Port Systems in Global Competition: Spatial-economic Perspectives on the Co-development of Seaports 1003316654, 9781003316657

Citation preview

Port Systems in Global Competition

In a world where most international trade is carried by sea, each port can be seen as a unique chokepoint competing to attract ever more traffic and economic activities. However, ports can also be seen as parts of a wider system, which can be defined as a system of two or more ports located in proximity within a given area. Their fate and governance is jointly inf luenced when belonging to the same region, country, or transnational space. Investments, shocks, innovations, and delays occurring in one port often affect other ports within a certain spatial range and time lapse. Further understanding of such co-developments in port systems is necessary to go beyond local specificities, through a multidisciplinary and multi-level contribution. Port Systems in Global Competition is an answer to the strong and urgent need for reviewing the relevant theories, concepts, methods, and sources that can be mobilized for the analysis of port systems. With contributions from reputable scholars coming from no less than 11 countries in Europe, Asia, and North America, this book delves into the analysis of port systems from diverse disciplinary angles (geography, regional science, economics, management, engineering, and mathematics/computer sciences), covering innovative empirical approaches to various port systems in the world. The theoretical and empirical knowledge can support and enhance decision-making in relation with the development of ports, supply chains, and transport networks in general. This book is an ideal companion to academics and upper-level students interested in the analysis of transport and economic systems in general, as well as the effective ways to answer complex issues in transportation and socio-economic development. It will be a valuable resource for those researching or studying transportation and supply chains, maritime and port economics, as well as regional development and human geography. César Ducruet, geographer, is Senior Researcher at the French National Centre for Scientific Research (CNRS). He is currently working at the EconomiX laboratory (Paris-Nanterre) on the local impact of contemporary maritime globalization. His research focuses on technological innovation, connectivity, employment, vulnerability, environment, and health issues in a

port and port-city context. He is Principal Investigator of the ANR-funded research project “Maritime Globalization, Network Externalities and Transport Impacts on Cities” (MAGNETICS) (2023–2026). César has been expert for various international organizations (OECD, World Bank, WHO) and works regularly with numerous partners in Asia (Korea Maritime Institute, JETRO, ASEM, Chinese Academy of Sciences, ECNU, Fudan University, Shanghai Maritime University). His publications include two edited volumes on Maritime Networks (2015) and Shipping Data Analysis (2017) in the Routledge Studies in Transport Analysis. He is also associate member of porteconomics.eu, scientific board member of SFLOG, GIS Axe Seine, GDR OMER, RETE Association, international advisory board member of PortCityFutures, and editorial board member of the Journal of Transport Geography, Maritime Business Review, the International Journal of Transport Economics, and Portus. Theo Notteboom is Professor of port and maritime economics. He is Chair Professor ‘North Sea Port’ at Maritime Institute of Ghent University, and a professor at the Faculty of Business and Economics of the University of Antwerp and Antwerp Maritime Academy. He previously held positions as Professor and Foreign Expert at universities in Dalian and Shanghai, China, and as MPA Visiting Professor at Nanyang Technological University in Singapore. He is Vice-President (2022-ongoing) and past President (2010– 2014) of International Association of Maritime Economists (IAME). He is Co-founder and Co-director of Porteconomics.eu and Member of the Risk and Resilience Committee of International Association of Ports and Harbors (IAPH). He is Associate Editor of Maritime Economics & Logistics and a member of the editorial boards of eight other leading academic journals in the field. He published over 160 papers in first-tier academic journals and another 400 publications in the form of reports and contributions to books, proceedings, and specialized press. He is Editor/Author of a dozen of academic books, including the handbook Port Economics, Management and Policy (Notteboom, Pallis & Rodrigue, 2022; Routledge). He is one of the most cited maritime economists in the world. Theo Notteboom has been involved as promoter or co-promoter in more than 100 academic research programs on the port and maritime industry and logistics topics.

