GaWC Research Bulletin 349

This Research Bulletin has been published in P.J. Taylor, P. Ni, B. Derudder, M. Hoyler, J. Huang and F. Witlox (eds) (2011) Global Urban Analysis: A Survey of Cities in Globalization London: Earthscan, pp. 22-39.

Please refer to the published version when quoting the paper.

(Z)

Advanced Producer Service Centres in the World Economy

P.J. Taylor *

The service sector has grown immensely in the last century and now dominates all modern economies (Bryson and Daniels, 1998). Industries in this sector can be divided into those servicing individuals (e.g. hair dressing) and households (e. g. supermarkets) and those servicing businesses (e.g. advertising). These are commonly referred to as consumer services and producer services respectively. The latter include very high value services – professional, creative and financial – that are termed advanced producer services, the subject of this chapter.

Advanced producer services are very important for understanding contemporary cities in globalization. This was first clearly articulated by Saskia Sassen (1991) in her ‘global city thesis'. She argued that there are a select number of key cities – she focused on New York, London and Tokyo – that are ‘global cities' because they are the strategic locales of contemporary globalization. They are strategic because they house the management and servicing functions that have enabled globalization of economic activities. In particular, these cities have become both the production and innovation centres of advanced producer services, and the markets for this new service provision for large corporations. For instance, in the 1980s and 1990s leading London law firms and leading New York advertising agencies ‘went global' as they extended their professional and creative expertise to encompass the needs of clients operating in transnational markets. Both are selling customized high-value knowledge products to global corporate clients: multi-jurisdictional law and multi-national advertising are both examples of very advanced producer services, which Sassen sees as being specifically concentrated in global cities.

In previous eras producer service firms were based in one city and operated primarily in that city: success meant building up a local client base. Where clients had ‘outside' service needs, these were handled through loose partnerships or associations with firms in other cities or countries. There were numerous forms of such ‘association', the most common being ‘correspondence banks': selected foreign banks recommended by your home bank for overseas business in different countries. But as global service needs grew from the 1970s, the service firms themselves became large corporations with a consequent requirement for brand integrity: servicing became increasingly an ‘in-house' practice. But services cannot be properly dispensed ‘at a distance' so that ‘in house' had to be translated into a world-wide network of offices to match the globalization of client activities. Therefore, beyond Sassen's global cities, a world city network has developed (Taylor, 2004) physically represented by the huge tower blocks that form the familiar silhouettes of major cities across the world. These appear as powerful static structures but they function as dynamic nodal structures: from these offices electronic messages come in and out, to and from other offices in other tower blocks in other cities. Advanced servicing involves complex projects that involve inputs from several locales. This creates what Castells (1996) has called spaces of flows in a new network society. The flows are primarily electronic encompassing movement of creative ideas, financial information, professional knowledge, management instruction, team planning, client input, etc. that are supplemented by video conferencing and face-to face meetings (physical movement of key players) to facilitate a project from initiation to completion.

As described in Chapter 1, it is impossible to directly measure the myriad of service flows between cities in the contemporary global economy. Thus we rely on indirect measures derived from the size and functions of offices within individual firms' office networks. This is modelled as the interlocking network model that has guided the data collection. Here we focus upon 175 advanced producer service firms in the data. These are made up of the top firms in five service sectors: 75 financial services firms (including insurance); and 25 each from accounting, advertising, law and management consultancy. The offices of these firms have been scored from zero to five across 525 cities depending on the importance of offices within cities (no office in a city scores zero). The result is an array of cities and firms in a 525 x 175 service matrix.

In this chapter we report on two different analyses of these data. First, the key measurement is the network connectivity of cities. This indicates how well a city is integrated into the world city network of services. Cities housing many firms with large office networks will score high on this measure, and conversely, a city with few firms will have a lower score. To ease comparison all network connectivities will be reported as proportions of the highest scoring city. Thus we have connectivity scales ranging from one (the highest scoring city) to zero (a city with no services recorded). Second, we apply principal components analysis to the services matrix to investigate the structure of the network – which sets of cities are more inter-connected and through which particular firms. This generates sub-nets within the overall world city network, particular regional city arenas.

