This Research Bulletin has been published in Tijdschrift voor Economische en Sociale Geografie, 108 (6), (2017), 753-767.
Please refer to the published version when quoting the paper.
World City Network (WCN) core research focuses on advanced producer services or APS (Sassen, 1993; Taylor, 2001) serving large global firms. The main APS considered are finance, insurance, accountancy, law, advertisement, management consultancy (Beaverstock et al., 1999; Sassen, 2000; Taylor, 2004). These are the main activities that characterize the domination of the informational post-industrial economy, distinguished by the transnational circulation of high value information (Sassen, 2006). Other activities, that belong to the industrial sector but are knowledge intensive (Lüthi et al., 2010), such as R & D, engineering, etc. are also commonly taken into account (Pain and Hall, 2006; Growe and Blotevogel, 2011).
This paper ? in the context of the WCN methodology, as proposed by Taylor (2001) and the GaWC (Globalization and World Cities Research Network) ? focuses on the logistics sector. Large industrial groups have reorganized around the concept of value chain ? which has become more and more complex ? and, quite recently, they began to invest in logistics, as it represents a part of the value chain. The development of logistics involves third party service providers, the most important of which are newly established global firms (Dicken, 2007). These firms have some influence on their clients ? as they are essential intermediaries ? and also have a dominant position with respect to transport companies, as they fill their ships and planes (Bologna, 2012). Ultimately, the sector is relevant in the global economy as it manages the physical flow of commodities (Hesse and Rodrigue, 2006; Jacobs et al., 2010). Now, a specificity of the logistics sector is that, besides the material/operational activities, the sector is also characterized by high value informational activities and these two kind of activities are characterized by a gap in terms of value and geographical organization (Hall et al., 2011).
This paper aims to focus on the geography of these informational logistics activities, that represent the command and top management functions of the sector and that play, in our opinion, a similar role to that of APS, as they are knowledge-based activities with a fundamental role in the functioning of the global economy (Castells, 1996; Lüthi et al., 2010). Now, while transport and logistics sectors have already received some attention in the WCN related research (Verhetsel and Sel, 2009; O'Connor, 2010; Wang and Cheng, 2010; Jacobs et al., 2010 and 2011; O'Connor et al., 2012) the distinction between operational/low level functions and informational/managerial functions has been somewhat problematic because it is difficult to clearly separate these functions since numerous firms carry out both types of activities. In fact, the analysis of firms’ websites (Taylor et al., 2002) ? which is the data search method most commonly used in WCN type research ? is poorly suited to the analysis of the logistics industry. First, the websites of major Logistics Service Providers (hereafter LSP), unlike those of traditionally analyzed APS, are often poorly informed. If companies often give the location of their warehouses to communicate on the operational aspects of their activities, they do not provide information regarding the functional side of their organization. Besides ? from information provided by websites ? it is very difficult to circumscribe advanced logistics services (hereafter ALS) from low level functions, given that often they are both present in the same business unit.
This paper aims to enrich the research on logistics firms localization with a first attempt to draw an interlocking network dedicated to logistics. This is done through an innovative methodology, based on the use of the social network LinkedIn for information retrieval. This tool allows to select the managerial functions and supply chain management and related services within the logistics companies. As a consequence, the "low-level" part of logistics can be discarded from the survey. LinkedIn gives access to information of great detail, including information on the business divisions location. Therefore, the proposed methodology is particularly suitable for investigation of secondary offices within a national or regional scale analysis. The case study presented here focuses on the network of Italian logistics cities based on the intra-firm networks of main LSP present in Italy. The networks corresponding to the main business divisions are defined as well. This type of analysis, already proposed by Lüthi et al. (2013) with reference to APS and high-tech firms in Germany, allows to specify the determinants of the location and the characteristics of a national urban system with a greater level of detail.
The paper is organized as follows. Section two discusses the analogies between part of logistics services and APS and review the WCN literature related to the transport and logistics industry. Section three provides a brief framework of the Italian organization, as regards to demography, economy, basic logistics and its inclusion in the WCN. Section four is dedicated to the LinkedIn methodology and the construction of the interlocking network. In section five the results at the firm level are discussed while, in section six, the analysis is split at the division level. Section seven draws the conclusions.
Advanced Producer Services Functions Within Logistics Firms
Logistics includes many functions that are very heterogeneous, and in fact, there is no specific "logistics sector" in statistical classifications (Saglietto, 2013). Looking into the details of statistics sector classifications, a good part of the activities of logistics companies is included in the NACE Rev. 2 code "H Transportation and storage". This is mostly "low level" activities such as transport and warehousing that have little to do with APS. But it also includes transport support activities such as freight forwarding and freight brokerage, that clearly require more specialized skills. Besides, the activities of logistics companies also extend to other areas of statistical classifications. In particular, in the Italian classification ATECO 6 digits, the 702201 code stands for "consultancy for firms’ logistics management". These activities are included in the NACE Rev. 2 equivalent sector "M-professional, scientific and technical activities", which also includes several of the most commonly considered APS, such as activities of management and consulting, accounting and legal services, engineering, advertising. Saglietto (2013) further allocates some logistics activities to additional statistical categories, that are anyway related to the functions of consulting, management and engineering.