Routledge Studies in Transport Analysis

11

U.S. Freight Rail Economics and Policy Are We on the Right Track? Edited by Jeffrey T. Macher and John W. Mayo

12

Arctic Shipping Climate Change, Commercial Traffic and Port Development Edited by Frédéric Lasserre and Olivier Faury

13

Parking Regulation and Management The Emerging Tool for a Sustainable City Edited by Daniel Albalate and Albert Gragera

14

Human Resource Management in Shipping Issues, Challenges and Solutions Lijun Tang and Peng fei Zhang

15 Container Logistics and Maritime Transport Dong-Ping Song 16

Logistics, Transport and the COVID-19 Crisis Managing and Operating Logistics Processes Edited by Jacek Woźniak, Wioletta Sylwia Wereda and Bogdan Nogalski

17

Port Systems in Global Competition Spatial-Economic Perspectives on the Co-Development of Seaports Edited by César Ducruet and Theo Notteboom

For more information about this series, please visit: www.routledge.com/ Routledge-Studies-in-Transport-Analysis/book-series/RSTA

Port Systems in Global Competition Spatial-Economic Perspectives on the Co-Development of Seaports Edited by César Ducruet and Theo Notteboom

First published 2024 by Routledge 4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Third Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 selection and editorial matter, César Ducruet and Theo Notteboom individual chapters, the contributors The right of César Ducruet and Theo Notteboom to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-032-32773-0 (hbk) ISBN: 978-1-032-32777-8 (pbk) ISBN: 978-1-003-31665-7 (ebk) DOI: 10.4324/9781003316657 Typeset in Bembo by codeMantra

Contents

List of figures List of tables List of contributors Foreword

xi xv xvii xxvii

J AC Q U E S C H A R L I E R



PART I

The concept of port system

11

1

13

A systematic and critical review of port system research C É S A R D U C RU E T A N D T H E O N O T T E B O O M

2

Evolutionary models of port system development – an application to the Latin American and Caribbean port system

55

GOR DON W I LMSM EI ER A N D JA SON MON IOS

3

Winding paths through urban systems and urban networks

77

BEN JA MIN PR EIS

4

The implications of duality of trans(port) systems: Evidence from Wusongkou International Cruise Port J A M E S J . WA N G , A D O L F K .Y. N G , A N D Y U I -Y I P L AU

89

viii Contents PART II

The dynamics of port systems

101









PART III

Collaborative port systems

181



10 Cooperation and competition between container shipping networks and their impact on container hub ports in Southeast Asia

203

W E I -Y I M YA P

11 Port collaboration in the Greater Bay Area: Reality, challenge, and opportunity

223

D O N G YA N G , YA N G C H E N , A N D Q I A N G Z H A N G

12 Analyzing the US Gulf Coast petrochemical port polycentric region, and its connectivity with the European Amsterdam-Rotterdam-Antwerp (ARA) region K A R E L VA N D E N B E RG H E , A N T O I N E P E R I S , A N D WO U T E R J AC O B S

245

Contents  ix PART IV

Port systems as shipping networks

263

13 Discovering shipping networks from raw vessel movement

265

A L E X A N D RO S T RO U P I O T I S - K A P E L I A R I S , G I A N N I S  S P I L I O P O U L O S , M A R I O S VO DA S , A N D D I M I T R I S Z I S S I S

14 Ocean container network dynamics during the COVID-19 pandemic

283

C H R I S T O P H E R D I R Z K A A N D M I C H E L E AC C I A RO

15 Shipping network analysis: State-of-the-art and application to the global financial crisis

300

C É S A R D U C RU E T

Index

335

Figures

1.1 1.2 1.3 1.4 1.5 1.6 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 4.1 4.2 4.3 5.1 5.2 A5.1 6.1

Publication trend of port system articles, 1963–2022 Distribution of articles by journal background, 1963–2022 Distribution of articles by level of analysis, 1963–2022 Collaboration dynamics, 1963–2022 The co-authorship network of port system studies A multilevel modeling of ports and port systems Generic four-phase model of the evolution of liner shipping networks and ports Comparison of container vessel size evolution on the WCSA, 2012–2021 Comparison of vessel size evolution in primary and secondary ports on the WCSA, 2012–2021 (left: Callao; right: Paita) Growth dynamics comparison of main ports along the ECSA port range, 1997–2021 Container throughput (TEU) along the ECSA port range, 1997–2019 Container throughput (TEU) at Chilean ports, 1997–2019 Container throughput and traffic shares along Pacific and Caribbean ranges (Mexico, Central America, and NCSA), 1997–2019 Container throughput at transshipment and hybrid ports, 1997–2019 The duality of transport system: an overall view The duality of transport system: typical relationships seen in classical, mobility-centered transport economics The duality of transport system: concerns in urban or regional planner’s and place-stakeholders’ mind Relative transition path by port range (total traffic) Port rankings in 2000 and 2019 by traffic type Ports mobility for total traffic between 2000 and 2019 Spatial processes of port migration and relocation of container terminals

