上海交通大学：《物流地理学 Geography of Logistics》课程教学资源（教学资料）The transport geography of logistics and freight distribution

JOURNAL OF TRANSPORT GEOGRAPHY ELSEVIER Journal of Transport Geography 12(2004)171-184 www.elsevier.com/locate/jtrangeo The transport geography of logistics and freight distribution Markus Hesse a.,Jean-Paul Rodrigueb Department of Earth Sciences,Urban Studies,Free University of Berlin,Malteserstr.74-100 D-12249 Berlin,Germany Department of Economics and Geography,Hofstra University,Hempstead,NY 11549,USA Abstract Goods movement and freight distribution are widely underrepresented in regional science and geographical research.This is surprising since a large body of traditional spatial theory has been developed with respect to transportation costs or to trade areas: those aspects that were originally closely connected with the exchange of goods.Growing attention is being paid in geography to related subjects,such as the emergence of global production networks,to structural changes in retail or to the commodification of modern consumption.To a certain extent,these processes depend upon the efficient transfer of information,finance and physical goods.Yet,with a few exceptions,the freight sector appears to be neglected in contemporary research.This paper provides an overview of the emerging transport geography of logistics and freight distribution.It challenges the traditional perspective where transportation is considered as a derived demand with the idea that logistical requirements underline transportation as a component of an integrated demand.The paper provides an analysis of the evolution of logistics as it pertains to the core dimensions of transport geography (flows,nodes/locations and networks).The concept of logistical friction is also introduced to illustrate the inclusion of the multidimensional notion of impedance in integrated freight transport demand. 2003 Elsevier Ltd.All rights reserved. Keywords:Logistics:Geography:Freight Transport:Physical distribution:Globalization 1.Logistics and freight transport:from derived to the science of physical distribution.Although it repre- integrated demand sents an entire system of space/time interdependencies, we believe that physical distribution has been neglected 1.1.Introduction in current geographical,urban or regional studies. Up to recently,geography did not pay much atten- The growing flows of freight have been a fundamental tion to logistics and freight transportation,as the focus component of contemporary changes in economic sys- was mainly on passengers and individual mobility issues. tems at the global,regional and local scales.The con- Textbooks on urban or general transport geography, sideration of these changes must be made within a like those edited by Hanson (1995),Taaffe et al.(1996) perspective where they are not merely quantitative,but or Hoyle and Knowles(1998),now raise more freight structural and operational.Structural changes mainly related questions than they did in earlier editions,par- involve manufacturing systems with their geography of ticularly with regard to trade and ports.The latter is production,while operational changes mainly concern probably the only logistics subject that received major freight transportation with its geography of distribution. reference from academic geography.Other core spatial As such,the fundamental question does not necessarily implications of distribution and logistics have been di- reside in the nature,origins and destinations of freight rectly addressed in geography by few authors who movements,but how this freight is moving.New modes developed an insight into wholesale activities and their of production are concomitant with new modes of dis- geographical distribution(Glasmeier,1992;McKinnon, tribution,which brings forward the realm of logistics; 1983,1988,1998:Riemers,1998:Vance,1970).Fol- lowing the nature of retailing as an originally distribu- tive activity,geographic research on retail and 'Corresponding author.Tel.:+49-30-838-70209;fax:+49-30-838- consumption is of interest in the logistics context too. 70749. However,retail geography does not pay much attention E-mail address:mhesse@zedat.fu-berlin.de (M.Hesse). to distribution changes (Marsden and Wrigley,1996), 0966-6923/S-see front matter 2003 Elsevier Ltd.All rights reserved. doi:10.1016j-jtrangeo.2003.12.004

The transport geography of logistics and freight distribution Markus Hesse a,*, Jean-Paul Rodrigue b a Department of Earth Sciences, Urban Studies, Free University of Berlin, Malteserstr. 74-100 D-12249 Berlin, Germany b Department of Economics and Geography, Hofstra University, Hempstead, NY 11549, USA Abstract Goods movement and freight distribution are widely underrepresented in regional science and geographical research. This is surprising since a large body of traditional spatial theory has been developed with respect to transportation costs or to trade areas: those aspects that were originally closely connected with the exchange of goods. Growing attention is being paid in geography to related subjects, such as the emergence of global production networks, to structural changes in retail or to the commodification of modern consumption. To a certain extent, these processes depend upon the efficient transfer of information, finance and physical goods. Yet, with a few exceptions, the freight sector appears to be neglected in contemporary research. This paper provides an overview of the emerging transport geography of logistics and freight distribution. It challenges the traditional perspective where transportation is considered as a derived demand with the idea that logistical requirements underline transportation as a component of an integrated demand. The paper provides an analysis of the evolution of logistics as it pertains to the core dimensions of transport geography (flows, nodes/locations and networks). The concept of logistical friction is also introduced to illustrate the inclusion of the multidimensional notion of impedance in integrated freight transport demand.
2003 Elsevier Ltd. All rights reserved. Keywords: Logistics; Geography; Freight Transport; Physical distribution; Globalization 1. Logistics and freight transport: from derived to integrated demand 1.1. Introduction The growing flows of freight have been a fundamental component of contemporary changes in economic sys￾tems at the global, regional and local scales. The con￾sideration of these changes must be made within a perspective where they are not merely quantitative, but structural and operational. Structural changes mainly involve manufacturing systems with their geography of production, while operational changes mainly concern freight transportation with its geography of distribution. As such, the fundamental question does not necessarily reside in the nature, origins and destinations of freight movements, but how this freight is moving. New modes of production are concomitant with new modes of dis￾tribution, which brings forward the realm of logistics; the science of physical distribution. Although it repre￾sents an entire system of space/time interdependencies, we believe that physical distribution has been neglected in current geographical, urban or regional studies. Up to recently, geography did not pay much atten￾tion to logistics and freight transportation, as the focus was mainly on passengers and individual mobility issues. Textbooks on urban or general transport geography, like those edited by Hanson (1995), Taaffe et al. (1996) or Hoyle and Knowles (1998), now raise more freight related questions than they did in earlier editions, par￾ticularly with regard to trade and ports. The latter is probably the only logistics subject that received major reference from academic geography. Other core spatial implications of distribution and logistics have been di￾rectly addressed in geography by few authors who developed an insight into wholesale activities and their geographical distribution (Glasmeier, 1992; McKinnon, 1983, 1988, 1998; Riemers, 1998; Vance, 1970). Fol￾lowing the nature of retailing as an originally distribu￾tive activity, geographic research on retail and consumption is of interest in the logistics context too. However, retail geography does not pay much attention to distribution changes (Marsden and Wrigley, 1996), * Corresponding author. Tel.: +49-30-838-70209; fax: +49-30-838- 70749. E-mail address: mhesse@zedat.fu-berlin.de (M. Hesse). 0966-6923/$ - see front matter
2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.jtrangeo.2003.12.004 Journal of Transport Geography 12 (2004) 171–184 www.elsevier.com/locate/jtrangeo

172 M.Hesse,J.-P.Rodrigue Journal of Transport Geography 12 (2004)171-184 although the physical movement of goods appears to be trade theory neglects the role of transport and logistics one of the costliest parts of retail activities (Christoph- (Dicken,1998,p.74),particularly the fact that transport erson,2001).One exception to these observations is in costs have a fundamental impact on the amount of trade Ralston (2003).who does address issues of inventory and goods exchange,as do traffic constraints and carrying costs and supply chain analysis. opportunities in general.We argue that this perspective Despite the upcoming notion of volatility and place- is mainly the result of a misinterpretation of role of the lessness,and contrasting the enduring neglect of trans- transport sector,freight and passengers alike,as a de- portation by regional and geographical sciences,the rived demand.Under such circumstances,transporta- material world of physical distribution and the respec- tion is perceived as a residual consequence-derived-of tive locales is considered of geographical significance. other processes or a mere "space-shrinking"function The two traditional disciplines for investigating physical (Dicken,1998;Knox and Agnew,1998).However,the distribution are business administration (economics) same processes behind the focus of the globalization and transportation sciences.Both cover,to varying de- literature,such as international trade,multinational grees,aspects of space and location.However,it is corporations and the division of labor/production,are legitimate to state that both disciplines did not pay also revealing a different perspective,which is the much attention to the spatial character of their subject. management of supply chains and their underlying In turn,economic and transport geography,did not logistics.Consequently,the role of distribution in develop too large a focus on logistics-keeping in mind globalization remains partially unanswered and a geo- the broad geographical relevance of distribution.A graphical analysis of logistics may provide substantial substantial amount of research covers different planning evidence in that respect. aspects of freight transport particularly in the urban context,either from a transport engineering and plan- 1.2.Definition of the subject ning perspective or emphasizing related urban problems (Chinitz,1960;Odgen,1992;Woudsma,2001).Logis- Logistics consider the wide set of activities dedicated tics,as a geography,remains relatively unexplored. to the transformation and circulation of goods,such as Freight distribution is now considered with more the material supply of production,the core distribution attention as productivity gains in manufacturing are and transport function,wholesale and retail and also the increasingly derived from efficiency at terminals instead provision of households with consumer goods as well as of from the efficiency of transportation modes (Rodri- the related information flows (Handfield and Nichols. gue,1999).Because transport geography is traditionally 1999).These activities composing logistics are included more engaged in long distance trade issues,freight re- into two major functions which are physical distribu- lated work received significant attention.With emerging tion;the derived transport segment,and materials global trade,production networks and distribution management;the induced transport segment.More systems,particular emphasis was given to ports and specifically: related research covering many of these issues (e.g. Hoyle,1990,1996;Hoyle and Pinder,1992;Nuhn,1999; Physical distribution (PD)is the collective term for the Slack.1998).In this context,an increasing amount of range of activities involved in the movement of goods work on intermodal freight transport and terminal is- from points of production to final points of sale and sues appeared as well (van Klink and van den Berg, consumption (McKinnon,1988,33).It must insure 1998:Drewe and Janssen.1998).Generally.interna- that the mobility requirements of supply chains are tional trade increasingly contributes to the amount and entirely met.PD comprises all the functions of move- the nature of physical distribution.Thus globalization ment and handling of goods,particularly transporta- is now discussed as having a major impact on goods tion services (trucking.freight rail,air freight,inland exchange (Janelle and Beuthe,1997;McCray,1998; waterways,marine shipping,and pipelines),trans- Pedersen,2000;Woudsma,1999). shipment and warehousing services (e.g.consignment, Still,fallacies are noted in globalization discourses storage,inventory management),trade,wholesale within economic geography,undermining the assess- and,in principle,retail.Conventionally,all these ment of the role of transportation.Within the large activities are assumed to be derived from materials body of work referring to the globalization discourse or management demands the impacts of internationalization and free trade Materials management (MM)considers all the activi- agreements,transport is not seen as a major issue or is ties related in the manufacturing of commodities in de facto taken for granted(Holmes,2000).Even classic all their stages of production along a supply chain. MM includes production and marketing activities such as production planning,demand forecasting, IThe special issues on containerisation in GeoJournal 48.1999 and purchasing and inventory management.It must in- on terminals in the Journal of Transport Geography 7,1999. sure that the requirements of supply chains are met