Within the logistics sector there are firms, called second party logistics or 2PL, that only offer operational services such as transport and warehousing and others that offer only supply chain management services. The latter are typically defined as pure 4PL (fourth party logistics), although it should be stressed that, in the scientific literature and sector magazines, there is no shared definition of the 4PL concept (Langley et al., 2005). Anderson (now Accenture Consulting) provides the first definition of 4PL defining them as: "an integrator that assembles its own resources, capabilities and technology and those of other service providers to design and manage complex supply chains" (Saglietto, 2013). 4PL manage the entire supply chain for their customers "by mobilizing their business network and by ensuring their coherence through total control of the information flows" (Paché, 2007). However, if pure 4PL present, according to some authors (Razzaque and Sheng, 1998; Pigeon and Sirois, 2010), the specific characteristic of not having their own physical assets, there are other logistics operators that provide supply chain management services, while possessing assets ? to varying degree ? and providing transportation and warehousing related services (Paché, 2007). These operators ? and in particular large global 3PL (third party logistics) ? are therefore also relevant, for part of their business, in order to consider the APS function of logistics.
WCN researchers developed the concept of "multiple globalizations" (Taylor et al. 2004; Krätke, 2014), that is the coexistence of different networks of cities, representing the organization at a global level of specific supply chains. Investigation of these “specialized” networks is sometimes based on firms that do not belong to the service sector (Lüthi et al., 2013; Krätke, 2014). In the transport sector, the network of port cities designed by Verhetsel and Sel (2009) on the basis of the location of terminal operators and shipping companies can be mentioned. Some authors have instead tried to isolate these APS specialized in a particular field. Still with regard to maritime transport, an example is given by the contribution of Jacobs et al. (2010; 2011) that have selected insurance and law firms, consultancy and surveying, organizations specialized in the maritime sector.
Unlike the port APS considered by Jacobs et al. (2010; 2011) ? that primarily serve commodity traders (Jacobs, 2014) and other companies in the transport sector ? ALS cannot be considered as services dedicated to a specific industry. In fact, they serve every type of manufacturing industry but also the service sector, for example by organizing the logistics of supply of hotel chains or catering (Wang and Cheng, 2010). Nevertheless, it is likely that, because of specific characteristics ? as in the case of specialized APS and control centers of manufacturing industry ? the localization behaviour of ALS differ from that of other sectors. Indeed, it should be noted that if the analysis of main APS is split in order to consider them individually (financial services, legal services, etc.), differentiated localization patterns can be recognized here too (Taylor, 2005; 2012).
While the logistics functions of transport and handling remain constrained by the infrastructure network, it seems that advanced functions are influenced by other location factors, which are present only at some large infrastructure nodes. This phenomenon seems to affect not only the distinction between types of companies but also the network of offices of individual firms (Slack, 1996). With reference to command and knowledge intensive functions, the literature reveals how they seek proximity with APS and/or industry specific APS and customers. For example, Jacobs et al. (2011) highlights a network of World Port Cities dominated by both World Cities hosting a port or a port industry and port cities not highly ranked in the WCN. But even for this second kind of city it seems that the key factor is not so much the presence of a port as mere traffic hub but rather the local presence of a cluster that is specialized in the field (port APS, port industry headquarters, research centers). Regarding the logistics sector as a whole, Wang and Cheng (2010) show that the logistics industry in the World City of Hong Kong is specializing in supply chain management activities, while low level operational services tend to migrate towards other infrastructure nodes, that are less expensive and congested. Focusing on advanced logistics services and transport management, O'Connor et al. (2012) show how, in the US, their location presents a correlation both with employment in transportation and warehousing and the degree of specialization in APS. However, an interesting outcome of their survey is that location strategies of transportation advanced logistics services are different from those of logistics management. Transport services are more related to total employment in the transport and warehousing sector and to the degree of specialization in APS while logistics management services are partly released from infrastructure and appear less connected with the traditional US business core.
Firms’ search for specific location advantages also extends to the business unit level (Lüthi et al., 2013). In fact, big companies configure their value chain by moving towards a greater specialization in each of the activities that compose the chain and are led to choose the most appropriate economic environments ? in terms of infrastructure endowment, knowledge resources, proximity to clients ? for every activity (Dicken, 2007; Coe et al., 2010). This trend towards specialization seems quite suitable for LSP, considering the different activities that make up the sector: LSP are connected to the transport industry, but also to production and trade, through specialization in different supply chains for complex activities ranging from the supply to factories to the final distribution of goods in cities. Moreover, like other large companies, LSP have their internal management features where commercial functions (marketing, customer services, etc.) are distinguished from support activities (e.g. IT, HR), etc. These divisions have their own localization requirements and can be located at the headquarter or elsewhere (for example: IT at a technopolis). As mentioned above, the different location of activities ? also within the firm itself ? is widely acknowledged in logistics studies (Jacobs et al, 2011; Wang and Cheng, 2010). Yet, to our knowledge, there are no in depth studies which focus on the location of every specific division.
The Italian Context
When it comes to a focus on a regional (or national) scale, Italy is an interesting case study because, on the one hand, it has a polycentric economy, while on the other, its placement within the WCN is clearly biased in favour of Milan. This structure is helpful to understand the logic of ALS locational choices. In particular it allows to highlight if ALS follow the manufacturing industry and infrastructure nodes or whether, like traditional APS, they search these knowledge rich environments, concentrating in world cities. Also, in Italy, "low level" logistics assume considerable weight in terms of number of companies and employees, and therefore a methodology that properly isolates the functions of advanced logistics is essential for a WCN type analysis.
