globalising world. As a response to this, a growing demand for handy information can be recognised, which is saturated by a number of international business magazines or consultancies that base their investigations upon some specific and partly questionable indicators to produce rankings about the world- or global-‘cityness’ of a selected number of candidates (cf. the recent study by the international consulting company ATKearney in 2012).
The academic debate surrounding these issues has focused in particular on the emerging networks of cities and their (re-configured) hierarchical relations caused by (more or less) tangible labour divisions and functional differentiations. This has been followed by comparatively distinct selections and rankings of cities at both the national and, particularly, at the international scale (see for example Friedmann, 1986; Taylor et al., 2010; Alderson et al., 2010). As a response to the reported geo-economic transformations we can easily witness an increasing global intercity competition to attract first-class technological, social and cultural infrastructures (i.e. real urban assets), human resources and transnational firms. However, despite the flattening potential of information and communication technologies (ICT) and decreasing transport costs, the aforementioned flows are not initiated by, or directed to all cities in the same way, so that the world, in the words of Florida (2005), can be still described as a ‘spiky place’, since only a small number of cities can be characterised as ‘outstanding’ urban nodes in this respect. Other cities can be rather qualified as being of secondary (or even less) importance, be it at the national, at the European or at the global scale. These considerations formed the starting point for our study, namely to position a number of Northern European cities within this changing geography of uneven urban development at different scales.
Nonetheless, the aforementioned geo-economic transformation has inevitably also changed the perception of cities, or city-regions respectively, from sub-national, bounded areas to nodes in global networks (Sassen, 2001) and ‘regional motors of the international economy’ (Scott, 2001). In this paper we depart from such a relational perspective on cities in our multi-scalar ‘bottom-up’ analysis of the Nordic capital regions’ inter-city connectivities through finance and market service firms and ICT service firms. Methodologically the study is inspired by the work of the Globalization and World Cities Research Network (GaWC, cf. Taylor et al., 2010). Similar to the POLYNET-study (cf. Hall and Pain, 2006) a number of metropolitan areas (here Stockholm, Helsinki, Oslo and Copenhagen) were selected as geographic departure points. Through our ‘modified Nordic-based bottom-up study’ using data sources from 2010 we quantitatively explored the network connectivities of cities through the multi-scalar locational strategies of ‘Financial and Market service firms’ as well as ‘ICT service firms’. In other words, we analysed the extent to which our sample of firms in the given branches connect each of these four Nordic capital regions with other cities at the national, Nordic, European and global scales. The calculated values define aggregated linkages between each Nordic capital region and all other cities at the respective scale by the presence and functional importance of offices (in the respective sector). Hence the values express the extent to which cities are interconnected through the analysed sample of firms by considering only those intra-firm networks that incorporate (in terms of office locations) at least either the Stockholm, Helsinki, Oslo or Copenhagen metropolitan area.
The paper begins by discussing the underlying specific conceptual approach as well as the applied methodology, data sources and also the restrictions of this survey and reflects critically on the explanatory power of the approach. The results are then discussed and contrasted between the four Nordic capital regions, the two chosen sectors as well as the four scales (national, Nordic, European and global). Specific focus here is placed on the sector and scale sensitivity, i.e. the different patterns in regards to the connectivity of the Nordic capital regions related to other cities within their respective countries, at the Nordic, but also at the European and global scale. The paper finalises with discussing some implications for research and urban policy.
Conceptualising intercity connectivity
Cities (or city-regions respectively) are connected by flows of information, capital, goods, and persons, travelling along infrastructure routes such as roads, railways, airlines and, increasingly, telecommunication linkages. Contemporary urban studies place great emphasis on the significance of these networks in explaining the economic, social and cultural functioning of cities (cf. Castells, 2002; Devriendt et al., 2010; Sassen, 2010; Taylor, 2004). In so doing, the authors emphasise that by using such a relational perspective it is possible to complement a more static understanding of place-based functions and infrastructures and thus to overcome to focus solely on the city’s critical mass (e.g. in terms of population or jobs) and its infrastructural endowment. To study such flows empirically, as well as to assess the factual integration (the so-called connectivity) of the city at hand in the resulting inter-city networks is, however, a challenging task.
The Globalization and World Cities (GaWC) Research Network (cf. e.g. Taylor, 2004; Taylor et al., 2010), has approached to capture cities external networks through analysing and mapping ‘intra-firm office networks’. The main empirical focus have been on ‘Advanced producer service firms’ (APS) because of their international reach and importance in the formation of the global economy, which are by definition almost similar to those that we call here KIBS (see below). Since a direct measurement of the myriad of such flows is hardly possible (cf. the discussion by Derudder 2008 or the web-survey undertaken within the aforementioned POLYNET-study, which could not provide in all respects robust empirical results, cf. Hall et al., 2006) the so-called ‘interlocking network model’ developed within GaWC is used as a surrogate regarding the analysis of the intensity of flows between office locations and, more generally, between cities in the world economy. In this perspective, it shall help to interpret advanced producer service firms as key actors in world-city network formation.