although the physical movement of goods appears to be one of the costliest parts of retail activities (Christoph￾erson, 2001). One exception to these observations is in Ralston (2003), who does address issues of inventory carrying costs and supply chain analysis. Despite the upcoming notion of volatility and place￾lessness, and contrasting the enduring neglect of trans￾portation by regional and geographical sciences, the material world of physical distribution and the respec￾tive locales is considered of geographical significance. The two traditional disciplines for investigating physical distribution are business administration (economics) and transportation sciences. Both cover, to varying de￾grees, aspects of space and location. However, it is legitimate to state that both disciplines did not pay much attention to the spatial character of their subject. In turn, economic and transport geography, did not develop too large a focus on logistics––keeping in mind the broad geographical relevance of distribution. A substantial amount of research covers different planning aspects of freight transport particularly in the urban context, either from a transport engineering and plan￾ning perspective or emphasizing related urban problems (Chinitz, 1960; Odgen, 1992; Woudsma, 2001). Logis￾tics, as a geography, remains relatively unexplored. Freight distribution is now considered with more attention as productivity gains in manufacturing are increasingly derived from efficiency at terminals instead of from the efficiency of transportation modes (Rodri￾gue, 1999). Because transport geography is traditionally more engaged in long distance trade issues, freight re￾lated work received significant attention. With emerging global trade, production networks and distribution systems, particular emphasis was given to ports and related research covering many of these issues (e.g. Hoyle, 1990, 1996; Hoyle and Pinder, 1992; Nuhn, 1999; Slack, 1998). In this context, an increasing amount of work on intermodal freight transport and terminal is￾sues 1 appeared as well (van Klink and van den Berg, 1998; Drewe and Janssen, 1998). Generally, interna￾tional trade increasingly contributes to the amount and the nature of physical distribution. Thus globalization is now discussed as having a major impact on goods exchange (Janelle and Beuthe, 1997; McCray, 1998; Pedersen, 2000; Woudsma, 1999). Still, fallacies are noted in globalization discourses within economic geography, undermining the assess￾ment of the role of transportation. Within the large body of work referring to the globalization discourse or the impacts of internationalization and free trade agreements, transport is not seen as a major issue or is de facto taken for granted (Holmes, 2000). Even classic trade theory neglects the role of transport and logistics (Dicken, 1998, p. 74), particularly the fact that transport costs have a fundamental impact on the amount of trade and goods exchange, as do traffic constraints and opportunities in general. We argue that this perspective is mainly the result of a misinterpretation of role of the transport sector, freight and passengers alike, as a de￾rived demand. Under such circumstances, transporta￾tion is perceived as a residual consequence––derived––of other processes or a mere ‘‘space-shrinking’’ function (Dicken, 1998; Knox and Agnew, 1998). However, the same processes behind the focus of the globalization literature, such as international trade, multinational corporations and the division of labor/production, are also revealing a different perspective, which is the management of supply chains and their underlying logistics. Consequently, the role of distribution in globalization remains partially unanswered and a geo￾graphical analysis of logistics may provide substantial evidence in that respect. 1.2. Definition of the subject Logistics consider the wide set of activities dedicated to the transformation and circulation of goods, such as the material supply of production, the core distribution and transport function, wholesale and retail and also the provision of households with consumer goods as well as the related information flows (Handfield and Nichols, 1999). These activities composing logistics are included into two major functions which are physical distribu￾tion; the derived transport segment, and materials management; the induced transport segment. More specifically: • Physical distribution (PD) is the collective term for the range of activities involved in the movement of goods from points of production to final points of sale and consumption (McKinnon, 1988, 33). It must insure that the mobility requirements of supply chains are entirely met. PD comprises all the functions of move￾ment and handling of goods, particularly transporta￾tion services (trucking, freight rail, air freight, inland waterways, marine shipping, and pipelines), trans￾shipment and warehousing services (e.g. consignment, storage, inventory management), trade, wholesale and, in principle, retail. Conventionally, all these activities are assumed to be derived from materials management demands. • Materials management (MM) considers all the activi￾ties related in the manufacturing of commodities in all their stages of production along a supply chain. MM includes production and marketing activities such as production planning, demand forecasting, purchasing and inventory management. It must in￾sure that the requirements of supply chains are met 1 The special issues on containerisation in GeoJournal 48, 1999 and on terminals in the Journal of Transport Geography 7, 1999. 172 M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184

M.Hesse,J.-P.Rodrigue I Journal of Transport Geography 12(2004)171-184 173 Manufacturing Requirements The purpose of the paper consequently is to assess Materials Management how geography is related to logistics and how logistics enforces a specific geography of production/distribution Induced Logistics Derived by exploring the integrated transport demand function. Demand (Integrated Demand) Demand It will first investigate the processes that have led to the integration of different functions into what has become Physical Distribution Mobility Requirements known as supply chain management.Second,its under- lying geographical dimensions are introduced,namely Fig.1.Logistics and integrated transport demand. the core concepts of flows,nodes and networks,which supply chain management has substantially modified. by dealing with a wide array of parts for assembly Last,spatial impedance (friction)factors linked to and raw materials,including packaging(for transport logistics are discussed and retailing)and,ultimately,recycling discarded commodities.2All these activities are assumed to be inducing physical distribution demands. 2.The evolution of logistics The close integration of PD and MM through logis- Although logistics were initially applied to military tics is blurring the induced/derived demand distinction. operations,its most significant impact is being felt The reciprocal relationship between the induced trans- through the functions of production.distribution and port demand function of physical distribution and the consumption(Rodrigue and Slack,2002).The exchange derived demand function of materials management is of goods is a constant feature of human economic thus considered as the integrated transport demand of activity.It was once essential for the rise of the mer- logistics (Fig.1).This implies that distribution,as al- cantile economy in medieval Europe(Braudel,1982)and ways,is derived from materials management activities became a large scale activity during the industrial rev- (namely production),but also,that these activities are olution.The location of industrial activity and thus the coordinated within distribution capabilities.Production, geography of manufacturing in general evolved with distribution and consumption are thus difficult to sep- respect to accessibility improvements that were partic- arate. ularly offered by railroads (which were then predomi- The more connected the different actors along a nantly freight related).Vice versa,every "long wave"in supply chain are,the harder it is to make a clear dis- the process of industrialization embodies distinct tinction between PD and MM as distribution channels transportation orientations and appropriate infrastruc- extend from suppliers to consumers and as responsibility ture requirements (Hayter,1997,27).This was true for for transport and warehousing is shared between man- the railroad in the fordist economy,as it is for trucking ufacturers,wholesalers and retailers(McKinnon,1988). and air freight more recently.The origins of the modern Logistics must be consistent with the products it sup- distribution sector go back to the emergence of the ports as customers tend to not place any difference be- capitalist economy,the development of specific modes tween a product and the distribution system that of industrial production and the unfolding of a partic- supplies it.Consequently,it is becoming increasingly ular division of labor.This created a distinct"sphere of difficult to consider transportation solely as a derived circulation",situated between production and con- demand,or industrial production,manufacturing and sumption (Marx.1939/1953).To a certain extent,cir- consumption solely as factors inducing transport de- culation allowed for the transition from use-value to mand.We thus argue that the classic transport geogra- exchange-value,and thus made possible the large-scale phy concept of derived freight demand has been blurred capitalization of commodities.Mass distribution and by the diffusion and adaptation of logistics.Manufac- marketing became incorporated in the practice of turing and mobility requirements are both embedded as modern management (Chandler,1977)and have been what is being produced,a part,or a finished product, significant factors of wealth generation. has to be moved at a similar rate along the supply chain. The organization and technology of modern distri- This paradigm shift leaning on supply chain manage- bution are embedded in a changing macro-and micro- ment,materials flow management and freight trans- economic framework.It can be roughly characterized by portation thus requires the elaboration of its own the terms of flexibilization and globalization.Flexi- transport geography bilization represents,far beyond the narrow interpreta- tion of"flexible specialization",a highly differentiated, strongly market-and customer-driven mode of creating added-value.Contemporary production and distribu- 2This is often labeled as reverse distribution,or integrated repair tion is no longer subject to single-firm activity,but and return (Rodrigue et al.,2001). increasingly practiced in networks of suppliers and