17 17 25 32 33 35 59 61 62 63 64 67 68 69 91 94 94 109 112 121 124

xii Figures

6.2 Governance settings of port migration and relocation of container terminals 6.3 London’s inner and outer vessel traffic evolution 6.4 Correlation between urban population and vessel traffic by type, 1880–2010 6.5 Outer traffic share by main cargo type, 1977–2008 6.6 Outer traffic share by world region, 1950–2010 6.7 Outer traffic share by location type, 1950–2008 6.8 Morphologies of port cities 6.9 Typology of outer port traffic trajectories 7.1 Metrans network of rail services 7.2 Visualization of the main country-port pairs (97% of the total) 7.3 Preferential relationships 7.4 Barrier effects 9.1 Maritime traffic handled by ports in the Rhine-Scheldt Delta and economic importance in terms of employment and value added 10.1 Depiction of shipping services that operate on the Asia-Europe trade route pertaining to the ports of Singapore, Port Klang, and Tanjung Pelepas 10.2 Research framework for analyzing container shipping networks on the Asia-Europe trade route and their impact on container hub ports in Southeast Asia 10.3 Container shipping networks in 2011 for the Asia-Europe trade route involving the three container hub ports in Southeast Asia 10.4 Container shipping networks in 2017 for the Asia-Europe trade route involving the three container hub ports in Southeast Asia 10.5 Container shipping networks in 2020 for the Asia-Europe trade route involving the three container hub ports in Southeast Asia 11.1 Location of the GBA multi-port system 11.2 Port collaboration analysis framework 11.3 Port operators in Hong Kong port 11.4 Port collaboration map in the GBA 11.5 Ownership structure of ports in the GBA 11.6 Evolution of port throughput (TEU), 2015–2020 11.7 Evolution of number of port calls, 2015–2020 11.8 Average size of ships calling at the port (TEU), 2015–2020 11.9 Average port calls by routes, 2017–2020 11.10 Ratio of foreign ports to domestic ports of origins (FDO) and destinations (FDD), 2015–2020 12.1 Overview of the US Gulf Coast and ARA polycentric port regions

124 137 137 138 139 140 140 141 151 154 158 158 186 210 211 213 213 214 224 225 229 230 233 234 234 235 235 236 246

Figures  xiii

12.2 Morphological and functional monocentric and polycentric systems 12.3 The emergence of a higher-level polycentric system out of the interactions between its lower-level constituting dimensions 12.4 The USGC with the different ports and their refineries, terminals, and chemical plants 12.5 Relations between oil tanking terminals, refineries, and chemical plants in the USGC according to the analysis of the movement of tankers 12.6 Relations ports of the USGC Region according to the analysis of the movement of tankers 13.1 The workf low of our approach: raw AIS messages are filtered and then fed into a mechanism that generates a grid for the selected area and calculates the vessels density (tonnage), producing a density map 13.2 Noisy vessel trajectory. The first four positions of the vessel follow a normal path, but the fifth (t4) is clearly an erroneous message. By calculating the estimated speed required to travel this distance (t3 to t4) we are able to eliminate this message 13.3 Geometry of the Aegean and Levantine Seas 13.4 Grids for the area around Cyprus, with grid cell lengths 100 km (left) and 10 km. Smaller lengths provide more detailed results 13.5 Density maps for the region of Cyprus, showing the total time fishing (left) and passenger (right) vessels spent within the cells of a 1 km grid 13.6 Example trips between Palermo and Civitavecchia for determining the total crossings; each time a vessel enters a cell then its density is increased 13.7 Color ramp for density map designed to highlight major routes by including only the top 5% of density measurements as the darkest category 13.8 Density maps for cargo vessels at European waters for the month of November 2021 (total amount of time spent – left, vessel tonnage – right) with shipping routes more easily identified using the proposed metric 13.9 Density maps for tanker vessels at European waters for the month of November 2021 (total amount of time spent – left, vessel tonnage – right) 13.10 Density maps depicting shipping activities for the Aegean and Adriatic seas from November 2021 (Cargo vessels – left, Tanker vessels – right)