Based on the observation that many such APS-firms have created global networks of offices to provide a seamless service to their corporate customers, one can consider each office network as the outcome of a firm’s specific locational strategy. Namely, through flows of information as well as people and their tacit knowledge such firms build-up linkages between their offices and, more abstractly, between the cities where these offices are sited.
In recent years this interlocking network model of the Globalization and World Cities (GaWC) Research Network has been enlarged and fine-tuned. From a spatial and methodological perspective two approaches can be distinguished: a so-called ‘top-down approach’ that studies the world city network from the perspective of the world’s largest advanced producer service firms (cf. Taylor, 2001; Taylor et al., 2010; Taylor et al., 2011), and a ‘bottom-up approach’ that considers key knowledge-intensive firms located within a selected number of cities or metropolitan areas (cf. Hall and Pain, 2006; Growe and Blotevogel, 2011; Lüthi et al., 2010; Lüthi et al., 2011). The present study has used such a bottom-up approach by using the four Nordic capital regions (Stockholm, Copenhagen, Helsinki and Oslo) as starting points for our investigations in a multi-scalar perspective.
Nevertheless, one should not forget that the interlocking network model (and its several modifications and applications) represents just one approach to the approximation of very specific flows. It allows only for the construction of a rather rudimentary picture in respect of inter-city connectivities as it does not capture at all the imaginative multifarious flows related to commodity chains and their inherent logics nor all thinkable places which actually constitute the nodes and hubs for the origin, destination and crossing of such flows (cf. Castells, 2002; Sassen, 2010). A promising approach in this direction has been provided and empirically tested by Lüthi et al. (2010). Based on the value-chain approach the authors have obtained spatial information through a web-survey on firms to trace their extra-firm relationships. Other studies have, for instance, used transport flows to get an indication of the extent to which cities are interacting. Here indicators such as the origin and destination of flights have been used (cf. the analysis in Schmitt/Dubois 2008:58-60 for the metropolitan areas in the Baltic Sea Region). Other more in-depth approaches consider data on airline bookings as a better estimate of the factual flows of passengers between (airport) cities (cf. Derudder et al., 2010a), although the problem still remains to filter out, for instance, business related flights from those motivated by other reasons. This approach also overemphasises those cities with well-equipped airports and corresponding strategies by the operating airlines and thus tends to neglect second or third tier cities in national urban systems, as they are generally less integrated into international airline networks. In addition it is important to bear in mind that the hub and spoke system is not only shaped by e.g. the infrastructural endowments and market sizes of city-regions, but also by the specific strategies and capacities of the various airlines involved to exploit what they consider as being ‘profitable’ air links. But for all that, one can still agree with Smith and Timberlake (2002) who argue that the so far investigated connectivities represent an extremely small portion of the city’s complex functional profile as being a node of economic, social, cultural and political life.
Methodological notes and limitations
The methodology used in this study is as mentioned largely based on the work of the Globalization and World Cities Research Network (GaWC), but has been modified in some crucial respects in order to respond to the specific objectives of this study. In contrast to the GaWC top-down approach, which basically uses global lists of large firms (e.g. provided by the ‘Forbes Global 2000’ list) as a starting point to explore (in this sense ‘top-down’) the extent to which firms belonging to a specific sector inter-connect cities through their office networks, we decided instead on a ‘bottom-up approach’ similar to the POLYNET-study. Here two aggregated knowledge intensive business service sectors (KIBS) were considered: finance and market service firms (‘FMS-firms’) including financial and insurance activities, legal and accounting activities, activities of head offices; management consultancy activities and advertising and market research (NACE codes 64-66, 69, 70 and 73; and information and communication technology service firms (‘ICT-firms’) including telecommunications, computer programming, consultancy and related activities, and information service activities (NACE codes 61-63). Hence the analysis indicates the extent to which intra-firm office networks belonging to the two sectors mentioned above incorporate the four Nordic capital regions with other cities. In so doing, we have considered so-called ‘non-local firms‘ that have, for instance, at least one office in the Stockholm capital region and at least one other office elsewhere.