by dealing with a wide array of parts for assembly and raw materials, including packaging (for transport and retailing) and, ultimately, recycling discarded commodities. 2 All these activities are assumed to be inducing physical distribution demands. The close integration of PD and MM through logis￾tics is blurring the induced/derived demand distinction. The reciprocal relationship between the induced trans￾port demand function of physical distribution and the derived demand function of materials management is thus considered as the integrated transport demand of logistics (Fig. 1). This implies that distribution, as al￾ways, is derived from materials management activities (namely production), but also, that these activities are coordinated within distribution capabilities. Production, distribution and consumption are thus difficult to sep￾arate. The more connected the different actors along a supply chain are, the harder it is to make a clear dis￾tinction between PD and MM as distribution channels extend from suppliers to consumers and as responsibility for transport and warehousing is shared between man￾ufacturers, wholesalers and retailers (McKinnon, 1988). Logistics must be consistent with the products it sup￾ports as customers tend to not place any difference be￾tween a product and the distribution system that supplies it. Consequently, it is becoming increasingly difficult to consider transportation solely as a derived demand, or industrial production, manufacturing and consumption solely as factors inducing transport de￾mand. We thus argue that the classic transport geogra￾phy concept of derived freight demand has been blurred by the diffusion and adaptation of logistics. Manufac￾turing and mobility requirements are both embedded as what is being produced, a part, or a finished product, has to be moved at a similar rate along the supply chain. This paradigm shift leaning on supply chain manage￾ment, materials flow management and freight trans￾portation thus requires the elaboration of its own transport geography. The purpose of the paper consequently is to assess how geography is related to logistics and how logistics enforces a specific geography of production/distribution by exploring the integrated transport demand function. It will first investigate the processes that have led to the integration of different functions into what has become known as supply chain management. Second, its under￾lying geographical dimensions are introduced, namely the core concepts of flows, nodes and networks, which supply chain management has substantially modified. Last, spatial impedance (friction) factors linked to logistics are discussed. 2. The evolution of logistics Although logistics were initially applied to military operations, its most significant impact is being felt through the functions of production, distribution and consumption (Rodrigue and Slack, 2002). The exchange of goods is a constant feature of human economic activity. It was once essential for the rise of the mer￾cantile economy in medieval Europe (Braudel, 1982) and became a large scale activity during the industrial rev￾olution. The location of industrial activity and thus the geography of manufacturing in general evolved with respect to accessibility improvements that were partic￾ularly offered by railroads (which were then predomi￾nantly freight related). Vice versa, every ‘‘long wave’’ in the process of industrialization embodies distinct transportation orientations and appropriate infrastruc￾ture requirements (Hayter, 1997, 27). This was true for the railroad in the fordist economy, as it is for trucking and air freight more recently. The origins of the modern distribution sector go back to the emergence of the capitalist economy, the development of specific modes of industrial production and the unfolding of a partic￾ular division of labor. This created a distinct ‘‘sphere of circulation’’, situated between production and con￾sumption (Marx, 1939/1953). To a certain extent, cir￾culation allowed for the transition from use-value to exchange-value, and thus made possible the large-scale capitalization of commodities. Mass distribution and marketing became incorporated in the practice of modern management (Chandler, 1977) and have been significant factors of wealth generation. The organization and technology of modern distri￾bution are embedded in a changing macro- and micro￾economic framework. It can be roughly characterized by the terms of flexibilization and globalization. Flexi￾bilization represents, far beyond the narrow interpreta￾tion of ‘‘flexible specialization’’, a highly differentiated, strongly market- and customer-driven mode of creating added-value. Contemporary production and distribu￾tion is no longer subject to single-firm activity, but increasingly practiced in networks of suppliers and 2 This is often labeled as reverse distribution, or integrated repair and return (Rodrigue et al., 2001). Logistics (Integrated Demand) Materials Management Physical Distribution Derived Demand Induced Demand Manufacturing Requirements Mobility Requirements Fig. 1. Logistics and integrated transport demand. M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184 173

174 M.Hesse,J.-P.Rodrigue Journal of Transport Geography 12 (2004)171-184 subcontractors (Dicken and Thrift.1992:Gertler,1992: principles,firms were following a more integrated ap- Hudson,2001).The supply chain bundles together all proach,thus responding to the upcoming demand for this by information,communication,cooperation,and, flexibility without raising costs.At the same time,many last but not least,by physical distribution (Bowersox firms took advantage of new manufacturing opportu- et al.,2000).Globalization means that the spatial frame nities in developing countries.As production became for the entire economy has been expanded,implying the increasingly fragmented.activities related to its man- spatial expansion of the economy,more complex global agement were consolidated.Spatial fragmentation be- economic integration,and an intricate network of global came a by-product of economies of scale in distribution fows and hubs(cf.Dicken.1998:Held et al..1999:Knox In the 1990s.with the convergence of logistics and and Agnew,1998).Logistics thus developed against the information technologies,this principle was increasingly background of long-term structural change in economy, applied to the whole supply chain,particularly to the technology and society affecting all major industrialized function of distribution.In some highly efficient facili- countries (IMF,2001).These interrelated changes com- ties,the warehousing function went down as far as 15 prise sectoral and structural changes,mainly the rise of min worth of parts in inventory.It is now being intro- service economies,the increasing share of goods with duced in service functions such as wholesale and retail high value and low weight,consumerism,the upcoming where inventory in stores are kept at a minimum and high tech and knowledge based sectors (Castells.1996): resupplied on a daily basis. they also include a new political framework,namely the Whereas contemporary logistics was originally dedi- policies of deregulation and liberalization that were cated to the automation of production processes,in effective for the US in the late 1970s and early 1980s, order to organize industrial manufacturing as efficiently and for Europe since the introduction of the Single as possible,the subsequent modernization of logistics European Market in 1992 (Knowles and Hall,1998). can been characterized by an increasing degree of inte- These policies are now also being adopted by many gration.This trend was already on the way in the 1960s, developing countries,with varied success. as a key area for future productivity improvements The principles of modern logistics can be traced back (Bowersox et al.,1968).However,only with the imple- to Taylor (1947),the conceptual creator of fordism. mentation of modern information and communication Although his ideas were related to improving efficiency technologies did this assumption become possible.They within the factory by organizing the worker's tasks allow for the integrated management and control of along an assembly line,the temporal dimensions intro- information,finance and goods flows and made possible duced.such as sequence.duration,schedule.rhythm. a new range of production and distribution systems synchronization and time perspective are of high sig- (Abernathy et al.,2000).Step by step,and according to nificance to logistical management.What Taylor's improvements in information and communication principles of management were able to achieve within technologies,the two ends of the assembly line became the factory in terms of productivity improvements, integrated into the logistics of the supply chain:the logistics are able to achieve between elements of the timely supply of raw materials and components from supply chain:a system of integrated factories.A mile- outside,and the effective organization of distribution stone that marked rapid changes in the entire distribu- and marketing (Fig.2). tion system was the invention of the concept of lean Flexible order and supply behavior is actually made management,primarily in manufacturing (Womack possible by new technologies,primarily through the real- et al.,1990;Harrison,1997).One of the main premises time exchange of information.Because of information of lean management is eliminating inventories and and communication technologies,firms are able to order organizing materials supply strictly on demand,replac- from point-of-sale,to adjust inventories to meeting de- ing the former storage and stock keeping of inventory. mand (if not to eliminate them entirely)and to reduce During the 1980s,the application of this"principle of redundancy almost totally.Features such as electronic flow"permitted the reduction of inventories in time- data interchange(EDI).automated product flow in dis- sensitive manufacturing activities from several days' tribution centers (DCs)and warehouses,or the recent worth to several hours.Much of these efforts initially computer based tracing-and-tracking systems-which took place within the factory,while supply and output offer on-line control of shipped parcels via the web-are flowed as batches from suppliers and to distributors. primary sources of enormous productivity gains over last High rack storages,which later became automatically two decades(Fig.3).They still seem to be more important driven,or the internal movement of packages by flat than innovations associated with the upcoming electronic robots were early expressions of logistical engineering. commerce(Hesse,2002a;OECD/ECMT,2001). Initially,logistics was an activity divided around the While cycle time requirements substantially decreased supplying,warehousing,production and distribution from the 1960s to 1980s,this came at the expense of functions,most of them being fairly independent from growing logistics costs,notably inventory.From that the other.With the new organization and management point on,the major achievements were related to pro-