249 250 252 254 254

267

269 270 273 274 275 276

277 277 277

xiv Figures

13.11 Density maps depicting shipping activities at the Black Sea for the month of November 2021 (Cargo vessels – left, Tanker vessels – right) 13.12 Top 100 busiest European ports for the month of November 2021, with the size of each port representing the total amount of cargo transferred at it, considering all types of vessels. The corresponding density maps for the shipping activity of vessels is included in gray 13.13 Detailed maps considering the total amount of cargo transferred toward main European ports, considering vessel movement from all vessels for the month of November 2021 (North Europe – top, North-central Europe – bottom) 14.1 Ocean shipping network dynamics, January 2020– February 2022 14.2 Public notices by liner operators, January 2020– December 2021 14.3 Paradigm shift – Public notices, port congestions and freight market 15.1 Publication trend about shipping networks, 2007–2022 15.2 Co-authorship patterns in shipping network studies 15.3 Relationship between centrality size and growth at world ports, 2007–2010 15.4 Daily containership traffic at the port of New York, January 2008–November 2010 15.5 Hub structure and centrality changes in the global shipping network, 2008–2009

278

278

279 289 293 294 302 303 311 316 316

Tables

1.1 Top 20 journals by the number of publications about port systems 18 1.2 Port system concentration and deconcentration factors, 1963–2022 20 1.3 Cooperation projects on sustainability among European ports, 2010–2023 30 2.1 Drivers of container port system evolution in Latin America and the Caribbean, 1997–2019 73 108 5.1 Gross weight (106) of commodities traffic in European ports 5.2 Rank-size model for container port systems 110 5.3 The quadratique model 111 5.4 Distribution of European ports in 2000 (by cargo) 113 5.5 Transition matrix 114 5.6 Mobility measures 115 5.7 Initial (2000) and ergodic distribution of European ports 115 7.1 Containerized exports of the countries in Central Europe to the US. Breakdown by port and country 153 7.2 Share of containerized exports of Central European countries to the U.S., broken down by forwarding port 155 7.3 Containerized exports of Central European countries to the U.S., Breakdown by type of commodity 156 7.4 Results of the production constrained spatial interaction model for all commodity types 157 7.5 Results of production constrained models by commodity type 160 7.6 Location coefficients of ports on commodity types 161 8.1 Global Logistics Index – All GLRs 169 8.2 Global Logistics Index for top ten ranked GLRs 170 8.3 Shares of global container movements: top ten ranked GLRs 170 8.4 Shares of air cargo movement: top ten ranked GLRs 170 8.5 Comparison of results, 1996–2006 and 2006–2020 171 A8.1 Container ports and airports in GLRs 177 A8.2 Global logistics index for GLRs, 2006 & 2020 180

xvi Tables



Contributors

Michele Acciaro is Associate Professor at the Department of Strategy and Innovation of the Copenhagen Business School. His main research interests are sustainability, the energy transition, global logistics, maritime policy, and climate change. Until December 2021, he was Associate Professor at the Department of Operations and Technology and the Director of the Hapag-Lloyd Center for Shipping and Global Logistics (CSGL) at the Kühne Logistics University (KLU) in Hamburg. He joined KLU in 2013 as Assistant Professor and between 2016 and 2018 he was also Head of the Logistics Department. Before accepting the appointment at KLU, Michele held the position of Senior Researcher – Green Shipping in the Research and Innovation Department of Det Norske Veritas AS (DNV) in Høvik, Norway. From 2004 until 2010, he was Deputy Director and Researcher at the Erasmus University Rotterdam Center for Maritime Economics & Logistics (MEL)/Erasmus SmartPort. Michele has a BSc and an MSc cum Laude in Statistics and Economics from University of Rome “ La Sapienza”, an MSc in Economics from Erasmus University Rotterdam (with a focus on Maritime Economics & Logistics, for which he was awarded the NOL/APL Prize for Student Excellence), and a PhD in Logistics from Erasmus University Rotterdam. Michele was awarded the Young Researcher Best Paper Prize at the IAME Annual Conference in Cyprus in 2005 and the Best Reviewer Prize at IAME 2018 in Mombasa. Jacques Charlier  is since 2015 an emeritus professor of geography at the School of Geography of the UCLouvain in Louvain-la-Neuve, Belgium. He spent most of its academic career there (including as a researcher for the Belgian National Fund for Scientific Research), but he was also an associate and visiting professor at the Paris-Sorbonne university. He is a member of the (Belgian) Royal Academy for Overseas Sciences, the president of which he was in 2000 and 2012. He published extensively (mostly in French) about maritime transport and port geography, with a focus on Western Europe, Canada, and Sub-Saharan Africa, and he has a special interest for the container and cruise shipping industries as well as for interoceanic canals. His PhD thesis (in 1981) was about the hinterlands