In order to limit our database for each Nordic capital region we had to define a meaningful roster of cities at the respective scale. At the national and Nordic level only municiapalities with more than 20,000 jobs where included (in total 102 Nordic cities outside the four captial regions). The data roster also included 341 European cities with a popualtion over 200,000 and 374 non-European cities with a population over 1,000,000. In total over 800 different cities where included in the study, which is more than double the amount of many other aformentioned related studies. The reason for this relatively large and strictly defined roster is that we wanted to base our mapping on a somewhat comparable criterium and eventually reveal a somewhat different picture (in particular at the globale scale) compared to former studies. Consequently apart from additional Nordic and European cities our study also included cities like Pune and Kolkota in India, Cali in Coloumbia, Busan in South Korea, San Antonio in the US, and Port Harcourt in Nigeria, and Chinesee cities such as Chengdo, Dalian and Hangzhou, non of which is usally included. Following our strict criterium, the study does however neglect for instance such US-cities with less than one million inhabitants which are normally included in this kind of studies, namely for instance Boston, Miami, San Francisco, Seattle, and Washington.
Otherwise, the research design in principal followed the data collection and processing for a ‘top-down approach’ as illustrated in Taylor et al. (2010: 412-425) as regards steps 6 and 7 (see table 1, below), whereas steps 1 to 5 are specific modifications for our ‘bottom-up’ approach which uses the four Nordic capital regions as ‘focal points’ . A crucial step here was undoubtedly to study the functional importance of each office within its intra-firm network. Here the question was to assess whether the particular office under consideration in the city at hand was, for instance, the firm’s headquarters or only a minor office with few significant employees. In sum, as proposed in Taylor et al. (2010: 413), we have studied:
Table 1: Research design: A ‘seven-steps’ process
Because the form of the information gathered is unique to each firm we needed to convert it into a common data matrix to enable comparisons. Service values indicate how important a city is within the operation of a firm’s overall network of offices. To illustrate the results of this exercise we calculated so-called ‘network connectivites’, which expresses the aggregated linkages between each Nordic capital region and the respective city at various scales in the respective sector, expressed in proportionate terms as a network connectivity index. The word ‘aggregated’ here means that the values incorporate not only the connectivity of e.g. the Finnish city of Turku in direct relation to the Helsinki capital region in the sector at hand, but also the extent to which Turku is connected to all other cities at this particular spatial scale (here for instance within Finland). Hence the network connectivity index value of 48 for Turku (here ICT sector and related to the Helsinki capital region, which is per definition valued as 100) indicates the extent to which Turku is interconnected within the ‘Finnish’ urban system through the sample of ICT firms (all of which have at least one office in Helsinki) used for this study.
As just touched upon above through the use of Turku as an example, this methodology allows us not only to generate some interesting insights in respect of the extent to which ‘other Nordic cities’ (e.g. Reykjavik, Odense, Gothenburg, Bergen or Pori) are interconnected with the Nordic Capital Regions in the first place, but also to which the intra-firm office networks that include the Nordic Capital Regions can link them with other cities at the respective spatial scale (e.g. national or Nordic). The same goes for the European (here non-Nordic) and global (here non-European) scales: for instance the proportionate connectivity of London related to the Copenhagen capital region is 135 in the FMS sector, which means that London is not only highly related to Copenhagen, it also shows a ‘very’ high connectivity in relation to all other cities at this particular spatial scale (here the European one). Despite the fact that in this particular dataset only non-local firms that have at least one office in the Copenhagen capital region were considered, London displays an even higher connectivity than that of Copenhagen (which is per definition 100). This indicates London’s central position in this sector per se (which is also confirmed by the aforementioned ‘top-down study’ by Taylor et al., 2011) as it is literally ‘a spider in the net’ that (at least potentially) inter-connects firms ‘across (and maybe even beyond) this particular sector’ in many European cities (and certainly also beyond). Using the same set of firms (and their corresponding intra-firm networks related to the Copenhagen capital region) at the global scale, New York City and Singapore can also be easily identified as such central hubs, since their network connectivities are also above 100 and thus higher than that of the Copenhagen capital region. This is noteworthy insofar as we should recall that we have considered here a ‘Copenhagen biased set of firms’, since all firms considered here have an office in the Copenhagen Capital Region, but apparently most of them also have an office in London, New York and Singapore which is, in many cases, of even higher functional importance.
To sum up, the network connectivity value can be understood as a potential approximation of ‘intra-firm’ linkages in the city at hand. A comparatively high value thus expresses the fact that the city at hand has a (relatively) high ‘intra-city connectivity for the sector that is under consideration expressed by the sample of firms studied. Having said this, it is worth mentioning once again that such ‘linkages’ are captured through the physical office location and functional importance of a particular firm (or a set of firms). Hence, these linkages are thus the result of strategic place of location decisions by the respective responsible persons of the firm(s) at hand. In addition, we should stress here once again that terms like linkages or even relations remain problematic, since we have not considered how those can be further characterised nor how (or to what extent) they are in fact articulated through particular flows (e.g. communication and information, different sorts of knowledge, financial means etc). The only indication in this respect is that not only the presence of the particular office of firm ‘a’ or ‘b’ in city ‘x ‘or ‘y’ has been analysed, but also its functional importance within its intra-firm network. Here the question was to justify whether the particular office under consideration in the city at hand is, for instance, the firm’s headquarters (which received the value 5) or an office of minor importance with few employees (which received the value 1). Consequently, as noted previously, the applied methodology can be viewed as a surrogate for the anticipation of linkages or relations between firms and, in this sense, also cities, due to the firm’s physical presence (and service work) there.