Italy has a polycentric structure both in terms of population and production, though heavily unbalanced in favour of northern Italy (46% of population and 56% of GDP in 2011, www.istat.it). In the North, the economic role of Milan is definitely predominant but still integrated in a macro area that presents the important automotive center of Turin, the country's main port (Genoa) and a large area of widespread industrialization (the so-called "Third Italy", i.e. birthplace of Italian industrial districts). In the Center (20% pop; 22% GDP), the capital plays an important economic role, supported by its institutional functions, while Southern Italy and Islands are characterized by a clear development lag (35% pop., 23% of GDP). The map of Italy changes, at least partially, if it is viewed from the WCN point of view (GaWC, 2012). In that case the dominance of Milan becomes much more evident, highlighting the coexistence in the country of a globalized and networked economy ? concentrated in Milan (Alpha in GaWC ranking) and Rome (Beta +) ? and of a poorly globalized economy, mainly based on small and medium firms, widespread on the whole country. Milan and Rome maintain strong ties as regards to advanced services (Taylor, 2012). The framework of the Italian positioning in the WCN 2012 is completed by Turin (Gamma -) and the other most populous Italian cities (Bologna, Genoa, Florence, Naples) ? except Palermo ? that get the sufficiency.
WCN research and this paper focus on networked cities and on this part of the economy that is directly or indirectly connected to the global network. However, to frame the Italian context, it is important to emphasize that, next to the Milanese and Roman realities, there is an important economic reality, sometimes also competitive, which is based on – strong – local or national networks but that is poorly open to the outside. According to Taylor (2007), the economy of the City works according to a combination between the urban internal dynamic of clusters, which deals with localized tacit knowledge and special “know-how”, and the urban external network dynamic which provides knowledge based on cosmopolitism. Italy, and in particular the areas of the North East and Centre, is famous for the existence and competitiveness of industrial districts (theorized in the 1970s by Becattini, 1979), i.e. production systems based on small and medium firms in which tacit knowledge and local contact networks are essential factors for competitiveness. Even outside the district areas, and with reference to larger companies, tacit knowledge and agglomeration economies are important factors of competitiveness. However, this initial endowment, strongly anchored in the productive history of the country, has been seldom enriched by the development of external connections (between Italian cities and the world).
With regards to Italian logistics, though the sector is now dominated by large global players, it stands for a strong presence of small and medium-sized firms (Confetra, 2007). Moreover, the distinction between high-level and low-level logistics functions in the Italian logistics sector is particularly relevant as it seems that, unlike other countries, competitiveness is pursued more through cost reduction (e.g. subcontracting) than through increases of productivity (infrastructure and supply chain management). In fact, an exploratory analysis at the European scale (based on LinkedIn data) on the Top ten Europeans 3PL (Scarzo, 2014), shows that Milan is the only Italian city to be ranked within the major European logistics centers. Milan is the only city ? thanks to the presence of important offices of the leading global logistics operators ? able to present strong connections with the centers of command and control of European logistics. In fact, the 50 leading companies whose head is in the province of Milan represent 40 per cent of the total turnover of the 500 largest LSP in Italy. Moreover, the non-Italian global logistics players are all based in the Milan area, except TNT based in Turin (Confetra, 2010). Apart from Milan, the geography of the main companies in the sector reflects that of the manufacturing sector and of major transportation hubs (Confetra, 2009). It highlights a concentration in the Centre-North and particularly in the area surrounding Turin, the axis joining Milan to Venice (Northeast widespread industrialization), the Emilian axis (Bologna region) and the Alpine passes. The areas of the northern Tyrrhenian ports (Genoa, La Spezia, Leghorn) and the Rome-Naples axis also stand out. The rest of the South is poorly represented and this can be explained both by its reduced contribution to the manufacturing industry and the decay of its road network.
In the following paragraphs this brief description of Italian logistics is enriched through a study devoted to the sole ALS functions and to their intra-firm and inter-city connection patterns. The objective is to clarify the role of Milan as the orchestrator of the Italian logistics and to understand how and how much other Italian logistics cities are connected between themselves and with Milan.
Construction of an Italian Logistics Interlocking Network Based on LinkedIn Data
In this paper, instead of companies’ website, we build the firms’ offices profiles through the information collected on LinkedIn concerning firms’ offices employees. This method allows to circumscribe the analysis to management and knowledge intensive functions and to provide more detailed information that is significant for an analysis at the national/regional scale.
LinkedIn is a professional social network of English expression which includes more than 400 million members in over 200 countries (http://press.linkedin.com/about; accessed January 2016). The advantage of LinkedIn is that it contains qualitative information on companies through the profiles of their staff registered on the site. The disadvantage is to work on a voluntary basis, like other social networks, so that companies are unevenly represented. In any case, the fact that participation in this social network is spontaneous is an interesting discriminatory criterion. Indeed, it can be said that LinkedIn members adhere to the “network society” by sharing the idea of an open and flexible job market and recognize themselves in a transnational and English-spoken economy, based on open access information on the web. Besides, if the role that corporate strategy can have in the spreading of the network among the company staff is considered (which means to encourage or not subscription to the network), adhering to the network society can probably be interpreted at the firm level, and not only individually.