xviii Contributors

of French seaports, and he is convinced since these days that a given port should be considered a component of several (sub)systems, horizontal (including the neighboring ports), vertical (including hinterlands and forelands), and functional (at terminal level, for each market segment). Yang Chen is Professor of Strategy and Innovation and Associate Dean of College of Transport and Communications at Shanghai Maritime University, China. He obtained his PhD degree from Fudan University and was once a visiting scholar at Harvard Business School. His research interest covers strategy, digitization, and innovation in shipping industry. Gabriel Figueiredo De Oliveira is a senior lecturer at the University of Toulon (France) since 2012. The main fields of his research are port and maritime economics and international trade. During his PhD, He spent 9 months at the World Maritime University (WMU) to improve his specialization in Maritime Economics. His thesis on the determinants maritime transport costs and their impact on trade was defended in 2011 and he has since published 14 scientific articles. Rania Tassadit Dial currently  is Teaching Assistant at the University of Rennes 1 (France). Her primary filed of research is port growth and regional development. In terms of methodology, she mainly uses spatial econometrics to understand how ports interact with their land and maritime environments in a context of inter-port competition. Rania is also interested in port concentration and hierarchy issues, mainly in Europe. Christopher Dirzka is a PhD student at the Department of Operations and Technology of the KLU. His main research interests are network designs, liner operations, sustainability, and anti-trust policy. Christopher joined the KLU in September 2018 as a PhD candidate under Prof. Dr. Michele Acciaro. He graduated with an MSc in Supply Chain Management from Copenhagen Business School in Copenhagen, Denmark, and a BSc in Economics and Business Administration from Aarhus University. Prior to joining the KLU, he gained professional experience as a junior analyst for the Baltic and International Maritime Council (BIMCO) in Bagsværd, Denmark, and as a research analyst for Mærsk Broker AS in Copenhagen, Denmark. David Guerrero  (PhD Univ Paris 7, 2010) has been Researcher at Gustave Eiffel University in Paris since 2011. Before working in academia, he has some operational experience in Public Policy as Project Manager for a French Prime Minister’s Delegation (Datar) in 2010–2011, where he contributed to the writing of a report for the parliament. As a transportation geographer, his particular areas of expertise include ports and maritime transport, global supply chains, and freight urban systems. David teaches transport geography and maritime transport at several universities in France and abroad, and has been invited Professor at Kwansei Gakuin