Global network connectivities

In this section, we report on six connectivity analyses: for all 175 advanced producer service firms, and for each of the five service sectors. All network connectivities are denoted as ‘global network connectivities' (GNC) because they encompass all 525 cities. However, to facilitate discussion in the space allocated, only connectivities of leading cities are reported here: the top 50 cities for all 175 firms and 75 financial services firms, and the top 25 cities for accountancy, advertising, law and management consultancy firms. Table 1 shows the overall connectivities of cities featuring all service firms. The following findings can be derived from this table:

• London and New York are clearly the most integrated cities and have very similar high levels of connectivity. ‘NYLON' is indisputably ‘ Main Street, World Economy'.


• Hong Kong is ranked third, separate from both New York and London above but also from cities below. Reflecting its role as economic gateway to China, the fastest growing services market in the world, this repeats earlier findings (Taylor, 2004) and contradicts the common assumption that Tokyo is the ‘third' most important city after New York and London.


• Tokyo ranks only sixth, one below Singapore, but both are part of an impressive showing by Pacific Asian cities: 5 in the top 10, 9 in the top 25.


• Within this region, the rise of Shanghai and Beijing into the top 10 is particularly noteworthy: the new ‘workshop of the world' is rapidly developing the advanced producer services to complement its huge production capacity.


• Western European cities are well-represented: 3 in the top 10 with Paris at number 4, and 9 in the top 25.


• In contrast, for US cities New York is alone in the top 10. Chicago is the second US city and ranks 19th, with Los Angeles ranking only 39th. Washington, Atlanta and San Francisco also appear in Table 1, but all are in the bottom ten. This finding repeats previous analyses showing surprising low levels of connectivity for US cities (Taylor and Lang, 2005; Taylor and Aranya, 2008). It appears to be related to the high national demand for services (by far the largest market in the world) that has resulted in more local- and national-level servicing than in other countries.


• Beyond Pacific Asia there are numerous ‘third world' cities represented as economic gateways to their respective countries: Buenos Aires 16th, Mumbai 17th, São Paulo 21st, Mexico City 24th, Caracas 38th, Santiago 41st, Johannesburg 44th, Manila 48th, and Bogotá 49th is an impressive list clearly showing the global dimension of the world city network.


• In the other ‘emerging market' region, Eastern Europe rising from its communist past, there are four cities featured including Moscow at 12th.


• Finally, the ‘old' Commonwealth is represented by Sydney, Toronto, Auckland, and Melbourne. The high ranking of Sydney, 7th, and the inclusion of Melbourne, suggests Australia is economically benefitting from its location next to Pacific Asia.

Table 1: Overall network connectivity.

Even though we have focussed on just the top 50 cities, it is remarkable how such a wide range of cities from different parts of the world is to be found.

Table 2: Financial network connectivity

Table 2 shows the financial connectivities based upon 75 firms. This contrasts with the overall connectivities (Table 1) in the following ways:

• There is now a tri-city apex to the list with Hong Kong joining New York and London with very similar scores at the top. This confirms the great importance of Hong Kong, not always recognised, within the world city network.
  
• Tokyo now rises to 4th and Shanghai to 6th in a pattern where Pacific Asian cities do even better than in Table 1: 5 in the top 10, 9 in the top 25 plus a Chinese new entry with Guangzhou at 41st.
  
• It is Western European cities that have moved down to make way for these Asian financial centres – Paris drops to 6th, but Frankfurt rising to 16th (from 32nd in Table 1) and Dublin to 23rd (from 26th) plus Geneva and Luxembourg entering at 35th and 37th respectively are the main exceptions.
  
• Nothing much changes for US cities – although San Francisco rises from 47th to 36th, and Boston is included at 50th.
  
• Emerging market areas continue to be represented approximately as before.
  
• One noteworthy entry is Dubai at 33rd, the first Arab Middle Eastern city to feature so far.

In summary, this table includes mostly the same cities as Table 1 but with an additional emphasis on Pacific Asia.

Table 3: Law network connectivity.

Table 3 shows the top 25 cities in terms of connectivity generated by law firms. Of all the services it is these firms that are generally the most concentrated in the most important cities (Taylor, 2004). The main findings from this table are as follows:

• For this service London dominates, with New York somewhat less connected but still easily ranked second.
  
• Western Europe is well represented in this table: Paris is a clear 3rd, with Frankfurt 4th and Brussels 6th. There are another six Western European cities in the top 25 including Dusseldorf, in our lists for the first time at 19th.
  