The investigation began by searching the LinkedIn profiles of people who are employed by the top 100 companies by turnover in the Italian logistics sector. This ranking of companies is provided by Confetra (2010) and include 3PL logistics, express delivery companies and intermodal operators. The 27 companies which are represented on LinkedIn by less than ten profiles have not been considered and this implied working with 73 companies. From July to September 2013, 3,581 profiles have been extracted.
The methodology that is applied in order to build the network corresponds to that of Taylor (2001), but is nevertheless adapted to the LinkedIn database. The change concerns the assessment of the value of companies in cities. In its regular application, the Interlocking Network Model ? on the basis of information provided by firms' websites ? measures the service value (v) by interpreting the symbolic function of the office inside the corporate network (global headquarters, regional headquarters, etc.). This is not the case for LinkedIn which deals with a population of individuals represented by their profiles, and which enables to represent the value of offices through the characteristics of local employees.
For the selection of profiles, we use a range of keywords, that brought to discard low level profiles while allowing to consider different titles for similar high level functions. Then, a value from one to four has been assigned to each selected profile as follow, depending on the position held in the company. Top management (e.g. a director): 4 points; Intensive knowledge management (e.g. project manager, business analyst, development manager, IT engineer, etc.): 3 points; Management functions without further indications: 2 points; Low management (e.g. supervisor, controller): 1 point.
Then, to obtain the service value (v) of a (j) firm in the (i) city, the sum of the value of all the profiles belonging to the same firm in the same city has been calculated. The resulting sum is called an “office”, by reference to the original vocabulary related to the Interlocking Network. This operation leads to highly variable scores from one office to another. Because the model is based on a product, it is necessary to weigh the values to avoid getting excessive results which distort the network. To do this, the distribution has been reduced in a range from 1 to 15 points. In the end, 24 Italian cities have been identified and in some case profiles from other small centers that belongs to their metropolitan area were allocated to them.
It must be stated that data collected on LinkedIn have a certain margin of error. In particular, the information related to the place of work is sometime incomplete or misleading. Also, even if new jobs and work promotions are usually underlined by users, there’s a possibility of incorrectly updated profiles. However ? as stated by Taylor et al. (2002) while acknowledging the presence of imperfection in the website data source ? the large number of firms (and profiles) surveyed reduce the influence of possible inaccuracies.
Once the service value (v) is assigned to each office, the “traditional” methodology is applied in order to calculate the nodal connections (Inter City Network Connectivity; ICNC) that measures the overall status of each city within the network (cfr: Taylor, 2001). Figure 1 graphically represents the ICNC value of the fourteen first cities of the Italian Logistics City Network: the value of 1 has been attributed to the first city, i.e. Milan, and the nodal connections of the other cities have been represented as portions of Milan’s nodal connection. The study proposes an oriented reading of the relationship between cities, through the movement of companies, represented by the arrows in the graph in Figure 1. To do this, the service value (v) of the (j) company in a couple of cities (a) and (b) have been compared. The city with the highest service value (v) is the “origin” which initiates the relationship toward the city whose service value (v) is lower. Then, the dominant values of all firms in the relation between (a) and (b) have been added in order to obtain the total value of the link from (a) to (b) and from (b) to (a). In Figure 1, the value of 1 is attributed to the highest link, i.e. Milan-Rome, and the other links are represented as a portion of it.
The Italian Logistics City Network
Fig. 1: Italian Logistics City Network
The configuration of the Italian logistics network presented in Figure 1 roughly reflects the WCN but in this map we can contemporarily distinguish the patterns of main centers of Italian industry and logistics and freight transport sector “as a whole”. As in the WCN, there is a clear predominance of Milan that is expressed both by its connectivity index and the direction of links between cities. In fact, out of the 73 surveyed, 28 firms have a representation in Milan, 26 of which have their headquarters there as well. It is clear that overall ? and especially for large multinational groups that highly contribute to the total connectivity within the network ? the other cities are functionally submitted to Milan (that, anyway seems to be herself a sub-node within the European logistics network; Scarzo, 2014). With regards to secondary cities, here Turin appears better ranked than in the WCN, and it is the only city to have equal links with Milan (this is mostly related to the presence of the TNT headquarter in Turin). With regards to links, it is also interesting to note that only Rome, Turin and Bologna have mutual links, while all the remaining cities are linked only to Milan, which clearly highlights a lack of functional policentricity (Lüthi et al, 2013). Always in confrontation with the WCN, Florence looses positions in respect to Bologna and Genoa (that have a well-known logistics vocation). At first sight the Northeast expresses its manufacturing importance and polycentric nature. As a matter of fact no center of the area overtops the others, if not slightly Bergamo that hosts an airport hub. In terms of networked logistics, Southern Italy is represented just by the city of Naples, though with a secondary role. If, compared to the rest of Italy, the logistics and transport sector as a whole was already under-represented in the South, this trend is even more evident with regards to ALS. This configuration reinforces the image of functional remoteness of this area.
In this study, it is possible to say that the total value of LinkedIn profiles in each of the analyzed cities provides information on the internal cluster, while the connectivity indicator gives information on the external dynamic of networks (Taylor, 2007). The intersection of these two types of information allows identifying the degree of openness or closure of urban economies, that is particularly relevant in order to understand the different characteristics of the smallest logistics cities. In Figure 2 both cities connectivity and sum of profiles value are represented. The Figure highlights ? in the sum of “profile value” trend ? some “depressions”, which indicate situations in which the connectivity is particularly high compared to the profiles’ value, and “peaks”, which show, on the contrary, a “deficit of connectivity”.
