Evolution of regional inequality in the global shipping network

Journal of Transport Geography 44 (2015) 1–12

Contents lists available at ScienceDirect

Journal of Transport Geography journal homepage: www.elsevier.com/locate/jtrangeo

Evolution of regional inequality in the global shipping network Xu Mengqiao ⇑, Li Zhenfu, Shi Yanlei, Zhang Xiaoling, Jiang Shufei Transportation Management College, Dalian Maritime University, Dalian, Liaoning, China

a r t i c l e

i n f o

Keywords: World regions Liner shipping Global shipping network Evolution Network analysis

a b s t r a c t Global shipping is a backbone of the global economy, and as such, it evolves alongside the development of trade and the elaboration of commodity chains. This paper investigates the evolution of regional inequality in the global shipping network by analyzing the changing positions of world regions during the period from 2001 to 2012. This was a period of both prosperity and recession in maritime shipping. Using data on inter-regional flow connections, the positions of seventeen regions in the global shipping network are analyzed in terms of their traffic development, centrality, dominance and vulnerability. The East Asian, Northwest European and Europe Mediterranean regions have consistently held the highest positions, while East African and North African regions have held the lowest positions. By commanding the largest flows in the network, East Asia assumes a dominant position. The Australasian, North American West Coast, Northwest European and Southern African regions show an increasing dependency on East Asia. The analysis also identifies a few emerging regions that have had the highest growth rates in total traffic volume and connectivity for the studied period, namely South American North Coast, South American East Coast, West Africa, Southern Africa and West Asia. The empirical results of this paper supplement existing research on global shipping network evolution. One implication of the analysis is that the traffic growth of East Asia does not imply that, there is an equivalent improvement in its position in the global shipping network. The paper also shows that indicators from network analysis may be used to provide a more nuanced understanding of port-regional development than existing measures based solely on total traffic volume. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Global shipping evolves alongside the development of trade and the elaboration of commodity chains, thus making it a meaningful looking glass for analyzing the global economy (Valentine et al., 2013). The significance of liner shipping to global trade can be inferred from the fact that over 70% of the seaborne trade, in terms of value, is transported by container ships (WTO, 2008). With the development of technology and the global economy, the liner shipping industry has developed in several ways, e.g. increasing ship size, prevalence of strategic alliances and other collaborative arrangements among carriers. Rapid economic growth of emerging countries in the last two decades (e.g. China, India and South Africa) had prompted shipping carriers to adjust their container deployment worldwide for better coverage of their service networks and higher revenues. As a consequence, the structure of ⇑ Corresponding author at: Dalian Maritime University, No. 1 Linghai Road, Dalian City, Liaoning Province, China. Tel.: +86 15842635122. E-mail addresses: [email protected] (M. Xu), [email protected] (Z. Li), [email protected] (Y. Shi), [email protected] (X. Zhang), [email protected] (S. Jiang). http://dx.doi.org/10.1016/j.jtrangeo.2015.02.003 0966-6923/Ó 2015 Elsevier Ltd. All rights reserved.

the global shipping network (GSN) has dynamically evolved in the last two decades. The economic conditions and trade situations of world regions are two main factors that influence the container deployment of shipping carriers. For example, Asia, Europe and North America are regarded as the three biggest trade zones, liner services among which constitute the East–west belt of the global shipping activities, while ports in Africa attract much fewer container vessels. The evolution of regional inequality in the GSN can be seen from the different development processes of world regions, thus making it a meaningful way to look into the GSN from a regional perspective. Total traffic volume has long been a widely used indicator to evaluate regional development in terms of maritime shipping, which can be seen not only from various reports on maritime transports provided by shipping consultants like Drewry and Lloyd’s list, but also academic research, e.g. Notteboom (1997, 2010). However, total traffic volume fails to provide detailed information about the spatial distribution of the traffic and the proportions shared with each linked region. Therefore, total traffic volume does not fully reflect the position of a region in the global shipping network due to its essence of fairly high level of aggregation. In the context of inter-node flow connections, network

2

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

analysis provides valid evidences for node position in the structure of a network (Freeman, 1979). The effectiveness of applying such a network perspective to global shipping has been proved by many empirical observations of port position in various shipping networks; see Ducruet et al. (2010a), Wang and Cullinane (2014) for a more extensive discussion. Empirical evidence from traffic growth of global container port system suggests five main successive waves of containerization with a shift from advanced economies to developing economies in specific regions (Guerrero and Rodrigue, 2014), i.e. East Asia, South Asia, South America. With a rising position in global trade over the last decade, East Asia has seen its liner shipping activities improved in terms of total traffic volume. On one hand, some established transfer hubs in East Asia play a significant role in global shipping, such as port of Singapore, Hong Kong, Kaohsiung, Busan. On the other hand, there are many rapidly growing ports with large throughputs, of which are mainly Chinese ports, e.g. port of Shanghai, Shenzhen, Qingdao, Ningbo, Dalian. As reported by Clarkson Research Services (2014), traffic growth in the global container ports has focused mainly on East Asia since 2000 and will be increasingly relied on this region: On one hand, the majority of intra-regional container traffic growth is expected within intra-Asia, which will continue to be bolstered by strong growth in trades between China and rapidly developing Asian economies such as Indonesia, Malaysia, Thailand and the Philippines. On the other hand, non-mainlane trades involving Asia, especially East Asia, are expected to grow robustly into the medium-term, e.g. East Asia–South America, East Asia–West Africa. In terms of total traffic volume, it seems that East Asia has had an exclusively significant influence on the global liner shipping market for a long time. However, does East Asia really acquire an equivalent position within the structure of the GSN as is indicated by its total traffic volume? Research question here is: To what extent can total traffic volume be regarded as an accurate indicator of actual regional development in light of maritime shipping? In other words, does the position of a region in the GSN rise synchronously with its traffic growth? Hypothesis proposed in this study is that the growth of traffic in East Asia does not necessarily imply an equivalent improvement in its position in the GSN. Within this context, the objective of this paper is to measure and map the evolution of positions of world regions in the GSN based on inter-regional flow connections from 2001 to 2012. This was a period of both prosperity and recession in maritime shipping, i.e. 2003–2008 and 2009–2012 respectively. Such an aim is achieved by analysis of traffic development, centrality, dominance and vulnerability. Regions possessing larger traffic volume, higher level of network centrality and dominance are of a higher position in the GSN. In addition, this paper also tries to assess whether or not there are emerging regions of maritime activity in the GSN that differ from those of 2001. The remainder of this paper starts with a review of shipping network research in Section 2. Section 3 introduces the division of the world regions and data sources, and Section 4 deals with the methodology and provides an exposition of sub-analyses in the position analysis of world regions. Empirical results are presented in Section 5, namely traffic evolution, centrality, dominance and vulnerability of word regions. Further discussions on some emerging regions are made in Section 6. Finally, the implications of the research findings and some conclusions are drawn. 2. Review of shipping network research With the wide prevalence of complex network theory and its application to transportation systems in recent decades, network analysis of real shipping systems is rising (Ducruet and Lugo, 2013). Most current studies can be classified into three categories in terms of the geographical coverage of the studied shipping