Contributors  xix

University ( Japan) in two occasions (2017 and 2019). In the last years, David has contributed to several research projects related to ports and maritime transport. The results of these works have been published in journals such as the Journal of Transport Geography, Regional Studies, or Transport Policy. Wouter Jacobs is the founder and director of the Erasmus Commodity & Trade Centre at Erasmus University Rotterdam, a dedicated thinktank supported by industry (launch Q3 2022). He initiated and leads the executive program Leadership in Commodity Trade and Supply Networks at Erasmus University Rotterdam. Wouter is a member of the advisory council of J.P. Morgan Centre for Commodities at the CU Denver Business School of the University of Colorado, and for over 8 years, he has been involved as Fellow at Erasmus UPT. Wouter has a PhD in Management from Radboud University Nijmegen, based on an international comparison of port competition strategies in Los Angeles-Long Beach, Rotterdam, and Dubai. Wouter has a specialization in teaching scenario planning methodologies, amongst others for the MEL program, which result in strategic options for industry and governments. He is also the program director at Erasmus University for the Maritime Studies Abroad program of Nanyang Technological University from Singapore. Yui-yip Lau is working at Division of Business and Hospitality Management, College of Professional and Continuing Education, The Hong Kong Polytechnic University. Until now, he has published more than 240 research papers in international journals and professional magazines, contributed 10 book chapters and 2 books, and presented numerous papers in international conferences. He has collaborated with scholars from more than 20 countries and regions spreading over five continents on research projects. He has also secured over HK$10 million research grants. Recently, he has been awarded a Certificate of Appreciation by the Institute of Sea Transport in recognition of his outstanding performance on research and the Best Paper Award in international leading conferences. His research interests are cruise, ferry, maritime transport, air transport, impacts of climate change, maritime education and training, transport history, sustainability issues, supply chain management, health logistics, and regional development. Jason Monios is Professor of Maritime Logistics at Kedge Business School, Marseille, France. His research revolves around three key areas: maritime transport (port system evolution, collaboration and integration in port hinterlands, port and shipping governance and policy, institutional and regulatory settings), intermodal transport (corridors, dry ports, terminal development, business models and logistics strategies, also including urban logistics), and sustainability and environmental concerns (maritime sustainability, decarbonization and environmental policy, green ports,

xx Contributors

climate change adaptation, autonomous and electric vehicles). He has led numerous research projects on these topics with a total budget of over €1m and contributed to many others. He has over 100 peer-reviewed academic publications with over 2,500 citations. He has worked with national and regional transport authorities and authored technical reports with UNCTAD, UN-ECLAC, and the World Bank. Adolf K. Y. Ng  is the associate department head and professor at the Department of Management, Faculty of Business and Management, BNU-HKBU United International College, China. Obtaining DPhil from the University of Oxford, his research interests include port management, transport geography, supply chain resilience, climate adaptation planning, institutional and organizational change, and global value chains. His scholarly outputs include eight books, more than 100 papers in leading journals, and other forms of scholarly and professional publications. He is an associate editor of Maritime Policy and Management, the editor-in-chief of The Maritime Economist, and a founding member of the international consortium on resilient transport and supply chain networks “CCAPPTIA” (www.ccapptia.com). Kevin O’Connor is the emeritus professor of Urban Planning at the University of Melbourne. A PhD graduate of McMaster University, his research draws on a background in economic geography and regional science to explore links between selected elements of the economic system and the size and functions of cities. Recent research focuses on logistics services, beginning with the idea that they have elements in common with advanced services such as finance, law, and management. That commonality may inf luence their location in favor of major commercial centers (global cities), weakening their historical connection to the transport-oriented cities. Air transportation is a second research interest. Here the core issue is the way that travel markets, airlines, and aircraft technology interact with national regulation to shape the air service at a city. An historical perspective provided a framework to investigate how trade patterns, tourism f lows, and new long-haul smaller aircraft (some in the hands of low cost carriers) have led to changes in the air transport functions. Much of the research has been targeted at the Asia-Pacific region. Kevin O’Connor is an associate editor of the Journal of Transport Geography, jointly responsible for the Asia Pacific region, and also for papers on air transportation and logistics. When not engaged in those tasks he can be found hiking in the forests around Melbourne, reading spy and crime novels, and following the fortunes of the Liverpool Football Club. Antoine Peris  (PhD from Delft University of Technology) is an assistant professor in quantitative human geography at Avignon Université and a member of the CNRS Lab Espace. His research deals with the analysis of urban systems and their dynamics at different spatial scales, methods to