subcontractors (Dicken and Thrift, 1992; Gertler, 1992; Hudson, 2001). The supply chain bundles together all this by information, communication, cooperation, and, last but not least, by physical distribution (Bowersox et al., 2000). Globalization means that the spatial frame for the entire economy has been expanded, implying the spatial expansion of the economy, more complex global economic integration, and an intricate network of global flows and hubs (cf. Dicken, 1998; Held et al., 1999; Knox and Agnew, 1998). Logistics thus developed against the background of long-term structural change in economy, technology and society affecting all major industrialized countries (IMF, 2001). These interrelated changes com￾prise sectoral and structural changes, mainly the rise of service economies, the increasing share of goods with high value and low weight, consumerism, the upcoming high tech and knowledge based sectors (Castells, 1996); they also include a new political framework, namely the policies of deregulation and liberalization that were effective for the US in the late 1970s and early 1980s, and for Europe since the introduction of the Single European Market in 1992 (Knowles and Hall, 1998). These policies are now also being adopted by many developing countries, with varied success. The principles of modern logistics can be traced back to Taylor (1947), the conceptual creator of fordism. Although his ideas were related to improving efficiency within the factory by organizing the worker’s tasks along an assembly line, the temporal dimensions intro￾duced, such as sequence, duration, schedule, rhythm, synchronization and time perspective are of high sig￾nificance to logistical management. What Taylor’s principles of management were able to achieve within the factory in terms of productivity improvements, logistics are able to achieve between elements of the supply chain: a system of integrated factories. A mile￾stone that marked rapid changes in the entire distribu￾tion system was the invention of the concept of lean management, primarily in manufacturing (Womack et al., 1990; Harrison, 1997). One of the main premises of lean management is eliminating inventories and organizing materials supply strictly on demand, replac￾ing the former storage and stock keeping of inventory. During the 1980s, the application of this ‘‘principle of flow’’ permitted the reduction of inventories in time￾sensitive manufacturing activities from several days’ worth to several hours. Much of these efforts initially took place within the factory, while supply and output flowed as batches from suppliers and to distributors. High rack storages, which later became automatically driven, or the internal movement of packages by flat robots were early expressions of logistical engineering. Initially, logistics was an activity divided around the supplying, warehousing, production and distribution functions, most of them being fairly independent from the other. With the new organization and management principles, firms were following a more integrated ap￾proach, thus responding to the upcoming demand for flexibility without raising costs. At the same time, many firms took advantage of new manufacturing opportu￾nities in developing countries. As production became increasingly fragmented, activities related to its man￾agement were consolidated. Spatial fragmentation be￾came a by-product of economies of scale in distribution. In the 1990s, with the convergence of logistics and information technologies, this principle was increasingly applied to the whole supply chain, particularly to the function of distribution. In some highly efficient facili￾ties, the warehousing function went down as far as 15 min worth of parts in inventory. It is now being intro￾duced in service functions such as wholesale and retail where inventory in stores are kept at a minimum and resupplied on a daily basis. Whereas contemporary logistics was originally dedi￾cated to the automation of production processes, in order to organize industrial manufacturing as efficiently as possible, the subsequent modernization of logistics can been characterized by an increasing degree of inte￾gration. This trend was already on the way in the 1960s, as a key area for future productivity improvements (Bowersox et al., 1968). However, only with the imple￾mentation of modern information and communication technologies did this assumption become possible. They allow for the integrated management and control of information, finance and goods flows and made possible a new range of production and distribution systems (Abernathy et al., 2000). Step by step, and according to improvements in information and communication technologies, the two ends of the assembly line became integrated into the logistics of the supply chain: the timely supply of raw materials and components from outside, and the effective organization of distribution and marketing (Fig. 2). Flexible order and supply behavior is actually made possible by new technologies, primarily through the real￾time exchange of information. Because of information and communication technologies, firms are able to order from point-of-sale, to adjust inventories to meeting de￾mand (if not to eliminate them entirely) and to reduce redundancy almost totally. Features such as electronic data interchange (EDI), automated product flow in dis￾tribution centers (DCs) and warehouses, or the recent computer based tracing-and-tracking systems––which offer on-line control of shipped parcels via the web––are primary sources of enormous productivity gains over last two decades (Fig. 3). They still seem to be more important than innovations associated with the upcoming electronic commerce (Hesse, 2002a; OECD/ECMT, 2001). While cycle time requirements substantially decreased from the 1960s to 1980s, this came at the expense of growing logistics costs, notably inventory. From that point on, the major achievements were related to pro- 174 M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184

M.Hesse,J.-P.Rodrigue I Journal of Transport Geography 12(2004)171-184 175 1960s Demand Forecasting 1980s Purchasing Materials Requirements Planning Management Production Planning 1990s Manufacturing Inventory 2000s Warehousing Materials Handling Supply Chain Logistics Management Packaging Inventory Distribution Planning Physical Order Processing Information Technology Distribution Transportation Marketing Customer Service Strategic Planning Fig.2.Evolution of logistical integration,1960-2000 20 40 plier,the producer,the distributor (e.g.a wholesaler,a 18 Logistics Costs (GDP) freight forwarder,a carrier),the retailer,the end con- 口Imventory Cos克%GDP) 35 Cycle Time Requirements(days sumer,all of whom represent important players and 30 particular interests (Bovet et al.,2000;Bowersox et al., 25 2000).Compared with the former,more traditional shape 20 of the freight transport system,the evolution of supply chain management and the related emergence of the logistics industry is mainly characterized by four features: First,a fundamental restructuring of goods merchan- dising by establishing integrated supply chains with 1960s 1970s 1980s 1990s 2000s integrated freight transport demand. Fig.3.Logistical improvements,manufacturing sector,1960-2000. Second,whereas transport was traditionally regarded as a tool for overcoming space,logistics is critical in terms of time.This was achieved by shifts towards vertical integration,namely subcontracting and out- ductivity gains in distribution,accompanied by a sourcing,including the logistical function itself(Har- reduction of cycle time requirements,but as impor- vey,1989). tantly,of inventory costs.Another important require- Third,according to macro-economic structural ment was containerization,which conferred substantial changes,demand-side oriented activities are becom- flexibility to production systems in addition to the ing predominant.While traditional delivery was pri- container being its own storage unit.The expansion of marily managed by the supply side,current supply classical infrastructure such as highways,terminals and chains are increasingly managed by demand. airports was also essential for the development of Fourth,the logistics services are becoming complex modern logistics.The remarkable growth of freight and time-sensitive to the point that many firms are transport could not have happened without extensive now sub-contracting parts of their supply chain man- networks of freeways for regional and long-distance agement to third-party logistics providers.These pro- traffic,just as the railway system had provided the basis viders benefit from economies of scale and scope by for industrialization decades earlier. offering integrated solutions to many freight distribu- The flow-oriented mode of corporate management tion problems and organization currently affects almost every single activity within the entire process of value creation.The core component of materials management is the supply 3.The core geographical dimensions of logistics chain,the time-and space-related arrangement of the whole goods flow between supply,manufacturing,dis- The structural change in distribution and logistics tribution and consumption.Its major parts are the sup- has distinct geographical dimensions,investigated by

ductivity gains in distribution, accompanied by a reduction of cycle time requirements, but as impor￾tantly, of inventory costs. Another important require￾ment was containerization, which conferred substantial flexibility to production systems in addition to the container being its own storage unit. The expansion of classical infrastructure such as highways, terminals and airports was also essential for the development of modern logistics. The remarkable growth of freight transport could not have happened without extensive networks of freeways for regional and long-distance traffic, just as the railway system had provided the basis for industrialization decades earlier. The flow-oriented mode of corporate management and organization currently affects almost every single activity within the entire process of value creation. The core component of materials management is the supply chain, the time- and space-related arrangement of the whole goods flow between supply, manufacturing, dis￾tribution and consumption. Its major parts are the sup￾plier, the producer, the distributor (e.g. a wholesaler, a freight forwarder, a carrier), the retailer, the end con￾sumer, all of whom represent important players and particular interests (Bovet et al., 2000; Bowersox et al., 2000). Compared with the former, more traditional shape of the freight transport system, the evolution of supply chain management and the related emergence of the logistics industry is mainly characterized by four features: • First, a fundamental restructuring of goods merchan￾dising by establishing integrated supply chains with integrated freight transport demand. • Second, whereas transport was traditionally regarded as a tool for overcoming space, logistics is critical in terms of time. This was achieved by shifts towards vertical integration, namely subcontracting and out￾sourcing, including the logistical function itself (Har￾vey, 1989). • Third, according to macro-economic structural changes, demand-side oriented activities are becom￾ing predominant. While traditional delivery was pri￾marily managed by the supply side, current supply chains are increasingly managed by demand. • Fourth, the logistics services are becoming complex and time-sensitive to the point that many firms are now sub-contracting parts of their supply chain man￾agement to third-party logistics providers. These pro￾viders benefit from economies of scale and scope by offering integrated solutions to many freight distribu￾tion problems. 3. The core geographical dimensions of logistics The structural change in distribution and logistics has distinct geographical dimensions, investigated by Demand Forecasting Purchasing Requirements Planning Production Planning Manufacturing Inventory Warehousing Materials Handling Packaging Inventory Distribution Planning Order Processing Transportation Customer Service Strategic Planning Materials Management Physical Distribution Logistics Supply Chain Management Information Technology Marketing 1980s 1990s 2000s 1960s Fig. 2. Evolution of logistical integration, 1960–2000. 0 2 4 6 8 10 12 14 16 18 20 1960s 1970s 1980s 1990s 2000s % of GDP 0 5 10 15 20 25 30 35 40 Days Logistics Costs (% GDP) Inventory Costs (% GDP) Cycle Time Requirements (days) Fig. 3. Logistical improvements, manufacturing sector, 1960–2000. M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184 175