In Fig. 2 both cities connectivity and sum of profiles value are represented (i.e. city value before reducing the distribution to a 1-15 range). From this figure, we can highlight - in the sum of "profiles' value" trend - some "depressions" which indicate situations in which the connectivity is particularly high compared to the profiles value, and "peaks" which show, on the contrary, a "deficit of connectivity".
Fig. 2: Connectivity and dimension of local clusters
Except for Milan, which can appear under-connected due to the mass of population, the under-connected cities are clearly those in which most of the profiles are employed by “native” companies (e.g. Reggio-Emilia). It is important to underline that if these cities are poorly integrated in the networked national and global economy, they however provide an important contribution to the regional and national economy that is illustrated by the existence of advanced logistics in the firms they host. When it comes to over-connection, this can be mainly explained by the dominance of large multinational companies in these cities. Yet, the over-connectivity is not necessarily associated with a functional predominance, but simply by a large presence of small or medium sized branch offices belonging to global firms (e.g. Naples).
The Organization of Logistics Companies National Value Chains
As stressed in the previous paragraph, the Italian logistics network roughly coincides with the Italian WCN. However, this is the “big picture” and the importance of WCN ranking for logistics firms’ location ? with respect to other location factors such as infrastructure and industry clustering ? can be further specified. In fact, with regards to Milan, it must be borne in mind that it is not only a World City, but also a city of important demographic dimensions, located within a manufacturing region and that is well positioned on the infrastructure axes. Besides, with reference to other Italian cities and their relation with Milan, it must be clarified if their secondary role in intra-firms networks coincides with a mere functional subordination, or with a firms’ strategy of functional specialization, dictated by the particular endowment of each city. A network analysis at the division level can allows to better distinguish the influence of these different location factors and firms’ strategy, between centralization and dispersion of functions. For this purpose, LinkedIn is once again a useful tool as people enrich their profiles with the details of their job, in particular the workplace and the division or service where they work.
Figure 3 decomposes the main intercity links between Milan and each of the 13 best connected Italian cities (see Fig. 1), as well as between Rome, Bologna and Turin. Considering that the city “home” is the headquarter of the company, Figure 3 lists management and valued specialized activities that are carried out in the city “target” by companies that contribute most heavily to the interurban relationship for each pair of cities.
Figure 4 depicts the internal organization of six activities representing a simplified value chain of companies (Lüthi et al., 2013) and portrays six city networks that describe the geographical organization of each of the six logistics activities in the Italian space. The first activity refers to basic logistics, i.e. the activities dealing with the management of regional markets and operational logistics. The three following deal with activities polarized around clients and infrastructures. It includes companies’ divisions dedicated to a specific commodity chain (vertical market), as well as the divisions specialized in ocean freight and air freight. The last two concern the activities of national and international management and the knowledge intensive jobs. A graph is drawn using the same methodology as in Figure 1 but in this case, only the population of profiles corresponding to the selected activity is considered. It must be noted that, if some of the networks illustrated in Figure 4 seem denser than the one represented in Figure 1, it is due to the fact that relations of minor intensity are represented here.
If Figures 3 and 4 are cross-read, it is possible to identify the main cities and interurban links which are involved in the different poles and activities which compose the Italian advanced logistics sector.
Fig. 3: Distribution of business activities
Fig. 4: Internal geographical organization of the activities that compose the Italian advanced logistics sector
Regional Centers and Operational Logistics
Logistics companies divide the domestic market in several territories from which they organize the sales management areas and localize platforms and warehouses. Figure 4.a shows the cities where the regional (subnational) head offices are located, plus the branch or site managers and sales area managers. These divisions are the ones ? within our selection of functions ? that are functionally the closest to “basic logistics” (storage, transport, distribution, sales and forwarding) and the farthest from ALS. In fact, while Milan is still dominant, the secondary cities assumes an higher relative weight because firms needs to locates offices of these divisions in different places across the country and the map quite properly reflect the location of the overall logistics and transportation sector.
Maritime and Airfreight Logistics
Port cities are preferential locations for shipping companies (Verhetsel and Sel, 2009). Main port cities also have a cluster of advanced maritime services (Jacobs et al., 2011) in which logistics can be inserted. In Europe, unlike the situation in North America and Asia, shipping companies did not substantially developed inland logistics, since they fail to compete with major 3PL located in continental metropolitan areas (Bologna, 2012). In this sense, it is interesting to see whether LSP firms localize some maritime functions in Milan, or instead, if they decentralize these functions in port cities. The same question also arises for air logistics activities, which are likely to be located in major airport cities (Milan, Rome and Bergamo).
Figure 4.b shows that maritime logistics is predominantly shared by Milan and Genoa, while the other cities, including other ports, are underrepresented. Clearly the specific cluster of knowledge that distinguishes Genoa in this sector (Ghiara and Sillig, 2008) encourage firms to locate the maritime division in this city in order to benefit from specific expertise and tacit knowledge. However, the high connectivity of Milan in the same sector highlights a trade-off with concentration of function and general expertise as Jacobs et al. (2010) point out. It is also interesting to underline that, for this division, Genoa has its own centrality, represented by its numerous connections, some of which (Bergamo and Ancona) are exclusive (i.e. these cities are not linked with Milan). Looking at air freight logistics (Figure 4.c), Bergamo stands out due to the “Mediterranean cluster” of DHL Aviation, which is shared with Milan. As for maritime logistics, the strongly dominant position of Milan (as well as the relative importance of Bologna, that does not account for an important airport) with respect to the other air hubs, point out how the presence of infrastructure influence the location of managerial functions only partially.