Contributors  xxi

deal with spatio-temporal data, f low, and network analysis. Antoine has published in academic peer-reviewed journals such as Computers, Environment and Urban Systems, Networks and Spatial Economics, Cybergeo, and Territory, Politics, and Governance. Benjamin Preis is a fourth-year PhD candidate in the MIT Department of Urban Studies and Planning. His dissertation research focuses on the relationship between the rental housing market and inequality in US cities. He is particularly concerned with how a changing rental market produces inequality and opportunities for extraction and exploitation. His other work has focused on gentrification, local government policy, and cybersecurity. Ben received his Master in City Planning from MIT in 2019. His master’s thesis focused on the relationship between cities and universities in the deployment of “smart city” technology. He has previously worked with the Nowak Metro Finance Lab at Drexel University, the Urban Redevelopment Authority of Pittsburgh, the Cambridge Housing Authority, and the Center for Public Studies in Santiago, Chile. Prior to his time at MIT, he worked on science and higher education policy at Lewis-Burke Associates in Washington, DC. Ben received his BS in Physics and Peace and Justice Studies from Tufts University. Alexandra Schaffar is Professor of Economics in the University of Toulon. She specializes in spatial and regional economics and econometrics. She is the head of the Institute of Development Economics of Toulon. She is also the vice-president of the French Speaking section of the International Regional Science Association. Brian Slack holds the title of Distinguished Professor Emeritus at Concordia University, Montreal, Canada. He has been recognized by the American Association of Geographers for his contributions to Transport Geography and received an Honorary Doctorate from the University of Le Havre. Throughout his career his research focused on maritime transport and ports, and has published extensively on topics including port competition, supply chains and intermodal transportation, maritime shipping networks, and port governance. He has served as External Examiner at the World Maritime University, Malmo, Sweden and has undertaken a wide range of consultancies in Canada, USA, France, and China. Giannis Spiliopoulos  holds a five-year diploma (2008) and an MSc (2010) from the Electronics and Computer Engineering Department at the Technical University of Create. He is a member of Technical Chamber of Greece since 2009, and he has worked as Software Engineer and Analyst in the private sector for six years. His academic interests focus on statistical analysis of spatial data and recently he has been working on extracting data-driven patterns using unsupervised methods on spatial and spatio-temporal data.

xxii Contributors

Jean-Claude Thill is Knight Distinguished Professor of Public Policy at the University of North Carolina at Charlotte, USA. His primary affiliation is with the Department of Geography and Earth Science, and he is also a faculty affiliate of the Public Policy Program and the School of Data Science. The study of interregional and international freight transportation systems has been one of his areas of research for more than 20 years. His interests include networked structures, the critical role of ports, and external barriers that shape hinterlands and containerized freight services. He has received a number of national and international awards and prizes for his scholarship. Alexandros Troupiotis-Kapeliaris received the BSc and MSc degrees from the Department of Informatics and Telecommunications, National and Kapodistrian University of Athens, in 2017 and 2020, respectively. From 2019 he works as an associate researcher for NCSR ‘Demokritos’; he has participated in a few European projects, including INFORE and VesselAI. His research interests include machine learning, big data, and streaming systems. He is currently pursuing a PhD degree with the Department of Product and Systems Design Engineering of the University of the Aegean. Karel Van den Berghe is currently Assistant Professor in Spatial Planning and (circular) Economy at the Faculty of Architecture and the Built Environment, Delft University of Technology, the Netherlands. Karel holds a master in Geography, a master in Spatial Planning and Urbanism, and a PhD in Spatial Planning and Economy (Ghent University, Belgium). His PhD “Planning the Port City” (2018) developed a relational perspective on port cities. More specifically, Karel’s articles analyzed if and how urban and maritime economies are related, and how in these local to global networks development emerges. His results have been used by the Flemish Government, and the Port Authorities of Ghent/North Sea Port and Amsterdam. After his PhD, Karel moved to Delft and has since then increasingly focused on the local-toglobal nexus, and in particular spatial policymaking as a (non-)driver of change. In particular, the circular economy is used as a perspective to understand how the changing global context (cf. slow/de-globalization) has an effect on concrete daily policymaking and, in turn, how far current “standard” (locational) private and public policymaking in port and cities is still appropriate in light of the changing global context. Karel has published in academic, professional, and newspaper journal. He regularly acts as expert for political parties, research institutes, municipalities, provinces, regional/national governments, and the European Commission. Next to his position in Delft, Karel is Lecturer at Erasmus University Rotterdam, part of the Leiden-Delft-Erasmus cooperation, and the Delft-Erasmus Convergence initiative. Finally, Karel is part of the advisory board of the Dutch organization of industrial areas.

