5.2 The City-Based Approach
5.2.1 The Studies by Peter J. Taylor
Drawing mainly, at theoretical level, on the ‘‘world city hypothesis’’ of Friedmann (1986), the ‘‘global cities’’ of Sassen (1991), and the ‘‘space of flows’’ of Castells (1996), Peter J. Taylor has developed various instruments for the measurement of globalization based on ‘‘a city-centric view of the world’’ expressly opposed to the
‘‘state-centric view of the world that emanates from most macro-level social sci- ence’’ (2004, p. 27).
The first and principal of these instruments aims in particular to quantify the Global Network Connectivityof 315 cities around the planet1—that is, it measures
1 This section refers to the methodological indications and the results set out in Taylor (2004).
These same methodological indications, however, were previously published in Taylor et al.
(2002), which considered 316 cities rather than 315.
their degree of interconnection. To this end, Taylor starts by identifying a set of
‘‘global service firms’’ for which it is possible to identify the location of their commercial offices by referring to information available on their websites.
He assumes, on the basis of previous research, that a firm can be called ‘global’ if it has offices in at least fifteen different cities, of which one or more is situated in one of the ‘‘prime globalization arenas’’ (Northern America, Western Europe, and Pacific Asia). The firms are therefore selected on the basis of the availability of the information required, according to a technique which Taylor (2004, p. 65) calls
‘‘scavenging’’.
Moreover, to enable decomposition of the index according to the sectors in which firms operate, and thereby compare their degrees of connectivity, Taylor considers only those sectors for which it is possible to obtain the information required for at least ten firms. Finally, on the basis of these criteria, Taylor selects 100 firms, the complete list of which is given in Table5.1, with regard to six different sectors of activity: accountancy, advertising, banking & finance, insur- ance, law, and management consultancy.
Overall, the firms selected have offices in thousands of cities. However, for theoretical reasons, and in order to handle the data, Taylor considers only 315 cities, as said. This selection is made—also on the basis of previous empirical studies—by referring mainly to two criteria: the first is territorial representative- ness; the second is economic importance.
Once the presence or otherwise of each of the 100 firms considered in all the 315 cities has been determined, the next step is to attribute a weight to this presence. For this purpose, Taylor constructs a scale that varies from 0 to 5, where 0 denotes that a city has no office of the firm considered, and 5 is assigned when the city hosts the firm’s headquarters. The intermediate scores are attributed according to the following criteria: presence of a regional headquarters, 4 points;
offices of large size, 3 points; offices of conventional size, 2 points; offices of small size or connected with other offices, 1 point.
On completion of this step, however, the value of a city’s Global Network Connectivityvalue is not simply given by the sum of the scores attributed to the offices of the firms present in that city.2Measuring the intensity of the networks, in fact, requires investigation of the connections among the 315 cities considered.
The procedure followed to identify and quantify the connections between two cities A and B consists in multiplying the score relative to the presence of a certain firm in city A by the score relative to the presence of the same firm in city B. Thus obtained is the score relative to the intensity of a single connection which—as far as one deduces from Taylor’s treatment—may vary from 0 (when the firm is not present in one of the two cities, so that there is no connection) to 20 (when the firm’s global headquarters are in one of the two cities, and the regional head- quarters are in the other). The score for a city’s Global Network Connectivity is then obtained by summing the intensities of the links of all the firms present in that