As such then, the results presented here illustrate only an understanding of how firms, in relation to the geographical outreach of ‘their’ offices in a particular pre-defined set of cities, try to serve specific markets (e.g. to be close to their clients/other firms etc). So we can see that firm ‘a’ has developed (perhaps over the period of some decades) a unique internal office network. We assume that there are some strategic decisions behind this though without knowing the exact rationales, which could provide a useful starting point for a follow-up study. As noted previously, by calculating the network connectivity values we consider how the city under consideration inter-connects a specific sample of firms (at least potentially) ‘across this particular sector’. Again, we can say for instance that in London with a comparatively high connectivity value of 135 (here expressed through intra-firms networks that incorporate the Copenhagen capital region) there is apparently a high potential for factual ‘inter’ (or extra) -firm linkages across this sector (here financial and market services, and maybe beyond), although we have not studied such particular relations per se.
A final remark which will help to set our results in their proper context is that we have selected our sample of firms based on a specific Nordic database that gave us data about the physical presence of firms in the respective Nordic Capital Region and sectors (ranked by its annual turnover). As the selected set of non-local firms with the highest annual turnover differ for the most part between the four Nordic Capital Regions, the direct comparability between them is limited. Naturally, the same goes for comparisons between the two chosen sectors as the set of firms also differ. Direct comparisons across scales within one Nordic Capital Region are, however, reasonable as they stem from the same set of firms.
Intercity connectivities of the Nordic capital regions
In the following section the results of the quantitative study exploring the ‘inter-city connectivities’ of the Nordic capital regions through intra-firm office networks are presented. The empirical quantitative study produced a large amount of data that can be analysed and presented in various ways, for example in thematic maps as illustrated below. The presentation here focuses on the 25 cities with the highest intercity connectivity index values of each of the four Nordic capital regions on the four different scales. Due to the restricted space here, we have selected only one map per scale. NOTE: In case the paper will be published we would like to add here a hint to a working paper that includes more maps, data and background information of this study.
Intercity Connectivities at the National and Nordic Scale
The intercity connectivity of the Nordic capital regions on the national scale is relatively weaker than on the other levels (Nordic, European, global). According to the calculated connectivity values we can also recognise a significant gap on the national level between the Nordic capital regions and the other national cities, since most of them have a network connectivity index value below 50. Their relatively low connectivity becomes even more visible if the national maps are compared with the global maps, where we can find the highest network connectivity index values. One reason is that we have selected our ‘Nordic-biased sample of intra-firm networks’ (in total 322) based on the highest annual turnover in 2009. In other words, the focus was on large, often multi-national firms which tend to span their intra-firm networks over a number of larger cities worldwide instead of incorporating a number of rather mid-sized Nordic cities. Nevertheless, the ‘gaps’ at the national level in each Nordic country confirm the monocentric national urban systems there on the one hand, and indicate that the Nordic capital regions are highly interconnected with non-domestic cities on the other. Despite of this, we can easily discern a second tier of national cities: Aarhus and Alborg in Denmark; Bergen, Trondheim and Stavanger in Norway; and Tampere, Turku and Oulo in Finland. In the Swedish case, Gothenburg and Malmö stand out after Stockholm followed by a range of what might be considered as a third tier of national cities (see map 1).
Figure 1: Stockholm’s national intercity connectivity through FMS and ICT-service firms
If we consider the five Nordic countries (Iceland, Norway, Denmark, Sweden, and Finland) as one macro-region, we can observe that Swedish cities ‘dominate’ Nordic maps of intercity connectivity (see for example map 2). In addition, Stockholm stands out at the Nordic scale in terms of relatively high intercity connectivity through FMS firms. Compared to this, there are considerably more variations existing in respect of the ICT service sector, since we see a rather uneven geographical pattern when looking at the connectivity values for each of the Nordic capital regions at the Nordic scale.
Figure 2: Oslo’s Nordic intercity connectivity through FMS and ICT service firms
European – Nordic Intercity Connectivities
As in many related studies, London comes out to be the main European finance and market service city. London’s network connectivity is above 100 in all four cases within the FMS sector. Other important cities with strong linkages to the Nordic capital regions within the finance and market service sector are Moscow and Warsaw. Paris appears to be the major European ICT city seen from the perspective of the Nordic capital regions. Strongly interconnected ICT service cities in Europe, alongside Paris are Madrid and to some extent Milan Brussels and Munich, but to a much lesser extent London for instance. This general pattern reappears throughout the Nordic capital regions, but is perhaps most clearly visible from a Copenhagen perspective.