network: Global level (Hu and Zhu, 2009; Kaluza et al., 2010; Woolley-Meza et al., 2011), regional level (Ducruet et al., 2010b; Low et al., 2009; McCalla et al., 2005; Wang and Cullinane, 2014) and specific shipping lines (Fremont, 2007; Mu et al., 2009). Based on service data from global shipping lines, these studies are fruitful in two main aspects: Firstly, centrality indicators, mainly degree and betweenness, were proved as effective parameters in evaluating port position in the structure of shipping networks. Secondly, some statistical properties of the overall network structure were revealed, e.g. small world and scale free. As most of these studies focus on the static state of shipping networks in one year, further questions about the evolution dynamics of shipping networks can hardly be answered. In regard to investigation into the evolution of regional inequality in the GSN, there are a few studies that have done some relevant works on a port level. For instance, Ducruet et al. did a series of illuminating works on the evolution of liner shipping network from 1996 to 2006, a period of rapid change in port hierarchies and liner service configurations in the world. Starting with investigating the modification of hub-and-spokes structure in the Atlantic container shipping system (Ducruet et al., 2010b), Ducruet et al. (2010a) then explored the position changes of major hub ports within Northeast Asia and their respective tributary areas, and eventually the changing port hierarchy of the global liner shipping network (Ducruet and Notteboom, 2012). Fremont and Soppe (2004) examined changes in the position of Europe in the global shipping network during the period from 1994 to 2002, as well as the position of European ports within regional and global shipping network. These studies added empirical evidence to the ongoing process of regional integration in maritime shipping activities, where ports tend to exchange relatively more with ports within their regions rather than with ports outside their regions. Meanwhile, inter-regional shipping services are fundamental to commodity trade among the world regions. As such, investigation into the evolution of inequality in the GSN with a renewed interest in regional perspective analysis, which focuses on inter-regional flow connections, may provide a nuanced understanding of regional development within the context of global shipping. However, most of the current regional shipping network studies tend to focus on intra-regional flows and especially on regions with large volume of seaborne trade, i.e. Europe, East Asia and North America. For example, Notteboom (1997) examined concentration and deconcentration tendencies to illustrate how load center development had occurred in the European continental container port system for the period from 1980 to 1994. Notteboom (2010) again updated the above study by extending the research period into 1985–2008, and confirmed that containerization would not necessarily lead to further port concentration in Europe. Meanwhile, there are several studies interested in the evolution process comparison of different regions. For instance, in terms of port throughput, number of container ports and the concentration level, similar evolution processes were detected between the container port systems in China from 1979 to 2009 and the USA from 1970 to 2009 (Li et al., 2012). Traffic inequality in container port systems of Europe during a period from 1975 to 2003 and North America from 1985 to 2003 was investigated by a Gini decomposition analysis (Notteboom, 2006), and an obvious traffic concentration tendency was observed in the latter. Wilmsmeier and Notteboom (2011) compared the evolution processes between Northern Europe and South American West Coast in terms of their liner shipping network configurations, and sought for the determinants of maritime network development between two differently developed regions. Although the fact that evolution of world regions, in terms of liner shipping development, is unequal is well admitted in most of the existing studies, the way to explore their evolution processes

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

from these studies suffers from two main shortcomings. Firstly, as mentioned above, the evolution of regional development is mainly investigated through intra-regional liner shipping activities, i.e. inter-port linkages within individual regions. However, positions of world regions in the GSN should be determined by comparisons with each other, which require analysis based on inter-regional interactions and could not be achieved separately. Secondly, liner shipping development of a region is primarily indicated by its total traffic volume, which lacks detailed information on the spatial distribution of the traffic. Due to the complexity of the GSN realized by a large number of liner services, the positions of world regions in the GSN can hardly be accurately reflected by solely individual traffic measures. Therefore, when attempting to undertake a formal investigation of the positions of world regions within the context of the global shipping network, as is the intention herein, these previous studies are unable to answer. It is for this reason that it has not been possible for these studies to address issues such as the extent that total traffic volume can be regarded as an accurate indicator of actual regional development in maritime shipping. Secondly, inconsistency between traffic growth and position improvement of world regions, and the dominant positions of those regions possessing large volume of seaborne trade, are the gaps in the literature that this work seeks to plug. Covering a consecutive research period from 2001 to 2012, additionally, this study could supplement to the scarce existing literature on the evolution of the GSN with a study period over a decade.

3. World regions and data sources 3.1. World regions So far there has been no clear definition or division on world regions in terms of global shipping, neither in practices nor in academic research field. On one hand, shipping lines define their service regions and design the shipping routes mainly depending on their different market focuses. On the other hand, most shipping network studies confirm their study areas according to their individual research interests while paying very little attention to relevant division criteria. Among them, discussions on the concepts of maritime facade, maritime region and maritime range from Ducruet and Zaidi (2012) are inspiring: The maritime facade implies the notion of coastal continuity (e.g. North America’s East Coast), the maritime region may include two or more neighbouring coasts, and maritime range implies a certain level of regional integration which results in interdependency among neighboring ports. In this paper, a region or maritime region is defined as a geographical area with a group of ports together serving the area. Thus it is necessary to look into some common criteria for defining a port range: geographical adjacency, belonging to the same set of loops of inter-continental liner services and sharing an overlapping hinterland, see e.g. (Fleming, 1997; Notteboom, 2006; Verhoeff, 1977) for a more extensive discussion. As cargo ships travel under restrictions of the geographical distribution of coastlines, oceans and seas, physical geography is among the most fundamental factors affecting the emergence of the GSN and its structural change. Following the zone-differentiation in global liner shipping practice (Degerlund, 2006) as well as Containerisation International,1 seventeen regions are finally confirmed in this research from a geographic perspective (Fig. 1). One of the main problems in defining regions is to decide on the specific geographical coverage of a region or which ports to group into it. Admittedly, the number of ports within these regions differs a lot due to their difference in economic development and natural 1

Containerisation International was acquired by Lloyd’s list in 2013.