176 M.Hesse.J.-P.Rodrigue Journal of Transport Geography 12(2004)171-184 transport geography,and which are expressed in terms 3.1.Flows of flows (information,freight,transportation,vehicles), nodes and networks within the supply chain.Recent The growth of geographical areas of interaction and commentators also noted a rising interest in hubs,flows the temporal flexibilization of freight flows result in a and networks in the broader sense (Crang,2002;Cre- rising amount of freight transport.The traditional swell,2001).In the context of this paper,the system of arrangement of goods flow included the processing of physical distribution can be regarded as a material raw materials to manufacturers,with a storage function foundation for such mobility of people,goods,and usually acting as a buffer.The flow continued via information.These elements interact in space,but also wholesaler and/or shipper to retailer,ending at the final in time (Fig.4). customer.Delays were very common on all segments of Space/time convergence is a well-known concept in this chain and accumulated as inventories in ware- transport geography,where time was simply considered houses.There was a limited fow of information from as the amount of space that could be traded with a the consumer to the supply chain.This procedure is now specific amount of time,which included travel and going a different way,mainly by eliminating one or more transshipment.Logistics has expended this concept to of the costly operations in the supply chain organiza- include activities that were not previously considered tion.An important physical outcome of supply chain fully in space/time relationships.They now imply an management is the concentration of storage or ware- organization and synchronization of flows through housing in one facility,instead of several.This facility is nodes and network strategies.For instance.the con- increasingly being designed as a flow-and throughput- ventional expansion of a distribution system (Fig.4) oriented distribution center (DC),instead of a ware- involves a trade-off between marginal improvements of house holding cost intensive large inventories(Fig.5). spatial coverage (AS)and the associated marginal time Recent freight flows tend to be of lower volumes.of change (AT).Supply chain management enables a more higher frequency,often taking place over longer dis- efficient space/time convergence since the marginal dif- tances.These flows have been associated with modal ferences are larger for space (AS)than for time (AT1). adaptation.The magnitude of change can be charac- The emerging paradigm also includes in the space/time terized by the growth of geographical areas of interac. convergence the notion of time as a management con- tion,and by the temporal flexibilization of freight flows, straint of transportation.In addition of being a factor of both resulting in a rising amount of freight transport. distance/friction/impedance,time is a component of The distribution center thus becomes the core compo- flows,synchronized at nodes and the expression of a nent of such a distribution system(Fig.5). network structure.Consequently,the expansion and Truck vehicle miles have almost kept pace with GDP improvement of a distribution system from DSI to DS3 and have more than tripled between 1970 and 1999 in (Fig.4)could either imply an extended spatial coverage the US and in Europe (USDOT,2000).At the same with a similar amount of time or a similar spatial cov- time,freight by ton-miles developed more dynamically erage with a reduction of time,or ideally,a combination than by freight tons,which indicates both a demateri- of both.The distribution system is not only providing alization of the economy and rising average transport for the constant flow of commodities.but ensuring their distances.Structural change is associated with modal availability and accessibility in a synchronized-timely- shift away from rail and waterway transport,supporting fashion. road and air modes.Whereas intercity truck ton-miles DS1 Conventional ==DS2 Raw Materials Parts Manufacturing Distribution ■ 1T1 -DS3 2 Contemporary Supply Chain Managemen Supply Chain Management Flows Nodes Networks AS Space Fig.4.Geographical dimensions of logistics. Fig.5.Conventional and contemporary arrangement of freight flows

transport geography, and which are expressed in terms of flows (information, freight, transportation, vehicles), nodes and networks within the supply chain. Recent commentators also noted a rising interest in hubs, flows and networks in the broader sense (Crang, 2002; Cre￾swell, 2001). In the context of this paper, the system of physical distribution can be regarded as a material foundation for such mobility of people, goods, and information. These elements interact in space, but also in time (Fig. 4). Space/time convergence is a well-known concept in transport geography, where time was simply considered as the amount of space that could be traded with a specific amount of time, which included travel and transshipment. Logistics has expended this concept to include activities that were not previously considered fully in space/time relationships. They now imply an organization and synchronization of flows through nodes and network strategies. For instance, the con￾ventional expansion of a distribution system (Fig. 4) involves a trade-off between marginal improvements of spatial coverage (DS) and the associated marginal time change (DT ). Supply chain management enables a more efficient space/time convergence since the marginal dif￾ferences are larger for space (DS) than for time (DT 1). The emerging paradigm also includes in the space/time convergence the notion of time as a management con￾straint of transportation. In addition of being a factor of distance/friction/impedance, time is a component of flows, synchronized at nodes and the expression of a network structure. Consequently, the expansion and improvement of a distribution system from DS1 to DS3 (Fig. 4) could either imply an extended spatial coverage with a similar amount of time or a similar spatial cov￾erage with a reduction of time, or ideally, a combination of both. The distribution system is not only providing for the constant flow of commodities, but ensuring their availability and accessibility in a synchronized–timely￾fashion. 3.1. Flows The growth of geographical areas of interaction and the temporal flexibilization of freight flows result in a rising amount of freight transport. The traditional arrangement of goods flow included the processing of raw materials to manufacturers, with a storage function usually acting as a buffer. The flow continued via wholesaler and/or shipper to retailer, ending at the final customer. Delays were very common on all segments of this chain and accumulated as inventories in ware￾houses. There was a limited flow of information from the consumer to the supply chain. This procedure is now going a different way, mainly by eliminating one or more of the costly operations in the supply chain organiza￾tion. An important physical outcome of supply chain management is the concentration of storage or ware￾housing in one facility, instead of several. This facility is increasingly being designed as a flow- and throughput￾oriented distribution center (DC), instead of a ware￾house holding cost intensive large inventories (Fig. 5). Recent freight flows tend to be of lower volumes, of higher frequency, often taking place over longer dis￾tances. These flows have been associated with modal adaptation. The magnitude of change can be charac￾terized by the growth of geographical areas of interac￾tion, and by the temporal flexibilization of freight flows, both resulting in a rising amount of freight transport. The distribution center thus becomes the core compo￾nent of such a distribution system (Fig. 5). Truck vehicle miles have almost kept pace with GDP and have more than tripled between 1970 and 1999 in the US and in Europe (USDOT, 2000). At the same time, freight by ton-miles developed more dynamically than by freight tons, which indicates both a demateri￾alization of the economy and rising average transport distances. Structural change is associated with modal shift away from rail and waterway transport, supporting road and air modes. Whereas intercity truck ton-miles Space Time DS1 DS2 DS3 T1 Flows Nodes Networks S Supply Chain Management Supply Chain Management ∆ ∆T2 ∆T3 ∆ Fig. 4. Geographical dimensions of logistics. Raw Materials & Parts Manufacturing Distribution Raw Material Storage National Distribution Regional Storage Local Distribution Retailers Supply Chain Management Raw Material Distribution Center Manufacturing Retailers Customers Customers Conventional Contemporary Material flow (delivery) Information flow (order) Core component Fig. 5. Conventional and contemporary arrangement of freight flows. 176 M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184

M.Hesse,J.-P.Rodrigue I Journal of Transport Geography 12(2004)171-184 177 grew by 124%in the US 1970-1995,air carrier ton-miles and consumption (e.g.Hartwick,1998;Leslie and Rei- grew by 468%(Lakshmanan and Anderson,2002,p.9). mer,1999).Yet,it is still missing appropriate coverage Containerization has permitted lower volume flows of its freight and logistics dimension. while offering economies of scale by the consolidation of numerous shipments in batch flow units (e.g.,cellular 3.2.Nodes and locations containerships and doublestack trains). International trade increasingly contributes to the Due to new corporate strategies,a concentration of amount and the nature of physical distribution,since logistics functions in certain facilities at strategic loca- world exports have grown much faster than world tions is prevalent.Many improvements in freight flows production.Whereas the average annual growth rate of are achieved at terminals(Trip and Bontekoning,2002; world production during the 1990s accounts for 2.7%, Rodrigue,1999).Facilities are much larger than before, global imports and exports grew by 6.7%annually the locations being characterized by a particular con- (Wienert,2002,based on IMF data).In this respect, nection of regional and long-distance relations.Tradi- globalization is now considered a major framework tionally,freight distribution has been located at major condition of goods exchange,along the "rivers of trade" places of production,for instance in the manufacturing (McCray,1998).The internationalization of corporate belt at the North American east coast and in the Mid- activity,e.g.by the activity of transnational corpora- west,or in the old industrialized regions of England and tions (TNC).through foreign direct investments or in continental Europe.Today,particularly large-scale particular as a consequence of free trade agreements goods flows are directed through major gateways and such as NAFTA or the European Single Market,has hubs,mainly large ports and major airports,and at substantially contributed to the emergence of global highway intersections with access to a market area.The flows (Janelle and Beuthe,1997;Pedersen,2000; changing geography of manufacturing and industrial Woudsma,1999)(see Fig.6). production has been accompanied by a changing geo- Regarding the nature of goods flows as originally graphy of freight distribution.This is notably the case in derived from economic exchange,a considerable Pacific Asia,where industrialization and integration to amount of research has been conducted on the pro- global trade have been accompanied by the emergence duction or commodity chain approach.This concept of large scale high throughput nodes such as Hong was originally developed by Wallerstein and Hopkins in Kong,Singapore,Shanghai,Busan and Kaohsiung the context of world-system analysis and received a (Comtois and Rimmer,1997).Not surprisingly,these broader distribution by Gereffi and Korzeniewicz hubs are the largest container ports in the world. (1994).The commodity chain represents a conceptual Distribution is increasingly planned and operated on structure that allows for the analysis and assessment the basis of nationally designed networks,due to the of interlinked processes,with particular regard to in- premise of cost reduction by economics of scale. terdependencies between production,distribution and According to this particular pattern of re-structuring, consumption.In terms of economic (geography)re- favorite locations are either those gateways or trans- search,this concept marked a substantial progress in portation corridors with access both to traditional linking the micro-and macro-dimensions of production gateways of trade (interfaces)and to large consumer 9801981 8319841985198619871988198919901991199219931994199519961997199819992000 Total Merchandise Trade World Merchandise Production 10 Fig.6.Growth in the volume of world merchandise exports and world production,1980-2001 (Source:WTO,2002)