In general, one can state that cities in Eastern Europe seem to be as significant for FMS firms that are linked to Nordic capital regions as Western European cities, while ICT service firms are more oriented towards Western Europe. This is most obviously displayed in Oslo’s European intercity connectivity pattern and least pronounced in the case of Helsinki (see map 3). Most of the European countries only have one city represented in the list of the top 25 connected cities when considering both sectors together. An exception is Germany with up to four cities, which corresponds to some extent to the in-depth analysis by Műnter (2011) on Germany using GaWC data from the ‘top-down’ study in 2008 (see Taylor et al., 2010). Italy with Milan and Rome, and Spain with Barcelona and Madrid have two cities represented. Other larger countries such as the United Kingdom, France and Russia only have one city represented which means that cities such as Birmingham, Lyon and Saint Petersburg are not particularly interconnected from a Nordic perspective.
According to our data Baltic Sea and East European cities are strongly interconnected with Nordic capital regions particularly through FMS firms. Copenhagen is for example strongly related to Baltic and East European cities. Warsaw stands out with a high network connectivity index value for both Oslo and Copenhagen. Oslo is furthermore strongly interconnected with other Baltic and East European cities such as Tallinn and Riga, Moscow and Prague, but there are also strong connectivities with Balkan cities through FMS firms too. From the perspective of FMS firms in Stockholm East European cities (Moscow, Warsaw and Prague) seem to be as important as Western European ones (Paris, Amsterdam, Madrid, Frankfurt and Zurich).
The cities of the Baltic States have relatively little importance for the Helsinki region’s European connectivity through ICT service firms (less than for example Oslo and Copenhagen) (see map 3). The spatial pattern of ICT service firms related to Oslo is compared to the other three Nordic capital regions less dispersed. Regarding Stockholm ICT service firms are more oriented towards Western Europe cities, particular Paris, but also Madrid and Brussels. In sum, it appears that from a Northern European perspective the former division in east and west has faded especially when considering the FMS sector.
Figure 3: Helsinki’s European intercity connectivity through FMS and ICT service firms
Global – Nordic Intercity Connectivities
Our specific ‘Nordic-biased dataset’ reveals that Singapore and New York are the two cities that stand out regarding global intercity connectivities and are thus consistent with related studies (e.g. Taylor et al. 2010). Although American and East Asian cities dominate at the global scale a number of other high connected cities can be recognised such as Dubai, Johannesburg and Sydney. New York, with a network connectivity value above 100, is as expected, an unambiguous global city for the FMS firms that are (also) located in the Nordic capital regions. Singapore also stands out in the FMS sector ranked as the second city, but more importantly to notice is that Singapore is the most connected city for the global intra-firm networks of the ICT service sector from a Nordic perspective. It is notable that Hong Kong, for example, seems to be of lesser importance from a Nordic perspective, in contrast to other studies (with the exception of the Helsinki capital region).
It is also notable that South American cities generally have higher network connectivity values than North American cities. The strong intercity connectivities of South America are most obvious in the case of Copenhagen (see map 4). It should, however, be noted that central ICT cities in Northern America, such as Boston, San Francisco and Seattle are not included in the study, since they do not have a population of one million or more (see methodological notes above). This does not however account for the fact that apart from New York, only in the ‘Stockholm related data set’ generally high ranked cities such as Los Angeles and Chicago are included. Among the Southern and Latin American cities especially Sao Paulo, Buenos Aires and Mexico City stand out. According to our results New York and Sao Paulo, for example, are equally significant regarding the global network connectivity of Copenhagen’s ICT service sector (see map 4). Toronto also seems to be relatively well connected, especially for FMS firms that are linked to the Nordic capital regions.
East Asian cities generally occupy rather high positions within the intercity networks in our ‘Nordic-biased database’. Singapore, as mentioned before, stands out as a city with network connectivity values around 100 in both sectors for all four Nordic capital regions. The three Chinese cities Beijing, Hong Kong and Shanghai are also strongly interconnected with all four Nordic capital regions. For Copenhagen’s ICT service sector Singapore, followed by Tokyo, Beijing, Hong Kong and Seoul seem to be strategic places, and more or less the same applies for ICT service firms linked to Stockholm (only with the addition of Bangalore, instead of Seoul). Dubai stands out as a comparatively strong connected city in the global network of ICT service firms that are related to Oslo. The same can be said, although to a lesser extent in the case of Copenhagen and Stockholm. From our ‘Oslo data set’ we can also read out that FMS firms are somewhat more oriented towards the Middle East, where Istanbul stands out as of being of particularly strategic importance. The latter is also visible in Copenhagen’s intercity connectivity pattern.