3

distribution of coastlines. We are well aware that it might influence the results. However, based on a concise analysis of existing literature on maritime region and market practice, we are confident that the division of these regions meets the research objective here. As a result, it is beyond this paper to conduct a sensitivity analysis to measure the effect of the different geographical coverage of individual regions. 3.2. Data sources Since actual container traffic data among world regions is quite unavailable, container deployment data is used to represent the inter-regional container traffic flow connections. The consecutive study period from 2001 to 2012 of this research is mainly determined by data availability to authors, and all the data used here are based on Containerisation International2 (CI-Online). This database provides container deployment data (in TEU) among world regions of the top 100 container lines in terms of the total TEU capacity, and it updates monthly. The combined container capacity of the top 100 in each studied year accounts for more than 96% of the world capacity, e.g. 97% in 2012.3 However, the main limitation of the data is the lack of detailed information about the container capacity on each direction between a pair of regions, which could reflect the global shipping activities more accurately. The weighted global shipping network of each year in this paper is constructed as follows. World regions are the nodes and inter-regional links with at least one connection form the edges. Inter-regional connections in the GSN are realized by liner circulations, where each liner circulation forms a complete graph among all the regions involved. The weight of an edge between two regions is the total container traffic between them, which is obtained by adding together the container capacities of vessels that have operated within the ports of the two regions on their individual service loops. In implementation, we take a liner circulation from East Asia to Southern Africa of a vessel named COSCO YINGKOU as an example. The container capacity of this vessel is 3534 TEU. SHA (Shanghai), NGB (Ningbo), KHH (Kaohsiung), HKG (Hong Kong), YTN (Yantian), TPT (Tanjung Pelepas), SIN (Singapore), DUR (Durban), CPT (CAPE TOWN), SIN, KHH, SHA are the sequential ports of call, where SHA, NGB, KHH, HKG, YTN, TPT and SIN belong to East Asia, DUR and CPT to Southern Africa. Therefore, this circulation contributes 3534 TEUs to the weight of the edge between East Asia and Southern Africa. Fig. 2 shows the container traffic distribution among world regions in 2001 and 2012. East Asia has an obviously superior position in the GSN in terms of total traffic volume, and shipping connection between East Asia and Northwest Europe has been greatly strengthened during the studied period, i.e. container traffic between them is much higher than other inter-regional shipping routes. In addition, there seems to be a lower concentration of container traffic in the GSN in 2012 than in 2001: Only shipping routes between East Asia, North America and Europe apparently stand out in 2001, while many other shipping routes outside this triangle have also seen their rapid traffic growth and rising market share during the period, e.g. South Asia–Europe, Australasia–East Asia, West Asia–East Asia, West Asia–Europe. 4. Methodology Fig. 3 gives a schematic representation of the methodology deployed. The concept of ‘‘position’’ of different regions is made 2 All the data used in this paper were provided by Containerisation International in November 2012, when it had not been acquired by Lloyd’s List. Now the data is available on Lloyd’s List, http://www.lloydslist.com/ll/. 3 Alphaliner. http://www.alphaliner.com/.

4

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

Fig. 1. World regions for global shipping. Note: Europe Med = Europe Mediterranean, NAWC = North American West Coast, NAEC = North American East Coast, NAGC = North American Gulf Coast, SANC = South American North Coast, SAWC = South American West Coast, SAEC = South American East Coast; Modified from Li et al. (2014).

operational by analyzing and comparing traffic development, centrality, dominance and vulnerability. Traffic development of a region is depicted by its total container traffic shared with the directly connected regions, which is a fundamental indicator to reflect the seaborne trade development of a region. That is, this paper measures the extra-regional container traffic (hereinafter, container traffic) of world regions from a perspective of inter-regional flow connections. Due to the inclusion of inland traffic and double counting of inter-port traffic within a region, traditional traffic indicator that simply aggregates container throughputs of ports within a region is not suitable here. Centrality of a region is reflected by connectivity and intermediacy, which are measured by degree (i.e. number of direct connections) and betweenness (i.e. number of possible shortest paths on which the region is positioned) respectively. Centrality in this research is defined from network analysis: The relative position of a given node or vertex with regard to the others in the network (Ducruet et al., 2010a). Although degree is a basic indicator of node position in the network structure, when combined with container traffic development indicator, it can further reveal the level of influence of individual regions. Regions with both highest degree value and largest container traffic are of maximum influence in the GSN; they are very likely to be the biggest trade regions, which hold sustainable container flows to and from the other regions of the world and attract plenty of liner services. Regions with lowest degree value and smallest container traffic are of minimum influence in the GSN; these regions not only possess small volume of seaborne trade, but also lack international hub ports to attract transshipment traffic. The influence of others is between the above two kinds, but regions with similar traffic volume may differ a lot in traffic diversification due to their different degree values. Betweenness is applied here to measure the extent a region is inserted into the GSN (Fleming and Hayuth, 1994). By analyzing the discrepancies between total traffic volume and the values of degree and betweenness of world regions, the discordance between traffic volumes and the actual regional positions in the GSN are revealed.

Dominance and vulnerability of the world regions are investigated by a classic network flow analysis, i.e. dominant flow analysis introduced by Nystuen and Dacey (1961). In implementation, the initial network of GSN is abstracted to a nodal network structure by focusing on the movement of the largest flows of individual regions. Dominant flow movements and their implications for network relationships are presented in Table 1. Regions possessing largest number of income flows show the highest dominant position or hub feature in the GSN. Meanwhile, the vulnerability of a region in this research is obtained by calculating the maximum percentage of container traffic that it shares with another (Gonzalez et al., 2012).