A vertical market consists of firms sharing the same product, service or process and having the same logistical need in a specific commodity chain. Developing specialized services in the most important industrial markets within logistics firms reflects a trend towards a growing proximity with the customer. Such services are likely to be located in cities inserted in different sectors such as automotive, fashion, etc. In fact Figure 4.d highlights the position of Milan, that is both a World City and a centre of manufacture, and Turin, that stand out for the automotive industry (and is Gamma – ranked in the WCN). Nevertheless, while the Italian industry is quite well distributed, at least in the Centre-North of Italy, this division network presents a remarkable concentration in few cities. It can be suggested that, if the proximity to the client is taken into account by firms for the location of managerial function in this division, this characteristic must be coupled with the presence of APS.
Top Management and International Management
The global economy can be defined as an unprecedented increase of cross-border relations. The most obvious function attached to this economy is therefore the management and coordination of trade and international organizations. This management is exercised from World Cities that act as articulators between domestic markets and the world market (Sassen, 2005). Figure 4.e highlights the places from where foreign markets (Europe, Middle East and Africa, etc.) are managed, and the location of national heads of the various corporate divisions (i.e. sectorial directions; e.g. Head of Marketing). This division graph shows a very centralized geography, where Milan is the most prominent. Moreover Figure 3 shows that numerous non-Milanese companies (TNT, SDA, etc.) keep many top management functions (IT, marketing, human resources) in Milan. The analysis of top management and international function underlines how the control of the Italian logistics network is firmly rooted in Milan that constitute the link between the national and international scale. A few other cities present top management and international function and this suggests that they have some skills in driving supply chains. Yet, they are often associated to specific value chains. Finally, it must be stressed out that the presence of infrastructures appears to be almost irrelevant for the location of these divisions.
Figure 4.f represents the geography of various jobs related to analysis (business analyst) and development (developer manager) as well as people who are responsible for maintaining relationships with major customers, whether on a commercial basis (key account manager, business developer) or on an operative basis (project manager). This is probably the division that better reflects the function of informational flow management within the logistics sector. Multilateral interactions between Milan, Rome, Turin and Bologna ? which emerge as nerve centers of knowledge logistics ? can be observed. The graph in Figure 4.f is quite similar to the one in Figure 1 and reflects the WCN (except for Florence) and the attraction of knowledge logistics near APS. It must be noted that every macro region is represented in the map. To this respect, the case of Bologna is interesting because it is associated to a strong presence in regional management – while top management and international functions are practically absents – and this probably expresses the necessity for firms to have knowledge logistics acting at the regional scale, confirming the significance of local ties.
This paper has analyzed how logistics firms link Italian cities. The Interlocking Network model, developed by the WCN, has been enriched through the use of LinkedIn. This data source allows selecting only highly skilled functions within an office as a means to overcome the specific problem that characterizes the logistics sector, which often includes high and low level functions within a single company. Moreover, LinkedIn provided qualitative information that allows concentrating the analysis at the divisions level. It was then possible to assess the positioning and the relationship between cities with a greater level of detail, which is fundamental at regional scale.
The study of the logistics industry at national scale ? besides highlighting a large presence of small and medium-sized firms that are active in transport and low-level logistics ? has shown that there are two different organizational models in Italy. On the one hand, Italy is inserted ? mainly through the presence of foreign large global operators – in the global logistics network, albeit with a subordinate role. On the other hand, it remains an economy that acts mainly on a national scale if not even local scale. This part of the economy is important as well because it holds some importance both in the creation of wealth and in shaping the national interlocking network (Taylor and Aranya, 2006).
The frame of the logistics city network, which is based on the largest companies and the sole advanced functions, differs clearly from the map of the transport and logistics sector as a whole. In fact, while we are witnessing a multiplication of operational sites (warehouses), management functions are concentrated in few places. The outcome is a network of advanced logistics strongly focused on Milan. Yet, Milan does not appear to be an exclusive command center, disconnected from its national hinterland. On the contrary, it acts as a hub (or "gateway"; Rossi and Taylor, 2006), that links the Italian network with the international scale. Value, information and power are on one side concentrated but, on the other side, are also very selectively distributed to certain places in the national space, depending on their sectorial and geographical specificities.
Asher points that "value is created less and less where physical flow take place [that would be attached to material investments and handling of goods] and increasingly in the areas where the tasks and functions are of a cognitive type, in the implementation of the reflexive potential, and in the exploitation by the company of available information " (Ascher, 2009, p 224, authors’ translation). Accordingly, infrastructural nodes, per se, lose importance in firms’ locational choices for management functions. While the presence of infrastructures may be relevant to be at least a “ranked” city, it is not sufficient to attract the most important functions. Indeed, warehouses and platforms are not systematically joined to extensive office spaces, and they are not always inhabited by high-level executives or knowledge managers. For example, while Piacenza is an important place in the logistics real estate market, the warehouses are not coupled with tertiary offices. On the contrary, Genoa stand out for its endowment of top managers in maritime logistics but have little storage capacity. Indeed ? in accordance to other studies on advanced logistics functions (Wang and Cheng, 2010; Jacobs et al., 2011) ? the study shows a preference for knowledge rich environment, both in the form of World Cities and specialized cities. APS endowment appears then as the first factor explaining the ranking of Italian logistics cities and the degree of the performed functions. In fact, the more the management functions are sophisticated and high ranked, the more they concentrate in few places. While regional centres and operational logistics ? that are the closest to basic logistics ? are widespread within the whole network, top management and international management are found mainly in Milan.