• In contrast Pacific Asian cities are relatively less well represented: Hong Kong drops to 7th and Tokyo to 9th although Beijing and Shanghai hold up at 10th and 11th. There is only one other city from this region in the list: Singapore at a lowly 17th.
  
• On this occasion, US cities fare much better than previously with Washington at a very high 5th position, Chicago at 15th and Los Angeles and San Francisco both included in the top 25. In addition Palo Alto (Silicon Valley) enters the list.
  
• More impressive are the Eastern European cities with Moscow at 8th and Warsaw, Budapest and Prague included. This relates to international law firms' contributions to converting state-owned economies to capitalist economies in the 1990s.
  
• However other emerging markets are less well represented: no Mumbai and no Latin American cities are included.

Generally speaking law firms have a distinct ‘western bias' plus an attraction to financial centres (e. g. Frankfurt) and political centres (e.g. Washington and Brussels), both of which offer specific markets for their wares.

Table 4: Advertising network connectivity.

Table 4 shows connectivities generated by advertising firms who tend to be more global than law firms (Taylor, 2004). Key findings are as follows:

• Advertising is an archetypal US industry in both its origins and development centred on New York. This is reflected here with New York being outstandingly the world centre of this service activity.
  
• However, this concentration in New York does not diffuse to other US cities: they are conspicuous by their absence from the top 25 cities listed here.
  
• London is ranked second but only just ahead of Paris; these two cities constitute a European second tier of activity in this industry.
  
• The familiar Pacific Asian trio of Hong Kong, Tokyo and Singapore come next but they lead a wide range of cities from across the world; for instance Moscow, Shanghai, Warsaw and Sydney make up the rest of the top 10.
  
• Pacific Asia dominates overall with 8 of the 25 cities listed.
  
• Western Europe is relatively under-represented after London and Paris but it is interesting that new cities appear: Athens at 16th, Stockholm at 17th, Vienna at 24th and Istanbul at 25th.
  
• Non-European emerging markets are represented only by Buenos Aires at a respectable 12th.

In summary, advertising activity appears to be concentrated where national TV markets are centred, usually capital cities but not always so (e.g. New York, Mumbai, Toronto and Istanbul). This produces quite a widespread distribution of this connectivity.

Table 5: Accountancy network connectivity.

Table 5 shows cities with the most connectivity due to accountancy firms. The latter appear to be the ubiquitous business service sector across the world with much larger office networks than the other services (Taylor, 2004). This has made the patterning of this service connectivity less clear-cut as the following shows:

• International spread of accountancy preceded globalization especially with respect to the old British Commonwealth, hence London easily outscores New York for this service and Sydney at 4th and Toronto at 11th have unusually high ranks.
  
• New York ranks a clear second but again there are no other US cities in the top 25.
  
• Western European cities are reasonably well represented but in new ways: Milan (6th) ranks above Paris (7th) and Lisbon and Berlin appear in the top 25 for the first time.
  
• The remainder are a mix of cities from emerging markets with Buenos Aires and Kuala Lumpur notable at 9th and 10th. Tel Aviv is a surprising entry at 12th.

There is little pattern to these connectivities except a tendency towards capital cities doing well especially at the lower rankings.

Table 6: Management consultancy network connectivity.

Table 6 shows the top 25 cities ranked by management consultancy connectivities. Like advertising, this is an archetypal US service that has globalized (Taylor, 2004). This is reflected in the findings:

• New York dominates this list of city connectivities even more so than for advertising.
  
• However, in contrast to advertising, in this case New York is not alone: Chicago is ranked a very high 4th, and Atlanta, Washington and Boston also appear in the top 25.
  
• Beyond the US, similar to advertising, London and Paris show a European second tier of cities below New York. There are 9 other Western European cities listed, most as before but with Munich included this time.
  
• Pacific Asia is again well represented with 7 cities, including the usual Hong Kong, Singapore and Tokyo in the top 10.
  
• Eastern Europe is not represented; cities in other developing countries are represented just by Mumbai at 11th and Mexico City at 15th.

In summary, this service is the one most concentrated in the main globalization arenas of Northern America, Western Europe and Pacific Asia.