The Indian cities of Bangalore, Mumbai, and Delhi seem to be relatively well linked in the ICT service sector, since they have relatively high network connectivity index values from a Nordic perspective. Bangalore, for example has the second highest value in our calculations for Stockholm’s global connectivity through ICT service firms. Regarding FMS firms, India is of less significance for the Nordic capital regions’ connectivities with the exception of Mumbai in the case of the global connectivity pattern related to Copenhagen (see map 4).
Johannesburg is the African city with relatively strong linkages in both sectors related to all four Nordic capital regions (an exception is Oslo regarding the FMS sector). In addition, Casablanca is also represented among the top 25 cities in the Helsinki data set (both sectors), and also through ICT service firms that are linked to Copenhagen. Last but not least, from the perspective of the Nordic Capital Regions, Sydney is a significant city for both sectors that have been studied here.
Figure 4: Copenhagen’s global intercity connectivity through FMS and ICT service firms
Discussion: a new geography in the making?
Using this ‘bottom-up’ approach, which conceptualises the external relations of the four Nordic Capital Regions through the lenses of KIBS-firms, reveals a number of interesting results. Firstly, in relative terms, it confirms for the most part the latest results of the global ‘top-down’ approach (cf. Taylor et al., 2011), namely the comparatively strong global significance of Stockholm, followed by Copenhagen and then Oslo and Helsinki (although the latter two seem to be much closer in terms of their factual inter-city connectivity in our ‘bottom-up’ approach). However, one can also conclude that our ‘scale-sensitive’ approach complements the aforementioned top-down study in many respects. It should also be noted here that we have defined our two sectors slightly differently as compared to the GaWC approach (cf. Taylor et al., 2010).
Nevertheless, our results at the national scale indicate quite strongly the relative monocentricity of the Nordic national urban systems. One explanation for this is the unitary nature of the Nordic nation states compared, for instance, to federal systems which might have fortified this picture. In addition, perhaps related to this, the relatively small critical mass of the Nordic national cities here in terms of their city-regional market potentials for KIBS-firms is another important aspect for consideration here. Moreover, it is clear that the four Nordic countries (Denmark, Finland, Norway and Sweden) themselves comprise small markets. In other words, many larger KIBS-firms serve the national market with only a small number of offices (in many cases with only one office located in the capital region).
In this context then it is perhaps not that surprising that Swedish cities are more prominent , at the Nordic scale,, since they belong to a country with an overall market size that is double that of any of the other Nordic countries. Another potential factor here might be the comparatively high critical mass (in terms of market potentials, here other firms as clients) of the Stockholm region. The Copenhagen capital region has more or less the same weight in this respect and might also profit from its closeness to the European mainland and the higher significance of its international airport (as compared to Stockholm, but also to the other two Nordic capital regions, cf. Schmitt and Dubois 2008). Nevertheless, these are rather speculative assumptions based on a highly rationalistic outlook in relation to market sizes, infrastructure endowment and geographic position. Hermelin (2010) for instance argues in her commentary on the global top-down approach of the GaWC approach (cf. Taylor et al., 2010) that the ‘integrated history’ of the Nordic states has resulted in relatively dense economic, social and cultural connections. This may then also support the perceptions held by strategic decision-makers, regarding intra-firm networks that shall cover the Nordic countries that only a small number of offices is required to service this (comparatively) well integrated market.
At the European scale it appears that the Iron Curtain has to a large extent been overcome. Indeed London and Paris stand out due to their high-degree of world ‘cityness’ (cf. also Taylor et al., 2010; 2011). But looking at the other non-Nordic European cities we see a rather balanced west-east picture. It would be interesting to compare these results with similar ‘bottom-up’ studies taking other European cities (such as Madrid, Brussels, Warsaw or Prague) as starting points to see if they come up with likewise results. Nevertheless, it appears that the four Nordic capital regions are (in this context) well-connected with these relatively fast growing markets. Perhaps the most eye-catching result in this respect is that the actual geographic proximity to the cities of the Baltic States (particularly the comparatively small markets in and around Tallinn, Riga and Vilnius) is mirrored by strong ‘intra-firm proximity’ in the FMS sector, whereas these linkages are less pronounced regarding the ICT service sector.