5. Empirical results 5.1. Traffic development analysis Fig. 4 shows the container traffic evolution of world regions from 2001 to 2012. Generally, there are three different periods for the traffic growth (Fig. 5): A slow growth period from 2001 to 2004, a rapid growth period from 2005 to 2008 and a declining period from 2009 to 2012. East Asia, Northwest Europe and Europe Mediterranean always ranked the top three during the period, with their market share ranging from 23.1% to 27.50%, 13.14% to 16.13% and 10.04% to 12.83% respectively. Therefore, in terms of inter-regional container traffic, East Asia undoubtedly possessed the highest market position. This can be explained by the rapid economic growth in some East Asian countries such as South Korea, China, Indonesia, Philippines and Vietnam, amongst which the Chinese ports have contributed a lot to the traffic development of East Asia. Ports called by long-haul international liner services were mainly transshipment hubs like Busan, Yokohama, Hong Kong, Kaohsiung, Kelang and Singapore in 2001 (Yap et al., 2006); however, many Chinese mainland ports emerged in recent years, e.g. Shanghai, Shenzhen, Qingdao and Ningbo. Since the Northwest European port system has more advantages accessing

5

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

2001

SouthAmericanEastCoast SouthAmericanEastCoast SouthAmericaEastCoast

SouthAmericanNorthCoast SouthAmericanNorthCoast SouthAmericaNorthCoast

WestAfrica WestAfrica

Caribean/CentralAmerican Caribbean/CentralAmerica Caribbean/CentralAmerica

Australasia Australasia NorthAmericanWestCoast NorthAmericanWestCoast NorthAmericaWestCoast

WestAsia WestAsia

NorthwestEurope

EastAsia EastAsia

EuropeMeditterranean SouthAsia SouthAsia NorthAmericanEastCoast NorthAmericaEastCoast NorthAmericanEastCoast

SouthernAfrica SouthernAfrica

NorthAmericanGulfCoast NorthAmericanGulfCoast NorthAmericaGulfCoast SouthAmericanWestCoast SouthAmericaWestCoast SouthAmericanWestCoast

2012

SouthAmericanEastCoast SouthAmericaEastCoast SouthAmericanEastCoast

EastAfrica EastAfrica

NorthAfrica

SouthAmericanNorthCoast SouthAmericanNorthCoast SouthAmericaNorthCoast

WestAfrica WestAfrica

Caribean/CentralAmerican Caribbean/CentralAmerica Caribbean/CentralAmerica

Australasia Australasia NorthAmericanWestCoast NorthAmericaWestCoast NorthAmericanWestCoast

WestAsia

NorthwestEurope NorthwestEurope EastAsia EuropeMeditterranean EuropeMeditterranean NorthAmericanEastCoast NorthAmericaEastCoast NorthAmericanEastCoast

SouthAsia

NorthAmericanGulfCoast NorthAmericaGulfCoast NorthAmericanGulfCoast SouthAmericanWestCoast SouthAmericanWestCoast SouthAmericaWestCoast

Container traffic(Unit:10,000TEU) 1,200 700 300 100

SouthernAfrica SouthernAfrica NorthAfrica NorthAfrica

EastAfrica EastAfrica

Edge weight(Unit:10,000TEU) 280

160

100

50

Fig. 2. Container traffic distribution among world regions in 2001 and 2012.

Traffic development analysis

Inputs˖ Container traffic on each directly linked edge of a region

Centrality analysis

Inputs: Number of shipping connections, number of possible shortest paths on which the region is positioned

Dominance and vulnerability analysis

Inputs: Number of largest inflows, maximum percentage of cargo shared with another

Fig. 3. Sub-analyses in the position analysis of world regions.

the European mainland hinterland than the Europe Mediterranean port systems (Gouvernal et al., 2012), traffic growth of the former region was probably caused by its seaborne trade development,

while the latter by its intermediate role for some geographically adjacent regions. For instance, container shipping between West Africa, North Africa and East Asia are often transshipped in port of Algeciras, Malaga, Marsaxlokk, Cagliari, Gioia Tauro, Piraeus. The market share of other individual regions is lower than 10% in each year. Among those with market share between 5% and 10%, i.e. Caribbean/Central America, North American East Coast, North American West Coast, South Asia and West Asia, Caribbean/Central America had experienced the highest traffic fluctuation during the period. Lying at the crossroads of major East–West and North–South liner shipping routes, Caribbean/Central America plays an important intermediate role in the GSN, which can be confirmed from the large proportion of transshipment containers in many ports within the region (McCalla et al., 2005; Rodrigue, 2012). As such, container traffic development of Caribbean/Central America is quite susceptible to

6

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

Table 1 Dominant flow movements and their implications for network relationships.a Spatial structure

Item description

Implications on inter-node relationship

Dominant flow of i is exclusively moving toward j i and j both have dominant flows toward each other Dominant flows of i heading to two or more destinations

i shows high dependence on j

i is the destination of dominant flows from two or more nodes

a

Complementary between i and j or mutual dependence Potential competition among destinations with the objective of flows from i

Hub features of i indicated by dominant flows from surrounding nodes

Adapted from Wang and Cullinane (2014).

these routes. Market share of North American East Coast and North American West Coast had fallen since the 2008 economic crisis, while South Asia and West Asia had increased. In spite of their small market shares (i.e. lower than 5%), the sustained traffic growth of South American East Coast, South American West Coast and West Africa since 2001 is noteworthy. Motivated by the economic and trade potential in these regions, shipping carriers had deployed more container capacity into them. Thus Durban, a Southern African port, appears to be a new relay hub between West Africa and Asia (Ducruet, 2012), as well as between South American East coast and Asia.

5.2. Centrality analysis With an average degree value of more than ten and a half of regions directly connected to at least ten other regions by liner services in each year from 2001 to 2012, most regions are well

connected to others (Fig. 6). Northwest Europe and Europe Mediterranean were always directly linked with the others, as well as East Asia after it was directly connected with North Africa in 2010. East Africa and North Africa had always been the least connected ones during the period, that their respective degree values were 6 and 5 in 2012. The connectivity of Australasia declined the most during the period, with the directly connected regions only restricted to East Asia, Europe Mediterranean, Northwest Europe, South Asia, North American East Coast, North American West Coast and Caribbean/Central America in 2012. On the contrary, the connectivity of South American North Coast and West Africa had significantly improved to the extent that, their degree values increased from 7 in 2001 to 10 in 2012. With the highest degree value and largest traffic volume during the period, East Asia, Europe Mediterranean and Northwest Europe have a big influence in the GSN. The top ten shipping routes in terms of container capacity are closely related with these regions especially East Asia (Table A.1), as well as the top twenty ports in container throughput. On the contrary, East Africa and North Africa are in the margin of the GSN due to their poor connectivity and small traffic volume, and the reasons may be as follows; one reason is that the volume of seaborne trade in the two regions is much smaller than the other regions as a result of their slow economic growth, the other is the general situation of poor port facilities as well as political instability that makes it difficult for them to attract more global liner services. Besides, the development of North African ports is constrained from its geographical adjacency to Europe Mediterranean, i.e. container cargo of North Africa is often transshipped in Mediterranean ports such as Algeciras and Valencia (Notteboom, 2012). The influence of the other regions is between the above two kinds, but the level of traffic diversification among those regions with similar traffic volume is not the same. For instance, with respective degree value of 11 and 8 but similar container traffic in 2012, Caribbean/Central America possessed higher traffic diversification than its neighboring region North American West Coast. Being at the crossroads of trans-Atlantic and North–south trade flows, the higher traffic diversification of Caribbean/Central America can be seen from its interlining

Fig. 4. Container traffic evolution of world regions from 2001 to 2012 on logarithmic scale.