These six tables illustrate the complexity of the world city network: the only constant is that New York and London always hold top positions. Otherwise, different services are responding to different markets producing distinctive ranges of cities. The top 10 cities of the overall pattern from Table 1 are found scattered throughout the other tables but beyond these ‘global cities', as Sassen might see them, there is variety across the city lists.

Locational strategies of firms

The complexity we have uncovered by the connectivity analyses can be investigated by returning to the service matrix, and performing a principal components analysis on it. This technique is a way of finding parsimonious results from large data sets. The idea is to find the basic dimensions in the data through reducing a large number of variables into a much smaller number of composite variables known as ‘factors' or ‘components'. For instance, in our services matrix there are 175 variables (each firm is a variable), which can be interpreted as 175 locational strategies (of where firms choose to place offices). Every firm has its own individual strategy consisting of the list of cities in which it has offices (and its scores for those offices). Thus nearly all firms include New York and London in their locational strategies with important offices, but far fewer include Munich and Bangkok, and even fewer are to be found in Naples and Brasilia. What our principal components analysis has allowed us to do is to reduce these 175 variables to just 8 composite variables, components that define a common locational strategy for a specific set of firms. One further note: because the technique can be overwhelmed by large numbers of zeros in the data, only the top 250 cities ranked by overall network connectivity are included. Therefore we begin with a service matrix of 250 cities x 175 firms (locational strategies) and conclude with a parsimonious 250 cities x 8 locational strategies. These 8 components include 52.54 per cent of the original variation in the data; that is to say, we reduce the number of variables from 175 to 8 (a 95 per cent reduction) for a loss of only 47 per cent in data variability. This is a good deal statistically: it is these eight locational strategies that are described below (Table 7).

Table 7: Locational strategies of advanced producer services.

(a) The advertising sector's global locational strategy (16.88%)

^{Sectors: AC – accountancy; AD – advertising; FS – financial services; LS – legal services; MC – management consultancy
Results are derived from principal components analysis with varimax rotation and components selected as in Taylor et al (2002). Scores are standardized variables (i.e with a mean of zero) and loadings are the equivalent of correlations between variables (firms) and components (common strategies). Note that these results should be read as two separate tables brought together to simplify interpretation. Thus the first row indicates the highest scoring city (Buenos Aires) and the highest loading firm (BBDO) but does not indicate a particular relation between Buenos Aires and BBDO.}

(b) Global city articulated strategies

LEGAL SERVICES LOCATIONAL STRATEGY ARTICULATED THROUGH NEW YORK (6.61%)

BUSINESS SERVICES LOCATIONAL STRATEGY ARTICULATED THROUGH LONDON (3.11%)

^{Sectors: AC – accountancy; AD – advertising; FS – financial services; LS – legal services; MC – management consultancy
For details of the model, see Table 7(a)}

(c) World-regional service strategies

FINANCIAL SERVICES PACIFIC ASIAN STRATEGY (7.63%)

LEGAL SERVICES PAN-EUROPEAN STRATEGY (6.62%)

FINANCIAL SERVICES UNITED STATES STRATEGY (5.75%)

^{Sectors: AC – accountancy; AD – advertising; FS – financial services; LS – legal services; MC – management consultancy
For details of the model, see Table 7(a)}

(d) Minor regional service strategies

FINANCIAL SERVICES CANADIAN STRATEGY (3.41%)

FINANCIAL SERVICES AUSTRALIAN STRATEGY (3.09%

^{Sectors: AC – accountancy; AD – advertising; FS – financial services; LS – legal services; MC – management consultancy
For details of the model, see Table 7(a)}

1. The Advertising Sector's Global Locational Strategy. This is the largest component accounting for 16.88 per cent of the overall variation. Table 7(a) shows its composition in terms of cities and firms. The former are ranked by their component scores, which define the new composite variable. The top 30 cities are listed. These are primarily world cities of medium importance; New York is included but is not at the top, London, Paris, Hong Kong, Tokyo, Singapore, Shanghai and Beijing are conspicuous by their absence. Firms are ranked by their component loadings – how correlated they are with the component. As can be seen the top nine firms are from advertising, and there are another 8 firms from this sector in the list. Hence, the new composite variable is interpreted as an advertising sector locational strategy that is worldwide in scope. As noted previously, advertising is an archetypal US service with a global range, but with no US cities other than New York (Table 4), and this is reflected in this component. Also there are other firms beyond advertising centred on New York but with worldwide scopes that share this strategy such as American Express and McKinsey & Company.
   