grew by 124% in the US 1970–1995, air carrier ton-miles grew by 468% (Lakshmanan and Anderson, 2002, p. 9). Containerization has permitted lower volume flows while offering economies of scale by the consolidation of numerous shipments in batch flow units (e.g., cellular containerships and doublestack trains). International trade increasingly contributes to the amount and the nature of physical distribution, since world exports have grown much faster than world production. Whereas the average annual growth rate of world production during the 1990s accounts for 2.7%, global imports and exports grew by 6.7% annually (Wienert, 2002, based on IMF data). In this respect, globalization is now considered a major framework condition of goods exchange, along the ‘‘rivers of trade’’ (McCray, 1998). The internationalization of corporate activity, e.g. by the activity of transnational corpora￾tions (TNC), through foreign direct investments or in particular as a consequence of free trade agreements such as NAFTA or the European Single Market, has substantially contributed to the emergence of global flows (Janelle and Beuthe, 1997; Pedersen, 2000; Woudsma, 1999) (see Fig. 6). Regarding the nature of goods flows as originally derived from economic exchange, a considerable amount of research has been conducted on the pro￾duction or commodity chain approach. This concept was originally developed by Wallerstein and Hopkins in the context of world-system analysis and received a broader distribution by Gereffi and Korzeniewicz (1994). The commodity chain represents a conceptual structure that allows for the analysis and assessment of interlinked processes, with particular regard to in￾terdependencies between production, distribution and consumption. In terms of economic (geography) re￾search, this concept marked a substantial progress in linking the micro- and macro-dimensions of production and consumption (e.g. Hartwick, 1998; Leslie and Rei￾mer, 1999). Yet, it is still missing appropriate coverage of its freight and logistics dimension. 3.2. Nodes and locations Due to new corporate strategies, a concentration of logistics functions in certain facilities at strategic loca￾tions is prevalent. Many improvements in freight flows are achieved at terminals (Trip and Bontekoning, 2002; Rodrigue, 1999). Facilities are much larger than before, the locations being characterized by a particular con￾nection of regional and long-distance relations. Tradi￾tionally, freight distribution has been located at major places of production, for instance in the manufacturing belt at the North American east coast and in the Mid￾west, or in the old industrialized regions of England and continental Europe. Today, particularly large-scale goods flows are directed through major gateways and hubs, mainly large ports and major airports, and at highway intersections with access to a market area. The changing geography of manufacturing and industrial production has been accompanied by a changing geo￾graphy of freight distribution. This is notably the case in Pacific Asia, where industrialization and integration to global trade have been accompanied by the emergence of large scale high throughput nodes such as Hong Kong, Singapore, Shanghai, Busan and Kaohsiung (Comtois and Rimmer, 1997). Not surprisingly, these hubs are the largest container ports in the world. Distribution is increasingly planned and operated on the basis of nationally designed networks, due to the premise of cost reduction by economics of scale. According to this particular pattern of re-structuring, favorite locations are either those gateways or trans￾portation corridors with access both to traditional gateways of trade (interfaces) and to large consumer -10 -5 0 5 10 15 20 25 1980198119821983198419851986198719881989199019911992199319941995199619971998199920002001 Total Merchandise Trade World Merchandise Production Fig. 6. Growth in the volume of world merchandise exports and world production, 1980–2001 (Source: WTO, 2002). M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184 177

178 M.Hesse.J.-P.Rodrigue Journal of Transport Geography 12 (2004)171-184 markets (destinations).Distribution firms respond to excellent accessibility,advanced terminal and transport structural change and competition by expanding the infrastructure,critical mass of logistics functions and infrastructure and rationalizing fows in order to meet attractive operating conditions(vis-a-vis its neighbours). the demand for quick and precise delivery.Owing to the Schiphol Airport and the Port of Rotterdam are among increased competition between distribution locations,all the most important hubs for international freight flows major freight hubs (large ports,freight airports,inland in Europe.Major population concentrations are well hubs)are currently committed to expanding their represented-Paris,London,the Ruhr area and Frank- infrastructure.In the case of North America,those furt (Europe's largest air cargo hub).Flanders in particular hubs are the strategic gateways at East and northern Belgium and the Nord-Pas de Calais region in West Coasts,e.g.the Ports of the San Pedro Bay in Los northern France also score highly.UK distributors tend Angeles,the Port of Seattle/Tacoma,the Port of New to prefer north-west Europe,due to improved access to York/New Jersey.The expansion of such places is pri- the continent via the Channel Tunnel (cf.JonesLang- marily due to the growth of trade and transport in LaSalle,2001). general,supported by economic growth and the The contemporary location of distribution centers is enlargement of market areas,both favoring scale econ- an outcome of high pressure on supply chains,caused by omies.Yet the strategy of concentrating freight at hub accelerated information transfers,changing consumer locations is increasingly becoming restricted,due to preferences and rising competition.Since many parts of density,land constraints,and congested traffic arterials. the supply chain are now globally integrated,distribu- Such limits to expansion and the scarce hinterland tion centers tend to be the link between global sourcing connections of major hubs are considered the most and regional distribution.The DC has become an important obstacle for further developing major hub interface between the geographies of manufacturing and locations. retailing,consequently handling the distribution scale As a consequence,.so called“Inland Hubs'”are and scope.Innovations such as containerization and becoming more and more important,where primarily particularly IT developments have integrated all com- road and air freight is consolidated.These new DC areas ponents of the chain.In response,major players in the are mainly affiliated with the interstate network and air distribution business (e.g.container shipping lines, cargo facilities.Consequently,warehousing,trucking, freight forwarders,warehousing firms,terminal opera- freight forwarding and air cargo activities are major tors)are trying to control as many parts of the logistics indicators and drivers of this new distribution economy. chain as possible.Not coincidentally,these firms are One of such new inland hubs is emerging along the Ohio challenged by vertical and horizontal linkages,by River Valley,particularly following a corridor from mergers,takeovers and strategic alliances (Slack et al., Ohio and Indiana to Tennessee."The 'first generation' 2002).For them,staying competitive means increasing e-fulfillment providers are gravitating towards the pre- the throughput and providing the demanded services at ferred location for a single,centralized distribution low rates.As a result,the activity space of main ports is facility,the greater Ohio River Valley,namely the states increasingly becoming relocated to low cost locations of Ohio,Indiana.Kentucky,and Tennessee.Industrial reaching far beyond traditional terminal sites and con- markets such as Columbus/OH.Indianapolis/IN.Heb- necting more distant places of their hinterlands. ron/KY (Cincinnati/OH)and Louisville/KY have seen Regarding the location issue,corporate decision substantial demand from these users."(Abbey et al., makers are used to carefully assessing advantages and 2001,15).In 1997,more than 150 distribution centers disadvantages of different locations.Compared with larger than 50,000 square feet were located only in the core urban areas,suburban sites offer larger and cheaper City of Columbus/OH.Both inventory and recent land resources,unrestricted transport access,a 'robust' absorption in the Columbus industrial real estate sub- environment for round-the-clock operations,and the market belongs to 80%to warehousing(SIOR database, locational advantage of intersections,connecting local 2001). and long-distance flows (Hesse,2002b).Existing facili- Recent European developments seem to be compa- ties often do not fit into the customers'profile,partic- rable.European Distribution Centers (EDC)are ularly with old buildings,or if they are surrounded by becoming larger,as the pressure to consolidate distri- sensitive neighborhoods.Trade-offs between inventory bution centers into pan-European centers continues. and transport costs are also highly supportive for sub- With access to a significant part of the European mar- urban locations,since mobilities(freight transport)and ketplace required,core Europe is the preferred loca- immobilities (land use)are closely intertwined.In order tion-most notably Benelux and eastern France. to find the optimal ratio between low land prices and National and regional centers are under pressure in all short distances to the point of final distribution,firms these countries,as distributors attempt to offload this move their DC location as far away from expensive land layer of warehousing.The Netherlands is emerging as the markets as necessary.For logistical and cost reasons, most favored location for European logistics,due to they also need to stay as close to their customers as