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

7

Fig. 5. A dynamic positioning of world regions based on market share and TEU growth rate. Note: Europe Med = Europe Mediterranean, NAWC = North American West Coast, NAEC = North American East Coast, NAGC = North American Gulf Coast, SANC = South American North Coast, SAWC = South American West Coast, SAEC = South American East Coast.

Fig. 6. Maritime degree and betweenness of world regions.

function between some long-haul shipping routes. However, ports in North American West Coast mainly serve seaborne trade within this region. As a result of the further expansion of the Panama Canal and heavy investment in creating modern port infrastructure (Rodrigue, 2012), e.g. port of Manzanillo, Panama, Colon and Caucedo, the influence of the Caribbean/Central America in the GSN will rise in the future. Generally, the betweenness values of world regions are fairly low (Fig. 6), which imply that their overall intermediate function

is weak. The reason can be seen, to some degree, from their relatively high connectivity: The widely inter-regional connections in the GSN make it convenient for liner shipping between all pairs of regions, thus resulting in a relatively low degree of dependency on some intermediate regions. However, by comparing the total traffic volumes of some regions that possess similar betweenness values, the probability of the intermediate function of a few regions can be revealed. For instance, with the same highest betweenness value but a different level of the total traffic volume

8

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

in 2012, Europe Mediterranean has a more obviously intermediate function than East Asia that serves the GSN mainly as a big trade region. Although degree and betweenness values are correlated with total traffic volume (i.e. the average correlation coefficients during the studied period are 0.736 and 0.768 respectively), they do not change synchronously with the total traffic volume. On one hand, traffic growth does not necessarily imply an equivalent improvement in centrality. For example, the total traffic volume of North American East Coast and Australasia grew during the studied period, but they suffered obvious decline both in degree and betweenness. As they are increasingly focusing on liner shipping with some specific regions, they serve the global shipping primarily as trade regions rather than intermediate regions, thus their relative positions in the structure of the GSN are declining. Taking a closer look at East Asia, its position improvement apparently lags behind the traffic growth: Although East Asia had always possessed the definitely largest total traffic volume during the research period, its degree and betweenness values fell behind Northwest Europe and Europe Mediterranean before 2010. On the other hand, traffic decline does not necessarily imply an equivalent centrality loss. Despite that the total traffic volume of both Caribbean/Central America and South American North Coast went down from 2008 to 2012, the degree values were quite stable for the former and even went up for the latter. These observations, to some extent,

show a relative robustness of centrality of world regions in the structure of the GSN as compared to the total traffic volume. 5.3. Dominance and vulnerability analysis Dominant positions and traffic vulnerability of world regions in each year during the period from 2001 to 2012 were investigated through dominant flow analysis. With the dominant flows concentrating on a few regions (Fig. 7), there shows a hub and periphery structure in the GSN. Furthermore, the inequality in the distribution of dominant flows in the GSN had been strengthened during the period. As the number of largest inflows increased from 9 to 13, East Asia took up an absolutely dominant position in the GSN. That is, except for South American North Coast, North American Gulf Coast and North Africa, the largest traffic flows of the other regions were shared with East Asia in 2012. Among them, Australasia, Europe Mediterranean, North American West Coast, Northwest Europe, South Asia, Southern Africa and West Asia were always dominated by East Asia in terms of the inter-regional flow connections. On the contrary, the dominant position of Northwest Europe had declined since 2008: The largest flows of North Africa, West Africa, North American Gulf Coast and South American East Coast were shared with Northwest Europe before 2008, while in 2012 it was only North Africa and North American Gulf Coast that shared. However, benefited from the stable container traffic

Fig. 7. Dominant flow distribution of the GSN. Note: The vulnerability of a region is obtained by calculating the maximum percentage of container traffic it shares with another.

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

9

Fig. 7 (continued)

growth on the East Asia–Northwest Europe shipping route, Northwest Europe had also become the dominant region of East Asia instead of North American West Coast since 2008. In addition, the intermediate role of Caribbean/Central America in the local scale can be seen from its dominant influence on the traffic flow of South American North Coast. Considering the rapid economic growth in many East Asian countries, the rising dominant position of East Asia may be explained as follows; on one hand, led by China, Vietnam, Philippines and Malaysia, East Asia has become the biggest world factory since the late 1990s and generates a huge volume of container cargo, on the other hand, as a region with a largest number of developing countries in the world, vast shipping demand for raw material and commodity trade is being created in the globalization processes of these countries. Traffic vulnerability changes of the world regions during the period are also shown in Fig. 7. The rising traffic vulnerability of Australasia, North American West Coast, Northwest Europe and Southern Africa reveals their increasing dependency on East Asia, i.e. their shared traffic with East Asia in 2012 were 65.8%, 72.9%, 44.1% and 45.1% respectively. With container traffic increasing from 275758 TEU in 2001 to 563168 TEU in 2012, container shipping between Australasia and East Asia has become the busiest south to north shipping route in the world. Although the traffic vulnerability of East Asia is much lower than most regions, container traffic of East Asia shared with Northwest Europe is rising, i.e. 19.9% in 2008 and 24.2% in 2012. Except for sustained trade

development between them, it should also be explained by the fact that a lot of large container vessels, ordered before the 2008 economic crisis, had been put into the Northwest Europe–East Asia shipping route since 2011 (UNCTAD, 2012). As the traffic volume shared with East Asia increased by 57% in the total external container traffic of Southern Africa during the period, the East Asia–Southern Africa shipping route is becoming an important corridor for south to north liner shipping activities. Except for East Asia, other regions possessing relatively low traffic vulnerability (i.e. the largest traffic shared with another is lower than 30%) in 2012 were Caribbean/Central America, Europe Mediterranean, North American Gulf Coast and South American East Coast. The reason for the relatively low traffic vulnerability for the former three regions lies mainly in their interlining roles in local scale, which is consistent with their relatively high traffic diversification. South American East Coast has seen its position in the GSN improved, and it will be further discussed in the next section.