2 This sum instead gives what Taylor (2004, p. 68) calls the ‘‘total service value’’ of a city.
5.2 The City-Based Approach 123
Table5.1GlobalservicefirmsconsideredintheconstructionoftheGlobalNetworkConnectivityofcities.Referenceyear:2000(Taylor2004, pp.215–217) AccountancyBanking/financeLaw 1Ernst&Young34WestLB(WestdeutscheLandesbankGironzentrale)68Latham&Watkins 2ArthurAndersen35DresdnerBank69MorganLewis 3MacIntyreStraterInternational(MSI)36Commerzbank70Baker&McKenzie 4IGAF37DeutscheBank71CliffordChance 5AGNNetwork38Chase72JonesDay 6BDO39BNPParibas73FreshfieldsBruickhausDeringer 7GrantThorntonInternational40ABN-AMRO74Allen&Overy 8HorwathInternational41RabobankInternational75Dorsey&Whitney 9KPMG42UBS76Linklaters—Alliance 10SummitInternational+BakerTilly43ING77White&Case 11RSMi44Barclays78CameronMcKenna 12MooresRowlandInternational45FujiBank79Morrison&Foerster 13HLBInternational46BayerischeHypoVereinsbank80Lovells 14MooresStephensInternationalNetwork47BayerischeLandesbankGirozentral81Skadden,Arps,Slate,Measher,&Flom 15NexiaInternational48SakuraBank82Sidley&Austin 16PKFInternationalAssociation49SumitomoBank83CoudertBrothers 17FiducialInternational50SanwaManagementconsultancy 18PricewaterhouseCoopers51J.P.Morgan84TowersPerrin Advertising52BTM(BankofTokyo-Mitsubishi)85LogicaConsulting 19Impiric53DKB(Dai-IchiKangyoBank)86WatsonWyatt 20TMP54HSBC87SemaGroup 21Hakuhodo55Citibank88CSC 22DraftWorldwide56CreditSuisse/FirstBoston89HewittAssociates 23DensuYoungandRubicam+YoungandRubicamInsurance90IBMWorldwide 24D’Arcy57AllianzGroup91MercerManagementConsulting 25FCB58SkandiaGroup92BostonConsultingGroup 26SaatchiandSaatchi59ChubbGroup93DeloitteToucheTohmatsu 27Ogilvy60Prudential94BoozAllen&Hamilton 28BBDONetwork61RelianceGroupHoldings95A.T.Kearney 29McCann-EricksonWorldGroup62Winterthur96McKinsey 30J.WalterThompson63Fortis97Bain&Company 31EuroRSC664CGNU98Compass 32CMG(CarlsonMarketingGroup)65LibertyMutual99AndersenConsulting 33AsatsuDK66Royal&SunAlliance100GeminiConsulting/CapGemini(Ernst&Young) 67Lloyd’s
city with all the other 314 cities considered by the study.3 The first column of Table5.2 contains the classification of the first 25 cities ordered according to Global Network Connectivity.
In developing his rich and composite analysis of the relations and intercon- nections among the world’s main cities, Taylor (2004, pp. 96–99) has then isolated the information in his database relative to firms operating in the banking & finance sector, the purpose being to create—using the same aggregation procedure as before—a measure of Bank Network Connectivity able to identify the world’s main financial centers. The classification of the first 25 cities according to this index is reported in the second column of Table5.2.
By means of an analogous technique,4but this time usingThe UN Yearbook of International Organizationsas his database, Taylor has also developed an index of NGO Network Connectivity determined for fully 600 cities (Taylor 2004, pp. 95–96). The first 25 cities according to this index are shown in the third column of Table5.2.
Finally, to complete his analysis, Taylor (2004, p. 94) also reports an index devised, with the same technique as already described, by Kratke (2002). The purpose of this instrument is to measure Media Network Connectivity through analysis of the presence of thirty-three ‘‘leading global media companies’’ in 196 cities. Also the first 25 cities according to this index are shown by Table5.2, in the fourth column.
By way of brief comment on the results—relative to the first 25 positions—of the four indices reported in Table5.2, first to be noted is a substantial degree of overlap among the classifications ofGlobal Network Connectivity,Bank Network Connectivity(though the overlap between these two is unsurprising, given that the latter is a partial version of the former), andMedia Network Connectivity. In fact, there are fully 14 cities which appear among the first 25 positions in all three classifications—classifications, moreover, which always rank the cities of London and New York in the first two places. Predominant among these 14 cities—which occupy the first 16 positions in regard toGlobal Network Connectivity—are ones located in the most advanced countries (London, New York, Paris, Tokyo, Chicago, Milan, Los Angeles, Madrid, Amsterdam, Sydney, and Brussels), added to which are the two Asian ‘tigers’, Hong Kong and Singapore, and the city of São Paulo.
However, the picture changes radically when one considers NGO Network Connectivity, which denotes the existence of ‘‘a quite different world city network’’
3 Taylor (2004, p. 69) reports that the score obtained by each city can be expressed in absolute form or, more conveniently, as a proportion of the overall value of all the connections identified (4,078,256), or again as a proportion of the largest individual connectivity value (in this case, the city at the top of the classification, London, assumes value 1). It is evident that, whatever solution is adopted, this does not alter either the relative order or the proportional relations among the cities in terms of Global Network Connectivity.
4 In this case, however, the scores relative to the presence of each NGO in the various cities have been attributed using a scale from 0 to 4 rather than 0 to 5.
5.2 The City-Based Approach 125
(Taylor2004, p. 100). Ranking among the first 25 cities in this classification, in fact, are only 4 of the 14 listed above (Brussels, London, New York, and Tokyo), while there are fully 15 cities that appear among the first 25 only in this classification. The majority of them are located in the developing countries (Nairobi, New Delhi, Manila, Harare, Accra, Cairo, Dhaka, Dakar, Santiago, Abidjan, and Dar es Salaam, added to which are Washington, Geneva, Moscow, and Rome). To be noted is the significant presence in this classification of African cities which are entirely absent from the first 25 places of the three other classifications.