markets (destinations). Distribution firms respond to structural change and competition by expanding the infrastructure and rationalizing flows in order to meet the demand for quick and precise delivery. Owing to the increased competition between distribution locations, all major freight hubs (large ports, freight airports, inland hubs) are currently committed to expanding their infrastructure. In the case of North America, those particular hubs are the strategic gateways at East and West Coasts, e.g. the Ports of the San Pedro Bay in Los Angeles, the Port of Seattle/Tacoma, the Port of New York/New Jersey. The expansion of such places is pri￾marily due to the growth of trade and transport in general, supported by economic growth and the enlargement of market areas, both favoring scale econ￾omies. Yet the strategy of concentrating freight at hub locations is increasingly becoming restricted, due to density, land constraints, and congested traffic arterials. Such limits to expansion and the scarce hinterland connections of major hubs are considered the most important obstacle for further developing major hub locations. As a consequence, so called ‘‘Inland Hubs’’ are becoming more and more important, where primarily road and air freight is consolidated. These new DC areas are mainly affiliated with the interstate network and air cargo facilities. Consequently, warehousing, trucking, freight forwarding and air cargo activities are major indicators and drivers of this new distribution economy. One of such new inland hubs is emerging along the Ohio River Valley, particularly following a corridor from Ohio and Indiana to Tennessee. ‘‘The
first generation’ e-fulfillment providers are gravitating towards the pre￾ferred location for a single, centralized distribution facility, the greater Ohio River Valley, namely the states of Ohio, Indiana, Kentucky, and Tennessee. Industrial markets such as Columbus/OH, Indianapolis/IN, Heb￾ron/KY (Cincinnati/OH) and Louisville/KY have seen substantial demand from these users.’’ (Abbey et al., 2001, 15). In 1997, more than 150 distribution centers larger than 50,000 square feet were located only in the City of Columbus/OH. Both inventory and recent absorption in the Columbus industrial real estate sub￾market belongs to 80% to warehousing (SIOR database, 2001). Recent European developments seem to be compa￾rable. European Distribution Centers (EDC) are becoming larger, as the pressure to consolidate distri￾bution centers into pan-European centers continues. With access to a significant part of the European mar￾ketplace required, core Europe is the preferred loca￾tion––most notably Benelux and eastern France. National and regional centers are under pressure in all these countries, as distributors attempt to offload this layer of warehousing. The Netherlands is emerging as the most favored location for European logistics, due to excellent accessibility, advanced terminal and transport infrastructure, critical mass of logistics functions and attractive operating conditions (vis-
a-vis its neighbours). Schiphol Airport and the Port of Rotterdam are among the most important hubs for international freight flows in Europe. Major population concentrations are well represented––Paris, London, the Ruhr area and Frank￾furt (Europe’s largest air cargo hub). Flanders in northern Belgium and the Nord-Pas de Calais region in northern France also score highly. UK distributors tend to prefer north-west Europe, due to improved access to the continent via the Channel Tunnel (cf. JonesLang￾LaSalle, 2001). The contemporary location of distribution centers is an outcome of high pressure on supply chains, caused by accelerated information transfers, changing consumer preferences and rising competition. Since many parts of the supply chain are now globally integrated, distribu￾tion centers tend to be the link between global sourcing and regional distribution. The DC has become an interface between the geographies of manufacturing and retailing, consequently handling the distribution scale and scope. Innovations such as containerization and particularly IT developments have integrated all com￾ponents of the chain. In response, major players in the distribution business (e.g. container shipping lines, freight forwarders, warehousing firms, terminal opera￾tors) are trying to control as many parts of the logistics chain as possible. Not coincidentally, these firms are challenged by vertical and horizontal linkages, by mergers, takeovers and strategic alliances (Slack et al., 2002). For them, staying competitive means increasing the throughput and providing the demanded services at low rates. As a result, the activity space of main ports is increasingly becoming relocated to low cost locations reaching far beyond traditional terminal sites and con￾necting more distant places of their hinterlands. Regarding the location issue, corporate decision makers are used to carefully assessing advantages and disadvantages of different locations. Compared with core urban areas, suburban sites offer larger and cheaper land resources, unrestricted transport access, a
robust’ environment for round-the-clock operations, and the locational advantage of intersections, connecting local and long-distance flows (Hesse, 2002b). Existing facili￾ties often do not fit into the customers’ profile, partic￾ularly with old buildings, or if they are surrounded by sensitive neighborhoods. Trade-offs between inventory and transport costs are also highly supportive for sub￾urban locations, since mobilities (freight transport) and immobilities (land use) are closely intertwined. In order to find the optimal ratio between low land prices and short distances to the point of final distribution, firms move their DC location as far away from expensive land markets as necessary. For logistical and cost reasons, they also need to stay as close to their customers as 178 M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184

M.Hesse,J.-P.Rodrigue Journal of Transport Geography 12 (2004)171-184 179 possible.Not coincidentally,most recent construction of capacity to handle large amounts of time-sensitive con- DCs and warehouses takes place in metropolitan re- signments.The logistical requirements of a hub-and- gions,at the urban fringe or beyond.Further,the spoke structure are consequently extensive as efficiency function of nodes has become more complex with sev- is dominantly derived at the hub's terminal.Routing eral distribution centers performing light manufacturing networks tend to use circular configurations where tasks such as assembly and especially packaging.The freight can be transshipped form one route to the other functions of production and distribution thus became at specific hubs.Pendulum networks characterizing blurred with logistical integration. many container shipping services are relevant examples of relatively fixed routing distribution networks. 3.3.Networks Achieving flexible routing is a complex network strategy requiring a high level of logistical integration as routes The spatial structure of contemporary transportation and hubs are shifting depending on anticipated varia- networks is the expression of the spatial structure of tions of the integrated freight transport demand. distribution.Network building leads to a shift towards larger distribution centers,often serving significant transnational catchments.However,this does not mean 4.The concept of friction in the transport geography of the demise of national or regional distribution centers. logistics with some goods still requiring a three-tier distribution system,with regional,national and international DCs. The concept of impedance,or the friction of space,is The structure of networks has also adapted to fulfill the central to many geographical considerations of eco- requirements of an integrated freight transport demand, nomic and social processes.Conventionally,this con- which can take many forms and operate at different cept was subjugated to issues concerning distance and scales (Fig.7). how to quantify it.Substantial economic research has Point-to-point distribution is common when special- focused on assessing impedance,the impacts of distance, ized and specific one-time orders have to be satisfied, time and elasticities on freight flows (Button.1993).As which often creates less-than-full-load as well as empty discussed so far,significant changes have incurred in return problems.The logistical requirements of such a freight transport nodes,flows and networks,which im- structure are minimal,but at the expense of efficiency. pacted on the concept of impedance.Logistics and Corridor structures of distribution often link high den- freight distribution,as a transport paradigm,require a sity agglomerations with services such as the landbridge review of this multidimensional concept to include four where container trains link seaboards.Traffic along the core elements,namely the traditional transport costs,but corridor can be loaded or unloaded at local/regional also the organization of the supply chain,and the trans- distribution centers.Hub-and-spoke networks have actional and physical environments in which freight dis- mainly emerged with air freight distribution and with tribution evolves.These four elements.which are high throughput distribution centers favored by parcel difficult to consider independently,jointly define the services(O'Kelly.1998:SRI International.2002).Such a concept of logistical friction and its possible improve- structure is made possible only if the hub has the ments 4.1.Transportllogistics cost Point-to-Point Fixed Routing Traditionally transports costs were considered as a distance decay function.The most significant consider- O 0 0 ations of logistics on transport costs are related to the functions of composition,transshipment and decom- Corridor Flexible Routing position.which have been transformed by logistics. More specifically,composition and decomposition costs, ◆ which involve activities such as packaging,warehousing, and assembly of goods into batches,can account to 10% Hub-and-Spoke ○Transshipment of production costs.A higher level of inventory man- ○Route node Route agement (e.g.lean management)can lead to significant ONetwork node Alter ative reduction in the logistical friction as well as terminal ○Unserviced node route improvements decreasing transshipment times and costs (Fig.8).Time is becoming as important as distance in Fig.7.Freight distribution and network strategies.Source:Adapted from Woxenius (2002)"Conceptual Modelling of an Intermodal Ex- the assessment of transportation costs and impedance. press Transport System",International Congress on Freight Transport As transport costs went down through space/time con- Automation and Multimodality,Delft,The Netherlands. vergence,the value of time went up proportionally.For