6. Emerging regions The uneven evolution process of world regions in the GSN during the period from 2001 to 2012 can be clearly seen from the above analyses. Meanwhile, there are also a few emerging regions whose positions improved a lot, i.e. South American North Coast, South American East coast, West Africa, Southern

10

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

Africa and West Asia. These regions not only get a traffic growth rate of over 100% during the period (i.e. 430.13%, 161.48%, 201.79%, 167.29% and 132.26% respectively), but also achieve an obvious improvement in connectivity. Attracted by the rapid economic development in South America in recent years, global shipping carriers are putting more container capacities into shipping routes related with this continent to provide it with better shipping connections with the other regions in the world; notwithstanding, global shipping lines entered the South American market typically configured to serve trades between North and South America before the twenty-first century (Guy, 2003). Along with the development of some transshipment ports in South American North Coast (Rodrigue, 2012; Veenstra et al., 2005), e.g. port of Cartagena and Curacao, the shared container traffic of South American North Coast with Caribbean/Central America declined from 42.8% in 2001 to 31% in 2012. That is, transshipment dependency of South American North Coast on the Caribbean/Central America is going down gradually. South American East Coast is being better integrated into the GSN, that it has been directly connected with twelve regions since 2009. As a consequence of the container traffic growth from 78604 TEU in 2001 to 294854 TEU in 2012, together with the rapid trade development of Brazil and Argentina, the South American East Coast– East Asia shipping route is playing a much important role in the global shipping activities. Although the total traffic volume of West Africa remains small, its traffic growth rate during the period ranks the second among the world regions. With the share of West Africa–East Asia shipping route in the total external traffic of West Africa increasing from 17% to 38%, this route has become one of the major focuses of the global shipping carriers in expanding their service networks. For example, the China–West Africa trade potential has been heightened by the establishment of several transshipment hub ports of shipping lines in the Mediterranean, such as Hanjin and Maersk Line’s terminals in Algeciras, ZIM line in Barcelona and COSCO in Genova, have been flagged as a shipping route with great prospects for the future. Southern Africa, a rising economic power in Africa, is supposed to have sufficient container shipping demand in the future resulting from the continued infrastructure construction. Besides, due to the rising Suez Canal transit fee and Somali piracy threat in recent years (Fu et al., 2010), more shipping carriers are devoting themselves to open up alternative routes of the Suez Canal routes (e.g. the Arctic routes and the Cape routes). As such, it may bring opportunities for some Southern African ports to be developed into transshipment hubs, e.g. port of Durban, East London, Elizabeth and Cape Town. Shipping market in West Asia has been growing rapidly since the twenty-first century. During the period from 2001 to 2012, container traffic on West Asia–East Asia shipping route achieved a growth rate of 181%, i.e. 466293TEU and 1309400TEU in 2001 and 2012 respectively; West Asia–South Asia achieved a growth rate of 275%, i.e. 178585 TEU and 668845 TEU in 2001 and 2012 respectively. The increasing economic cooperation among West Asia, South Asia and East Asia, e.g. the Silk Road Economic Belt strategy of China, will lead to their closer trade connections that will further boom the shipping market in West Asia. However, the betweenness values of these emerging regions had not increased much, which are still much lower than the established intermediate regions like Europe Mediterranean. This proves that the rapid traffic growth of these regions is mainly contributed by their trade development with other regions rather than transshipment traffic. Therefore, they remain trade regions that could hardly play intermediate roles for inter-regional liner shipping in the GSN, and their positions in the structure of the GSN have not gone up as much as their total traffic volumes.

7. Research implications and conclusions This paper investigated the evolution of regional inequality in the GSN during the period from 2001 to 2012 by measuring and mapping the position changes of the world regions based on inter-regional linkages of liner shipping activities. The core concept of this paper is that the positions of world regions within the structure of the GSN should be determined by comparisons with each other, which requires analysis based on inter-regional interactions and thus, could not be achieved separately. The scholarly contribution of this paper lies mainly in the theoretical aspect. First, with a renewed interest in regional perspective analysis focusing on inter-regional liner shipping connections, this paper proposed a synthetic methodology to analyze the positions of world regions in the GSN, which consists of analysis of traffic development, centrality, dominance and vulnerability. Second, by a formal investigation into the position changes of world regions within the context of the GSN during the research period, total traffic volume, a widely used indicator, was proved to be inaccurate in reflecting the actual regional development in terms of maritime shipping. Therefore, this paper contributes to existing literature in which the evolution of the positions of world regions in the GSN is seriously under researched. Major findings are as follows: In spite of the superiority of East Asia over the other regions in terms of the total traffic volume, traffic growth of East Asia does not imply an equivalent improvement in its position in the GSN; in fact, results of this research evidence that the latter lags behind the former. First of all, the centrality of East Asia in the structure of the GSN, measured by degree and betweenness indicators, had always been lower than that of Northwest Europe and Europe Mediterranean before 2010. Secondly, considering the low betweenness values of East Asia during the studied period, the rapid container traffic growth of this region over the last decade is primarily contributed by the trade development of some countries within the region (e.g. China, the Philippines and Malaysia), rather than the increase in the transshipment traffic. That is, rapid traffic growth of many East Asian ports (i.e. mainly the Chinese coastal ports) in the last decade did not really improve the intermediate function of East Asia within the structure of the GSN, thus such function is very likely still restricted to several established transfer hubs such as Hong Kong and Busan; with an unequal intermediate function as compared to the total traffic volume, East Asia remains a powerful trade region rather than an intermediary in the GSN. Nevertheless, East Asia has occupied a significant position in the GSN due to its wide influence on world regions as is revealed by dominant flow analysis in this paper: The number of largest inflows rose from 9 in 2001 to 13 in 2012, i.e. except for South American North Coast, North American Gulf Coast and North Africa, the largest flows of the other regions were shared with East Asia in 2012. As global trade and containerization evolved during the studied period, some trade regions had seen their positions in the GSN obviously changed, while the positions of those intermediate regions were relatively stable. North American West Coast, North American East Coast and Australasia are the three big trade regions whose positions declined the most. This can be seen from their low traffic growth rates for the studied period, the decrease of degree and betweenness values, as well as the increasing dependency on East Asia. A few emerging regions that once in the margin are attracting rising attention from the world, i.e. South American North Coast, West Africa, Southern Africa, South American East coast and West Asia. This can be proved by their highest growth rates in connectivity and total traffic volume led by their rapid trade development; their traffic growth rates for the studied period rank the top five, i.e. 430.13%, 201.79%, 167.29%, 161.48% and