On the other side, besides this hierarchical organization between scales of management, a trend towards specialization can be observed for divisions with specific requirements such as "vertical markets" and "ocean freight". In that case, even if Milan remains dominant, companies tend to localize certain activities in those cities that have localization economies (specialized knowledge, proximity to clients, etc.) related to that specific function or product. It must be emphasized, however, that cities such as Turin and Genoa, which stand out for the localization of specialized functions, are cities that also present some endowment of APS (i.e. presence in the WCN ranking). Now, it must be noted that there can be some kind of “chicken and eggs” conundrum regarding endowment of APS and the ability to attract specialized functions of large firms. For example, the dominant role of Genoa in Ocean Freight APS, already documented by Jacobs et al. (2010), suggests that it can be precisely its maritime vocation and the APS that gravitate around it that inserted it into the WCN. Similar considerations, to a lesser extent, can be applied also to Bologna and Turin, but for this last one considering its specialization in the automotive sector (Krätke, 2014). There are too few contributions on secondary Italian World Cities in the literature to provide additional responses to this issue, yet the role of industrial specialization should be usefully developed in order to understand the mechanisms of insertion of secondary cities in the WCN and the relation between specialized advanced services and main APS.
ASCHER, F. (2009), L’age des metapoles. L’Aube, La Tour-d’Aigue.
BEAVERSTOCK, J.V., R.G. SMITH, & P.J. TAYLOR (1999), A roster of World Cities. Cities 16, pp. 445-458.
BECATTINI, G. (1979), Dal “settore” industriale al “distretto” industriale. Alcune considerazioni sull’unità di indagine dell’economia industriale. Rivista di economia e politica industriale 1.
BOLOGNA, S. (2012), Banche e crisi. Dal petrolio al container. Roma: Derive Approdi.
CASTELLS, M. (1996), The Rise of the Network Society. Oxford: Blackwell.
COE, N.M., P. DICKEN, M. HESS & H.W.C. YEUNG (2010), Making connections: Global Production Networks and World City Networks. Global Networks 10, pp. 138–149.
CONFETRA (2007), Imprese e addetti nei servizi di logistica e di trasporto merci, quaderno n.2. Available at <http://www.confetra.it/it/centrostudi/doc_pdf/quaderni_QUADERNO%20N.%202-2007.pdf>. Accessed on 12 January 2016.
CONFETRA (2009), Analisi strutturale delle imprese del settore dei servizi logistici e di trasporto 2007, quaderno 1. Available at <http://www.confetra.it/it/centrostudi/doc_pdf/QUADERNO%20N.1-2009.pdf>. Accessed on 12 January 2016.
CONFETRA (2010), Le imprese leader del trasporto di merci in Italia ? anno 2010. Available at <http://www.confetra.it/it/primopiano/Imprese%20leader2010_fatturato.pdf>. Accessed on 12 January 2016.
DICKEN, P. (2007), Global Shift: Mapping the Changing Contours of the World Economy. 5th edn., London: Sage Publications.
GAWC (GLOBALIZATION AND WORLD CITIES RESEARCH NETWORK) (2012), The World According to GaWC 2012. Available at <http://www.lboro.ac.uk/gawc/world2012t.html>. Accessed on 12 January 2016.
GHIARA, H. & C. SILLIG (2008), Clusters and territorial multiscalarity: Genoa’s port and logistics cluster. In: Musso, E. & H. Ghiara, eds., Ports and Regional Economies, pp. 81-110. Milano: McGraw-Hill.
GROWE, A. & H. BLOTEVOGEL (2011), Knowledge hubs in the German urban system: Identifying hubs by combining networking and territorial perspectives. Raumforschung und Raumordnung 69, pp. 175-185.
HALL, P.V., W. JACOBS & H. KOSTER (2011), Port, Corridor, Gateway and Chain. Exploring the geography of advanced maritime producer services. In: Hall, P.V., R. McCalla, C. Comtois & B. Slack, eds., Integrating Seaports and Trade Corridors, pp. 81-100. Farnham- Burlington: Ashgate.
HESSE, M. & J.-P. RODRIGUE (2006), Global Production Networks and the Role of Logistics and Transportation. Growth and Change 37, pp. 499-509.
JACOBS, W., C. DUCRUET & P. DE LANGEN (2010), Integrating world cities into production networks: the case of port cities. Global Networks 10, pp. 92–113.
JACOBS, W., H. KOSTER & P. HALL (2011), The Location and Global Network Structure of Maritime Advanced Producer Services. Urban Studies 48, pp. 2749-2769.
JACOBS, W. (2014), Rotterdam and Amsterdam as trading places ? In search of the economic-geographical nexus between global commodity chains and world cities. Tijdschrift voor Economische en Sociale Geografie 105, pp. 483-491.
KRÄTKE, S. (2014), How manufacturing industries connect cities across the world: extending research on multiple globalizations. Global Networks 14, pp 121–147.