2. Global City Articulated Strategies. There are two other strategies that are global in scope but are very different from the last strategy: in these cases both are completed dominated by a global city, one by New York and the other by London. These are shown in Table 7(b). The first accounts for 6.61 per cent of the overall variation with a very large score recorded for New York. Of the top 10 firms listed by their loadings, 8 are law firms. Therefore this is named the Legal Services Locational Strategy articulated through New York. The other cities are important cities with Washington and London scoring reasonably high at ranks 2 and 3, followed by Los Angeles, Paris and Chicago. We found in the connectivity analysis that law services were concentrated (Table 3); here we have a composite variable illustrating this process. The second component in Table 7(b) is dominated by London but on this occasion there are a variety of sectors represented in the high loading forms. This component is named the Business Services Locational Strategy articulated through London, which accounts for 3.11 per cent of the overall variation. Note that this component includes US cities, notably Boston, but not New York. This is a truly worldwide strategy with South Asia, Australasia, Pacific Asia, Sub-Saharan Africa, Western Europe and Latin America all represented.
   
3. World-Regional Service Strategies. Most strategies are not global in scope; there are three that largely concentrate on just one major world region. These are shown in Table 7(c). The first accounts for 7.63 per cent of variation in the data and is fairly straightforward to interpret: it is made up of Pacific Asian cities and 7 of the top 10 firms are in financial services. Therefore this is named the Financial Services Pacific Asian Strategy. Note that it is Tokyo and not Hong Kong that is the first ranked city in this strategy. The second strategy in Table 7(c) is equally easy to fathom. Here all the cities are European (two from Eastern Europe) and 8 of the 10 firms loading high are law firms. Thus we name this the Legal Services Pan-European Strategy, which accounts for 6.62 per cent of the overall variance. Note that Frankfurt, London, Paris and Brussels are all fairly balanced with high loading on this component. The final strategy in Table 7 consists of US cities but not including New York and eight of the high loading firms are financial. This is, therefore, the Financial Services United States Strategy, which accounts for 5.75 per cent of the overall variation. Note that once again there is no single dominant articulator city; this role is shared by Chicago and Atlanta with other cities not far behind.
   
4. Minor Regional Service Strategies. There are two unexpected strategies that are each largely based upon a single country. These are shown in Table 7(d) and both are dominated in the loadings by financial service firms. In the first strategy all but one of the cities listed are Canadian. In the second strategy the top four are Australian cities, two of the other six are also from Australia, and there is one New Zealand city. Thus these components are named Financial Services Canadian Strategy and Financial Services Australian Strategy. They account for 3.41 per cent and 3.09 per cent of the overall variation in the data respectively. These really look like ‘national strategies' with the cities in each strategy roughly ranked in terms of their importance in their respective countries.
   

From amongst the complexity of 175 individual firms' locational strategies we have found 8 clear composite strategies that encompass between them more than half the original variation in the data. This has proven to be a good statistical deal as we knew from the start but we have also learned something important about cities in globalization: the world city network is worldwide but it is constituted by distinctive sub-nets not all of which are themselves global. The key general finding is that the composite locational strategies exist at three different scales: we have found three global strategies, three world-regional strategies and two large country strategies. Thus even after more than a quarter of a century of economic globalization, not all major advanced producer service firms are primarily worldwide in the scope of their service provision. Many, especially in financial services, continue to operate through concentrations of offices in their region/country of origin with other offices in the remainder of the world limited in number and importance.

REFERENCES

Bryson, J. and Daniels, P. W. (eds) (1998) Service Industries in the Global Economy (two vols), Edward Elgar, Cheltenham

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

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

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

Taylor, P. J. and Aranya, R. (2008) ‘A global “urban roller coaster”? Connectivity changes in the world city network, 2000-04', Regional Studies, vol 42, pp1-16

Taylor, P. J., Catalano, G. and Walker, D. R. F. (2002) ‘Exploratory analysis of the world city network', Urban Studies, vol 39, pp2377-2394

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

NOTES

^* Peter J. Taylor, Department of Geography, Loughborough University, UK; email: p.j.taylor@lboro.ac.uk