From our perspective the most surprising results, particularly compared to the most recent global top-down approaches (cf. Taylor et al., 2010; 2011), are to be found at the non-European (i.e. global) scale, which can be certainly traced back to our comparatively large roster of cities, which is as such fairly receptive regarding otherwise neglected cities. Our results differ from the recent GaWC studies, since they apportion comparatively more weight to cities in the BRIC countries and other rapidly growing city regions in Africa and Middle East, Central and South America as well as West and South West Asia. Most of the cities with the highest global network connectivities (GNCs) in the latest GaWC studies are nevertheless generally also to be found in our Top 25 lists (such as New York, Tokyo, Shanghai, Singapore). Perhaps most interesting to note is the extent to which the pattern of cities belonging to global macro-regions (such as Central and South America or Pacific Asia and Oceania) changes between the four maps of the Nordic capital regions. Here further investigations are required in order to incorporate for instance, historical and geo-political path-dependencies. It does however appear that the Nordic capital regions seem to be well-connected with these emerging or fast growing macro-regions (and respective cities). Still one needs to bear in mind here that the KIBS firms that connect the Nordic capital regions with these markets in our sample are for the most part ‘non-Nordic based firms’. In other words, we have to understand these maps and findings as representations of decision-makers that are located for instance in New York or London and thus consider the global market potentials of their firm not necessarily from a Nordic point of view. Also it is noteworthy that the two KIBS sectors studied here produce slightly different geographical patterns. Regarding ICT service firms it is Singapore and Paris which can be said to be the most globally interconnected cities in our Nordic biased data set in contrast to the rather usual suspects that are among the Top 5 regarding FMS firms. Hence our study offers also in such a sectoral perspective some complementary results compared to the GaWC approach, which have neglected so far ICT service firms as significant ‘world city networkers’.
Implications for research and policy
The modified Nordic bottom-up approach complements the results of the numerous top-down studies executed within the GaWC group and thus provides additional network geographies of cities that are sensitive to scales and sectors. It thus contributes to the ongoing development of research approaches and analysis of cities’ external relations. However, although we have anchored our Nordic-biased study in a large and extensive data base following mainly the bottom-up Polynet approach (Hall and Pain 2006) the methodological and structural limitations of the underlying GaWC ‘interlocking model’ still remain as recently debated by Liu and Derudder (2012), Hennemann and Derruder (2012) and particularly Neal (2012). Although we can concur with the addressed critics and the various proposals for improvement, we would like to add that further research needs to include in particular qualitative in-depth studies in order to better understand the rationales behind the observed intra-firm and, finally, anticipated intercity networks. When presenting our results from this study to policy-makers they have asked immediately a number of obvious questions such as ‘why do firms choose to establish an office in one Nordic city rather than another’ or ‘how many offices are actually needed to serve the specific market that the KIBS firm in question is operating in’. In other words, besides the more fundamental and structural critique we think there is also a need to better construe the results from such ‘macro studies’ as presented here by exploring the rationales and functioning of intra-firm networks and the role of ‘the city’ therein. Only in this way we can eventually re-draw our findings and better validate the factual position of the city at hand regarding the one or other intra-firm network (or when applying a larger study to give a better account about ‘the sector’s’ specific network geography). Hence, in more general terms, we want to remind that future global urban research needs to go beyond structures delving into practices, by exploring the contents of the spaces of flows and the related network actors.
One of the central lessons of this study regarding policy implications is undoubtedly the need to challenge established notions of the national, Nordic, European and even global urban systems from a Northern European perspective. Although critical mass aspects have also had a certain influence on our results (here regarding the selection of KIBS firms due to their annual turnover and the roster of cities) our study follows a ‘relational approach’ to the consideration of current national and international urban systems. As such it challenges traditional mental maps of the relational position of the one or other city. Within our interventions of Nordic policy-makers that have followed-up this study from the beginning to the end we could easily observe the extent to which their mental maps are anchored in absolute place-based attribute. Another eye-catching difficulty in formulating concrete policy implications for this target group is that neither ‘high’ nor ‘low’ indices of intercity connectivities provide any concrete indication of the city’s economic performance, social-economic well-being or factual competitiveness.
Nonetheless, in respect of policy implications we think that the fundamental message is simply the need to take into fuller consideration such ‘relational representations’. They challenge our traditional maps as well as our policy frameworks in so far as they focus on the networking agent (i.e. the firm) and the decisions/practices of specific key actors instead of a static ensemble of institutions and infrastructures which are qualified in terms of market sizes and/or the economic performance of spatial units within jurisdictional boundaries. This figure of thought still organises the spatial allocation of policies, programmes and projects – with the EU structural funds being a particularly eye-catching example of this. Certainly such a relational perspective is less tangible and before formulating network oriented policies to complement ‘place-based’ policies, it is imperative to vigorously interrogate the nature of such relations and what could be done in order to optimise them. This study represents then a first step in this direction by carefully considering the multi-scalar scope of these ‘interlocked urban systems’ as well as its sector sensitivity.
This paper bases upon work accomplished and initiated by the Working Group for the exchange of experience and knowledge development (Urban Policies) under the Nordic Council of Ministers. The authors would like to thank the members of this group, who represent national ministries from the five Nordic Countries (Denmark, Finland, Iceland, Norway and Sweden), for their feedback, support and for the stamina they showed throughout this study.
Alderson A, Beckfield J, Sprague-Jones, J (2010) Intercity Relations and Globalisation: The Evolution of the Global Urban Hierarchy, 1981-2007. Urban Studies 47: 1899-1923.