LANGLEY, C., E. VAN DORT, A. ANG & S.R. SYKES (2005), Third Party Logistics. Results and findings of the 10th annual study, pp. 1-46.
LÜTHI, S., A. THIERSTEIN & V. GOEBEL (2010), Intra-firm and extra-firm linkages in the knowledge economy: the case of the emerging mega-city region of Munich. Global Networks 10, pp. 114-137.
LÜTHI, S., A. THIERSTEIN & M. BENTLAGE (2013), The Relational Geography of the Knowledge Economy in Germany: On Functional Urban Hierarchies and Localised Value Chain Systems. Urban Studies 50, pp. 276-293.
O’CONNOR, K. (2010), Global City regions and the location of logistics activity. Journal of Transport Geography 18, pp. 354-162.
O’CONNOR, K., B. HOLLY & A. CLARKE (2012), A case for incorporating logistics services in urban and regional policy: some insights from US metropolitan areas. Regional Science Policy and Practice 4, pp. 165-177.
PACHE, G. (2007), Les 4PL capitalisent-ils les savoirs issus du pilotage des chaînes logistiques étendues ? In: Dubois, P.L. & Y. Dupuy, eds., Connaissance et Management. Paris: Economica.
PAIN, K. & P. HALL (2006), The polycentric metropolis. Learning from mega-city regions in Europe. Oxford: Earthscan.
PIGEON, S. & M. SIROIS (2010), Industrie de la logistique, Gouvernement du Québec. Available at <http://www.google.it/url?sa=t&rct=j&q=&esrc=s&source=web&cd=6&ved=0CEUQFjAF&url=http%3A%2F%2Fwww.bibliotheque.assnat.qc.ca%2FDepotNumerique_v2%2FAffichageFichier.aspx%3Fidf%3D120183&ei=l7tuVZfFL8qqywOfhoDABA&usg=AFQjCNGgXyQB6iX69lOyPG6HQVkHu8O5AQ&bvm=bv.94911696,d.bGQ>. Accessed on 12 January 2016.
RAZZAQUE, M. & C. SHENG (1998), Outsourcing of logistics functions: a literature survey. International Journal of Physical Distribution & Logistics Management 28, pp. 89-107.
ROSSI, E.C. & P.J. TAYLOR (2006), ‘Gateway cities’ in economic globalization: how banks are using Brazilian cities. Tijdschrift voor Economische en Sociale Geografie 97, pp. 515-534.
SAGLIETTO, L. (2013), Toward a classification of fourth party logistics (4PL). Universal Journal of Industrial and Business Management 1, pp 104-116.
SASSEN, S. (1993), The Global City. Princeton: Princeton University Press.
SASSEN, S. (2000), Cities in a World Economy. Thousand Oaks: Pine Forge.
SASSEN, S. (2005), The Global City: introducing a concept. Brown Journal of World Affairs 11, pp. 27-43.
SASSEN, S. (2006), A Sociology of Globalization. 1st ed., New York:W. W. Norton & Company.
SCARZO, G. (2014), La rete delle città logistiche europee: l’analisi del World City Network mediante l’utilizzo di LinkedIn (Master Thesis, Università di Genova).
SLACK, B. (1996), Services linked to transportation. Papers in Regional Science 75, pp. 253–263.
TAYLOR, P.J. (2001), Specification of the World City Network. Geographical Analysis 33, pp. 181–194.
TAYLOR, P.J. (2004), World city network: a global urban analysis. London, New York: Routledge.
TAYLOR, P.J. (2005), Leading world cities: empirical evaluations of urban nodes in multiple networks. Urban Studies 42, pp. 1593-1608.
TAYLOR, P.J. (2007), Cities within spaces of flows, Theses for a materialist understanding of the external relations of cities. In: Taylor, P.J., B. Derudder, P. Saey & F. Witlox, eds., Cities in Globalization: Practices, Policies and Theories, pp. 287-297. London: Routledge.
TAYLOR, P.J. (2012), Milano nella rete delle città mondiali. In: Perulli, P., ed., Nord. Una città-regione globale, pp. 177-194. Bologna: il Mulino.
TAYLOR, P.J., G. CATALANO & D.R.F. WALKER (2002), Measurement of the World City Network. Urban Studies 39, pp. 2367-2376.
TAYLOR, P.J., G. CATALANO & D.R.F. WALKER (2004), Multiple globalisations: regional, hierarchical and sectoral articulations of global business services through world cities. The Service Industries Journal 24, pp. 63-81.
TAYLOR, P.J. & R. ARANYA (2006), The UK's Provincial Cities: Using Global Network Connectivities to Assess Claims of Economic Revival. Town & Country Planning 75, pp. 309-314.
VERHETSEL, A. & S. SEL (2009), World maritime cities: From which cities do container shipping companies make decisions? Transport Policy 16, pp. 240-250.
WANG, J. & M.C. CHENG (2010), From a hub port city to a global supply chain management center: a case study of Hong Kong. Journal of Transport Geography 18, pp. 104-115.
* Sébastien Antoine, Département de Géographie – LOTERR, Université de Lorraine .
** Cécile Sillig, Dipartimento di Economia, Università di Genova.
*** Hilda Ghiara, Dipartimento di Economia, Università di Genova.
Note: This Research Bulletin has been published in Tijdschrift voor Economische en Sociale Geografie, 108 (6), (2017), 753-767