possible. Not coincidentally, most recent construction of DCs and warehouses takes place in metropolitan re￾gions, at the urban fringe or beyond. Further, the function of nodes has become more complex with sev￾eral distribution centers performing light manufacturing tasks such as assembly and especially packaging. The functions of production and distribution thus became blurred with logistical integration. 3.3. Networks The spatial structure of contemporary transportation networks is the expression of the spatial structure of distribution. Network building leads to a shift towards larger distribution centers, often serving significant transnational catchments. However, this does not mean the demise of national or regional distribution centers, with some goods still requiring a three-tier distribution system, with regional, national and international DCs. The structure of networks has also adapted to fulfill the requirements of an integrated freight transport demand, which can take many forms and operate at different scales (Fig. 7). Point-to-point distribution is common when special￾ized and specific one-time orders have to be satisfied, which often creates less-than-full-load as well as empty return problems. The logistical requirements of such a structure are minimal, but at the expense of efficiency. Corridor structures of distribution often link high den￾sity agglomerations with services such as the landbridge where container trains link seaboards. Traffic along the corridor can be loaded or unloaded at local/regional distribution centers. Hub-and-spoke networks have mainly emerged with air freight distribution and with high throughput distribution centers favored by parcel services (O’Kelly, 1998; SRI International, 2002). Such a structure is made possible only if the hub has the capacity to handle large amounts of time-sensitive con￾signments. The logistical requirements of a hub-and￾spoke structure are consequently extensive as efficiency is dominantly derived at the hub’s terminal. Routing networks tend to use circular configurations where freight can be transshipped form one route to the other at specific hubs. Pendulum networks characterizing many container shipping services are relevant examples of relatively fixed routing distribution networks. Achieving flexible routing is a complex network strategy requiring a high level of logistical integration as routes and hubs are shifting depending on anticipated varia￾tions of the integrated freight transport demand. 4. The concept of friction in the transport geography of logistics The concept of impedance, or the friction of space, is central to many geographical considerations of eco￾nomic and social processes. Conventionally, this con￾cept was subjugated to issues concerning distance and how to quantify it. Substantial economic research has focused on assessing impedance, the impacts of distance, time and elasticities on freight flows (Button, 1993). As discussed so far, significant changes have incurred in freight transport nodes, flows and networks, which im￾pacted on the concept of impedance. Logistics and freight distribution, as a transport paradigm, require a review of this multidimensional concept to include four core elements, namely the traditional transport costs, but also the organization of the supply chain, and the trans￾actional and physical environments in which freight dis￾tribution evolves. These four elements, which are difficult to consider independently, jointly define the concept of logistical friction and its possible improve￾ments. 4.1. Transport/logistics cost Traditionally transports costs were considered as a distance decay function. The most significant consider￾ations of logistics on transport costs are related to the functions of composition, transshipment and decom￾position, which have been transformed by logistics. More specifically, composition and decomposition costs, which involve activities such as packaging, warehousing, and assembly of goods into batches, can account to 10% of production costs. A higher level of inventory man￾agement (e.g. lean management) can lead to significant reduction in the logistical friction as well as terminal improvements decreasing transshipment times and costs (Fig. 8). Time is becoming as important as distance in the assessment of transportation costs and impedance. As transport costs went down through space/time con￾vergence, the value of time went up proportionally. For A B Point-to-Point A B Corridor A B Hub-and-Spoke A B Fixed Routing A B Flexible Routing Transshipment node Route node Network node Unserviced node Route Alternative route Fig. 7. Freight distribution and network strategies. Source: Adapted from Woxenius (2002) ‘‘Conceptual Modelling of an Intermodal Ex￾press Transport System’’, International Congress on Freight Transport Automation and Multimodality, Delft, The Netherlands. M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184 179

180 M.Hesse.J.-P.Rodrigue/Journal of Transport Geography 12 (2004)171-184 100% 4.2.Complexity of the supply chain 90% 80% An integrated freight transport system requires a high 70% level of coordination.The more complex the supply 60% Administrative Costs chain,the higher the friction since it involves both 50% Transportation Costs organizational and geographical complexity(see below). 40% Inventory Carrying Costs Under such circumstances,the logistical friction takes 30% 20% the form of an exponential growth function of the 10% complexity of the supply chain (Fig.9).A core dimen- 0% sion of this geographical complexity is linked with the 1980 1985 1990 1995 2000 level of spatial fragmentation of production and con- sumption.Globalization has thus been concomitant Fig.8.Logistics costs,US,1980-2000 (in billions of $)Source:Cass and ProLogic (2002). with a complexification of the supply chain and logisti- cal integration permitted to support it.Many industrial location concepts indirectly address this perspective by investigating how firms grow in space and how pro- instance,between 1950 and 1998,the average time in duction is organized to take advantage of comparative transit for imported goods fell from 40 days to about 10 advantages(Dicken,1998).The extended range of sup- days.Each transit day adds about 0.8%to the final cost pliers and the globalization of markets have put of goods.As such,20 days at sea adds the equivalent to increasing pressures on the supply chain,a problem a 16%tariff on international trade (Hummel,2001). partially solved by using high-throughput distribution Concomitantly,the costs of logistics in the American centers. economy went from about 16%of the GDP in 1980 to The geographical scale of the supply chain is linked 10%in2000(Fig.8). with a level of logistical friction as nationally oriented However,within the components of logistics costs, supply chains tend to be less complex than multinational the transportation segment has experienced absolute as supply chains,mainly because they are less spatially well as relative growth.While it accounted for 46.5%of fragmented.From an operational perspective,it con- total logistics costs in 1980,this share climbed to 58.6% siders a balance between the benefits derived from the in 2000 (Fig.8).Inventories are thus increasingly in increased fragmentation of the supply chain with the circulation and inventory costs were reduced propor- organizational costs that come along.At some point,it tionally.The issue of mobile inventories,as opposed to becomes excessively difficult to maintain the coherence the traditional concept of fixed inventories has blurred of the supply chain.The marginal costs of this function the assessment of logistics costs.Trade-offs between have substantially been reduced by information tech- fixed costs (inventories,warehouses,etc.)and variable nologies and corporate strategies such as mergers and costs (transportation)play a major role in corporate joint ventures,implying that increasingly complex sup- strategies,since the advancement of new technologies ply chains can be supported with the resulting allows for the mobilization of inventories and.sub- improvements in productivity,efficiency and reliability. sequently,the elimination of facilities-whereas the Consequently,it is possible to maintain or improve key deregulation of transport markets attracted firms to time-dependent logistical requirements over an extended expand their shipping and transportation activities,by geography of distribution,namely the availability of significantly lowering the freight rates.Thus companies parts and products,their order cycle time,and the fre- were able to reduce a considerable amount of total quency,on-time and reliability of deliveries.The con- distribution costs. solidation of logistical activities in high-throughput Decomposition Information technologi Clearance. Transshipment Regulation erminal imp ovement Inventory management Deregulation Composition Distance Tim Complexity Market area Transport Supply Chain Transactional Fig.9.Logistical friction

instance, between 1950 and 1998, the average time in transit for imported goods fell from 40 days to about 10 days. Each transit day adds about 0.8% to the final cost of goods. As such, 20 days at sea adds the equivalent to a 16% tariff on international trade (Hummel, 2001). Concomitantly, the costs of logistics in the American economy went from about 16% of the GDP in 1980 to 10% in 2000 (Fig. 8). However, within the components of logistics costs, the transportation segment has experienced absolute as well as relative growth. While it accounted for 46.5% of total logistics costs in 1980, this share climbed to 58.6% in 2000 (Fig. 8). Inventories are thus increasingly in circulation and inventory costs were reduced propor￾tionally. The issue of mobile inventories, as opposed to the traditional concept of fixed inventories has blurred the assessment of logistics costs. Trade-offs between fixed costs (inventories, warehouses, etc.) and variable costs (transportation) play a major role in corporate strategies, since the advancement of new technologies allows for the mobilization of inventories and, sub￾sequently, the elimination of facilities––whereas the deregulation of transport markets attracted firms to expand their shipping and transportation activities, by significantly lowering the freight rates. Thus companies were able to reduce a considerable amount of total distribution costs. 4.2. Complexity of the supply chain An integrated freight transport system requires a high level of coordination. The more complex the supply chain, the higher the friction since it involves both organizational and geographical complexity (see below). Under such circumstances, the logistical friction takes the form of an exponential growth function of the complexity of the supply chain (Fig. 9). A core dimen￾sion of this geographical complexity is linked with the level of spatial fragmentation of production and con￾sumption. Globalization has thus been concomitant with a complexification of the supply chain and logisti￾cal integration permitted to support it. Many industrial location concepts indirectly address this perspective by investigating how firms grow in space and how pro￾duction is organized to take advantage of comparative advantages (Dicken, 1998). The extended range of sup￾pliers and the globalization of markets have put increasing pressures on the supply chain, a problem partially solved by using high-throughput distribution centers. The geographical scale of the supply chain is linked with a level of logistical friction as nationally oriented supply chains tend to be less complex than multinational supply chains, mainly because they are less spatially fragmented. From an operational perspective, it con￾siders a balance between the benefits derived from the increased fragmentation of the supply chain with the organizational costs that come along. At some point, it becomes excessively difficult to maintain the coherence of the supply chain. The marginal costs of this function have substantially been reduced by information tech￾nologies and corporate strategies such as mergers and joint ventures, implying that increasingly complex sup￾ply chains can be supported with the resulting improvements in productivity, efficiency and reliability. Consequently, it is possible to maintain or improve key time-dependent logistical requirements over an extended geography of distribution, namely the availability of parts and products, their order cycle time, and the fre￾quency, on-time and reliability of deliveries. The con￾solidation of logistical activities in high-throughput 228 233 291 302 377 214 274 351 445 590 18 20 26 30 39 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1980 1985 1990 1995 2000 Administrative Costs Transportation Costs Inventory Carrying Costs Fig. 8. Logistics costs, US, 1980–2000 (in billions of $). Source: Cass and ProLogic (2002). Transport Supply Chain Transactional Composition Transshipment Decomposition Friction Terminal improvements Information technologies Deregulation Clearance, Inspection, Regulation Distance / Time Complexity Market area Inventory management } { { } } } } { Fig. 9. Logistical friction. 180 M. Hesse, J.-P. Rodrigue / Journal of Transport Geography 12 (2004) 171–184