Castells M (2002) Local and Global: Cities in the network society. Tijdschrift voor Economische en Sociale Geografie 93: 548-558.
Devriendt L, Derudder B and Witlox F (2010) Conceptualizing digital and physical connectivity: The position of European cities in Internet backbone and air trafﬁc ﬂows. Telecommunications Policy 34: 417–429
Derudder B (2006) On Conceptual Confusion in Empirical Analyses of a Transnational Urban Network. Urban Studies 43: 2027-2046.
Derudder B (2008) Mapping Global Urban Networks: A Decade of Empirical World Cities Research. Geography Compass 2: 559-574.
Derudder B, Devriendt L and Witlox F (2010a) A spatial analysis of multiple airport cities. Journal of Transport Geography 18: 345-353.
Derudder B, Taylor P, Ni P, De Vos A, Hoyler M, Hanssens H, Bassens D, Huang, J, Witlox F, Shen W and Xiaolan Y (2010b) Pathways of Change: Shifting Connectivities in the World City Network, 2000-2008. Urban Studies 47: 1861-1877.
Florida R (2005) The World is Spiky. The Atlantic 296: 48-51.
Friedmann J (1986) The world city hypothesis. Development and Change 17: 69–83.
Growe A and Blotevogel H H (2011) Knowledge Hubs in the German Urban System: Identifying Hubs by Combining Network and Territorial Perspectives. Raumforschung und Raumordnung 69: 175-185.
Hall P and Pain K (eds.) (2006) The polycentric metropolis. Learning from mega-city regions in Europe. London: Earthscan.
Hall P, Pain K and Green N (2006) The Informational Geography of Europolis: Mapping the Flow of Information. In: Hall P and Pain K (eds.) The polycentric metropolis. Learning from mega-city regions in Europe. London: Earthscan, 70-87.
Hennemann S and Derudder B (2012) An Alternative Approach to the Calculation and Analysis of Connectivity in the World City Network. GaWC Research Bulletin 401. URL (accessed 10 July 2012): http://www.lboro.ac.uk/gawc/rb/rb401.html
Hermelin B (2010) Nordic Cities. In: Taylor P, Ni P, Derudder B, Hoyler M, Huang J and Witlox F, (eds) Global Urban Analysis. A survey of Cities in Globalization. London: Earthscan, 300-305.
Liu X and Derudder B (2012) Two-mode Networks and the Interlocking World City Network Model: A Reply to Neal. Geographical Analysis, 44 (2): 171-173.
Lüthi S, Thierstein A and Bentlage M (2011) Interlocking Firm Networks in the German Knowledge Economy. On Local Networks and Global Connectivity. Raumforschung und Raumordnung 69: 161-175.
Lüthi S, Thierstein A and Goebel V (2010) Intra-firm and extra-firm linkages in the knowledge economy: the case of the emerging mega-city region of Munich. Global Networks 10: 114-137.
Ministers responsible for Spatial Planning and Territorial Development of the European Union (2011) Territorial Agenda of the European Union 2020: Towards an Inclusive, Smart and Sustainable Europe of Diverse Regions. http://www.eu-territorial-agenda.eu/Reference%20Documents/Final%20TA2020.pdf (assessed: 10 July, 2012).
Münter A (2011) Germany’s Polycentric Metropolitan Regions in the World City Network. Raumforschung und Raumordnung 69: 187-200.
Neal, Z (2012) Structural Determinism in the Interlocking World City Network. Geographical Analysis 44 (2): 162-170.
OECD (2006) OECD Territorial Reviews: Competitive Cities in the Global Economy. 446 p.
Sassen S (1991) The Global City. Princeton, NJ: Princeton University Press.
Sassen S (2010) Global inter-city networks and commodity chains: any intersections. Global Networks 10:150-163.
Schmitt P and Dubois A (2008) Exploring the Baltic Sea Region – On territorial capital and spatial integration. Nordregio Report 2008:3, Stockholm, 138 p.
Scott A (2001) Globalization and the rise of city-regions. European Planning Studies 9: 813–826.
Smith D A and Timberlake M (2002) Hierarchies of dominance among world cities: a network approach. In: Sassen S (ed.) Global networks, linked cities. London: Routledge, 117-141.
Taylor P (2004) World City Network: A Global Urban Analysis. London: Routledge.
Taylor P, Ni P, Derudder B, Hoyler M, Huang J and Witlox F, (eds) (2010) Global Urban Analysis. A survey of Cities in Globalization. London: Earthscan.
Taylor P, Derudder B, Hoyler M and Ni P (2011) New Regional Geographies of the World as Practised by Leading Advanced Producer Service Firms in 2010. GaWC Research Bulletin 392. URL (accessed 10 July 2012): http://www.lboro.ac.uk/gawc/rb/rb392.html.