11

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12 Table A.1 Container capacity of the top ten shipping routes in TEU. 2001

2004

East Asia–North American West coast East Asia–Northwest Europe

1385640 1364347

East Asia–South Asia East Asia–Europe Med

1064996 1054766

Europe Med–Northwest Europe

751468

East Asia–North American East Coast North American East coat– Northwest Europe North American West coat– Northwest Europe East Asia–West Asia

605612

Caribbean/Central America– North American East coast

517085 469887 466293 418841

East Asia–North American West coast East Asia–Northwest Europe East Asia–Europe Med East Asia–South Asia Europe Med–Northwest Europe East Asia–North American East Coast East Asia–West Asia East Asia–Caribbean/ Central America West Asia–Europe Med North American East coat–Northwest Europe

2008

2012

1526992

East Asia–Northwest Europe

2239923

1362234

1867483

1293357 962254

East Asia–North American West coast East Asia–Europe Med East Asia–South Asia

706446

Europe Med–Northwest Europe

1390996

677362

East Asia–West Asia

979784

629600

East Asia–Caribbean/Central America Caribbean/Central America– North American East coast East Asia–North American East Coast West Asia–Europe Med

838564

490644 477578 476952

1753220 1578320

768320 638855 633442

East Asia–Northwest Europe East Asia–North American West coast East Asia–West Asia East Asia–Europe Med East Asia–South Asia

2774807 1622901 1309400 1163146 1077911

East Asia–North American East coast West Asia–Europe Med West Asia–South Asia

885238

South Asia– Northwest Europe East Asia–Caribbean/ Central America

650494

670357 668845

608456

Note: Europe Med = Europe Mediterranean. Source: Calculated by authors based on CI-Online data.

132.26% respectively. Due to the widely direct connections with all the world regions, Northwest Europe and Europe Mediterranean maintained their relatively intermediate function in the GSN. Although the total traffic volume of Northwest Europe and Europe Mediterranean were much lower than that of East Asia, their individual positions in the structure of the GSN were not far behind East Asia. The reasons for the different evolution processes of the above regions are as follows. The total traffic volume of a trade region is very susceptible to the economic context and trade situation of the region itself, thus the positions of such trade regions in the GSN are much volatile. On the contrary, with many transfer hubs facilitating liner shipping between some geographically remote regions, it is relatively capable for the established intermediate regions (e.g. Europe Mediterranean) to maintain their positions in the GSN. This is mainly because of the efforts that their hub ports have committed to keep the advantages (e.g. improving container handling efficiency and reducing congestion), and also the memory effect of shipping lines in the port selection process. In summary, this article is exploratory that only focuses on the uneven evolution processes of the world regions in the GSN. The major limitation of this research is that container flow directions among these regions are not considered. Further studies are needed to reveal the driving forces behind such processes, and also to optimize the distribution of world container capacities among these regions for better serving the world trade and the common prosperity of all the regions. Acknowledgements Authors are grateful for the useful advices from the associate editor Peter Hall and the anonymous reviewers, which greatly improved this study at the revision stage. We also appreciate for the contribution of Werikhe Gerald Wanzala and Tetteh Evans Ago to improve the English in this paper. This research benefited from the financial support of National Natural Science Foundation of China (Grant Number: 61174166) and National Social Science Foundation of China (Grant Number: 13&ZD170). Appendix A See Table A.1.

References Clarkson Research Services, 2014. Container Intelligence Quarterly: First Quarter 2014. Clarkson Research Services, London. Degerlund, J. (Ed.), 2006. Containerisation International Yearbook. Informa Group, London. Ducruet, C., 2012. The polarization of global container flows by interoceanic canals. In: Interoceanic Canals and World Seaborne Trade: Past, Present and Future, Bruxelles, June 7–8. Ducruet, C., Lugo, I., 2013. Structure and dynamics of transportation networks: models, concepts, and applications. In: Rodrigue, J.P., Notteboom, T.E., Shaw, J. (Eds.), The SAGE Handbook of Transport Studies. SAGE Publications, pp. 347– 364. Ducruet, C., Notteboom, T., 2012. The worldwide maritime network of container shipping: spatial structure and regional dynamics. Global Networks 12 (3), 395– 423. http://dx.doi.org/10.1111/j.1471-0374.2012.00352.x. Ducruet, C., Zaidi, F., 2012. Maritime constellations: a complex network approach to shipping and ports. Marit. Policy Manage.: Flagship J. Int. Ship. Port Res. 39 (2), 151–168. http://dx.doi.org/10.1080/03088839.2011.650718. Ducruet, C., Lee, S.W., Ng, A.K.Y., 2010a. Centrality and vulnerability in liner shipping networks: revisiting the Northeast Asian port hierarchy. Marit. Policy Manage.: Flagship J. Int. Ship. Port Res. 37 (1), 17–36. http://dx.doi.org/10.1080/ 03088830903461175. Ducruet, C., Rozenblat, C., Zaidi, F., 2010b. Ports in multi-level maritime networks: evidence from the Atlantic (1996–2006). J. Transp. Geogr. 18 (4), 508–518. http://dx.doi.org/10.1016/j.jtrangeo.2010.03.005. Fleming, D.K., 1997. World container port rankings. Marit. Policy Manage.: Flagship J. Int. Ship. Port Res. 24 (2), 175–181. http://dx.doi.org/10.1080/ 03088839700000068. Fleming, D.K., Hayuth, Y., 1994. Spatial characteristics of transportation hubs: centrality and intermediacy. J. Transp. Geogr. 2 (1), 3–18. http://dx.doi.org/ 10.1016/0966-6923(94)90030-2. Freeman, L.C., 1979. Centrality in Social networks conceptual clarification. Social Networks 1 (3), 215–239. http://dx.doi.org/10.1016/0378-8733(78)90021-7. Fremont, A., 2007. Global maritime networks: the case of Maersk. J. Transp. Geogr. 15 (6), 431–442. http://dx.doi.org/10.1016/j.trangeo.2007.01.005. Fremont, A., Soppe, M., 2004. The Evolution of North-European Shipping Networks: from Inter-Continental Links to a Global System, 1990–2000. In: World Conference on Transport Research, Istanbul, July 4–8. Fu, X., Ng, A.K.Y., Lau, Y., 2010. The impacts of maritime piracy on global economic development: the case of Somalia. Marit. Policy Manage.: Flagship J. Int. Ship. Port Res. 37 (7), 677–697. http://dx.doi.org/10.1080/03088839.2010.524736. Gonzalez, F., Freire, M.J., Pais, C., 2012. Maritime degree centrality and vulnerability: port hierarchies and emerging areas in containerized transport (2008–2010). J. Transp. Geogr. 24 (2012), 33–44. http://dx.doi.org/10.1016/ j.jtrangeo.2012.06.005. Gouvernal, E., Rodrigue, J.P., Slack, B., 2012. The divergence of regionalization: the challenges of the Mediterranean ports of Europe. In: IAME CONFERENCE, Taipei, September 6–8. Guerrero, D., Rodrigue, J.P., 2014. The waves of containerization: shifts in global maritime transportation. J. Transp. Geogr. 34, 151–164. http://dx.doi.org/ 10.1016/j.jtrangeo.2013.12.003. Guy, E., 2003. Shipping line networks and the integration of South America trades. Marit. Policy Manage.: Flagship J. Int. Ship. Port Res. 30 (3), 231–242. http:// dx.doi.org/10.1080/0308883032000113271.

12

M. Xu et al. / Journal of Transport Geography 44 (2015) 1–12

Hu, Y.H., Zhu, D.L., 2009. Empirical analysis of the worldwide maritime transportation network. Physica A 388 (10), 2061–2071. http://dx.doi.org/ 10.1016/j.physa.2008.12.016. Kaluza, P., Kölzsch, A., Gastner, M.T., Blasius, B., 2010. The complex network of global cargo ship movements. J. Roy. Soc. Interf. 7 (Suppl_2), 1093–1103. http:// dx.doi.org/10.1098/rsif.2009.0495. Li, K.X., Luo, M., Yang, J., 2012. Container port systems in China and the USA: a comparative study. Marit. Policy Manage.: Flagship J. Int. Ship. Port Res. 39 (5), 461–478. http://dx.doi.org/10.1080/03088839.2012.705032. Li, Z., Xu, M., Shi, Y., 2014. Centrality in global shipping network basing on worldwide shipping areas. GeoJournal. http://dx.doi.org/10.1007/s10708-0149524-3. Low, J.M.W., Lam, S.W., Tang, L.C., 2009. Assessment of hub status among Asian ports from a network perspective. Transport. Res. Part A: Policy Practice 43 (6), 593–606. http://dx.doi.org/10.1016/j.tra.2009.04.004. McCalla, R., Slack, B., Comtois, C., 2005. The Caribbean basin: adjusting to global trends in containerization. Marit. Policy Manage: Flagship J. Int. Ship. Port Res. 32 (3), 245–261. http://dx.doi.org/10.1080/03088830500139729. Mu, X.W., Chen, Y., Yang, M., Li, T., 2009. Topological features of liner shipping network. J. Dalian Marit. Univ. 35 (2), 34–37 (in Chinese). Notteboom, T., 1997. Concentration and load centre development in the European container port system. J. Transp. Geogr. 5 (2), 99–115. http://dx.doi.org/ 10.1016/S0966-6923(96)00072-5. Notteboom, T., 2006. Traffic inequality in seaport systems revisited. J. Transp. Geogr. 14 (2), 95–108. http://dx.doi.org/10.1016/j.jtrangeo.2004.12.003. Notteboom, T., 2010. Concentration and the formation of multi-port gateway regions in the European container port system: an update. J. Transp. Geogr. 18 (4), 567–583. http://dx.doi.org/10.1016/j.jtrangeo.2010.03.003. Notteboom, T., 2012. Towards a new intermediate hub region in container shipping? Relay and interlining via the Cape route vs. the Suez route. J. Transp. Geogr. 22, 164–178. http://dx.doi.org/10.1016/j.jtrangeo.2012.01.003.

Nystuen, J.D., Dacey, M.F., 1961. A graph theory interpretation of nodal regions. Papers Reg. Sci. Assoc. 7 (1), 29–42. http://dx.doi.org/10.1007/BF01969070. Rodrigue, J.P., 2012. The Benefits of Logistics Investments: Opportunities for Latin America and the Caribbean. Inter-American Development Bank, Washington, DC. UNCTAD, 2012. Review of Maritime Transportation. UNCTAD, Geneva/New York. Valentine, V.F., Benamara, H., Hoffmann, J., 2013. Maritime transport and international seaborne trade. Marit. Policy Manage.: Flagship J. Int. Ship. Port Res. 40 (3), 226–242. http://dx.doi.org/10.1080/03088839.2013.782964. Veenstra, A.W., Mulder, H.M., Sels, R.A., 2005. Analysing container flows in the Caribbean. J. Transp. Geogr. 13 (4), 295–305. http://dx.doi.org/10.1016/ j.jtrangeo.2004.07.006. Verhoeff, J.M., 1977. Seaport competition: an analysis of its nature. Int. J. Transp. Econ. 4 (3), 271–284. Wang, Y.H., Cullinane, K., 2014. Traffic consolidation in East Asian container ports: a network flow analysis. Transp. Res. Part A 61, 152–163. http://dx.doi.org/ 10.1016/j.tra.2014.01.007. Wilmsmeier, G., Notteboom, T., 2011. Determinants of liner shipping network configuration: a two-region comparison. GeoJournal 76 (3), 213–228. http://dx.doi.org/10.1007/sl0708-009-9333-2. Woolley-Meza, O., Thiemann, C., Grady, D., Lee, J.J., Seebens, H., Blasius, B., Brockmann, D., 2011. Complexity in human transportation networks: a comparative analysis of worldwide air transportation and global cargo-ship movements. Europ. Phys. J. B 84 (4), 589–600. http://dx.doi.org/10.1140/epjb/ e2011-20208-9. WTO, 2008. World Trade Report 2008 – Trade in a Globalizing World. Technical Report, World Trade Organization. Yap, W.Y., Lam, J.S.L., Notteboom, T., 2006. Developments in container port competition in East Asia. Transport Rev. 26 (2), 167–188. http://dx.doi.org/ 10.1080/01441640500271117.