Contributors  xxiii

Marios Vodas received his BSc and MSc degrees in Informatics from the University of Piraeus, Greece, in 2011 and 2013, respectively. He is currently a researcher in MarineTraffic. His research interests include indexing, distributed processing, and mining on spatio-temporal big data. James J. Wang,  former Head of the Department of Geography of the University of Hong Kong, is currently the research director of the Belt and Road Hong Kong Centre/Bay Area Hong Kong Centre, focusing on the regional development of Guangdong-Hong Kong-Macau Greater Bay Area and the Belt & Road Initiative (BRI). Professor Wang is a steering committee member of the Transportation Geography Committee of the International Geographical Union (IGU), fellow of the Hong Kong Institute of Transportation and Logistics (FCILT), editing board member of four top international journals in transport research, consultant to China Port City Study by World Bank, and invited expert by a few city governments, including Hong Kong SAR, Guangzhou, Zhuhai, and Shanghai, for their transport and urban planning projects. Professor Wang received a bachelor’s degree in economics from the Renmin University of China, a master’s degree in geography from the University of Hong Kong, and a PhD from the University of Toronto. As a transport geographer and port-city development specialist, he has published widely in internationally referred journals, as well as manuscripts including “Port-city Interlays in China” (2014, Routledge), “Ports, Cities, and Global Supply Chains” (2007, Ashgate, editor), “World-class Hub: Hong Kong’s External Transportation” (2019, Commercial Press), and provided his expertise in research or consulting works for more than 40 projects in various port cities in China and abroad. Gordon Wilmsmeier is Associate Professor for Shipping and Global Logistics and Director of the Hapag-Lloyd Centre for Shipping and Global Logistics at the Kühne Logistics University (KLU) in Hamburg, Germany. At the same time, he holds the Kühne Professorial Chair in Logistics at the Universidad de los Andes in Bogotá, Colombia. From 2011 to 2017, he worked as Economic Affairs Officer in the Infrastructure Services Unit at the United Nations Economic Commission for Latin America and the Caribbean (UN-ECLAC). Previously he worked at Edinburgh Napier University’s Transport Research Institute (TRI), and as Consultant for UN-ECLAC, UNCTAD, UN-OHRLLS, the World Bank, Adelphi Research, JICA, IDB, CAF, OAS. He is an internationally recognized expert in maritime transport geography and economics, port economics, and sustainable logistics with particular interests in shipping networks, governance, competition, transport costs, sustainability, energy efficiency, and nautical electromobility. Gordon is Honorary Professor for Maritime Geography at the University of Applied Sciences in Bremen, Germany. He has published over 100 book chapters, journal papers, institutional publications, and working papers. He is Leader of the global port

xxiv Contributors

performance research network (PPRN – https://pprn.network). Gordon is Vice-President of the International Association of Maritime Economists (since 2020) and has been Council Member 2010–2016 and since 2018. He is also an Associate Member of PortEconomics. Dong Yang is an associate professor of Department of Logistics and Maritime Studies, Deputy Director of Shipping Research Centre, at the Hong Kong Polytechnic University. He is Associate Editor of the International Journal of Shipping and Transport Logistics and Maritime Business Review and current Council Member of International Association of Maritime Economist (IAME) (2019 to now). Dr. YANG obtained his PhD majoring in maritime logistics science from Kobe University in 2008. After that, he has successively served as Assistant Professor at Southern University of Denmark (2008–2010), Research Fellow at the Centre for Maritime Studies, National University of Singapore (2010–2012), and Senior Research Fellow at China Waterborne Transport Research Institute (2013–2016). Wei Yim Yap is Associate Professor and Head of Maritime Management at the Singapore University of Social Sciences. In his career, he has worked closely with the international maritime, shipping, and port community to deal with various issues covering maritime cluster development, port development, port marketing, performance benchmarking, traffic analysis, competition studies, commercial feasibility studies, investment appraisal, scenario planning, economic impact analysis, and environment scans across different cargo types as well as the cruise business. He was former Head of Strategic Planning at the Maritime and Port Authority of Singapore and remains actively involved in the industry with government agencies and private sector entities. A/P Yap obtained his PhD in Maritime Economics from the University of Antwerp in 2009. His book entitled Business and Economics of Port Management: An Insider’s Perspective is a must read for anyone who wants to have an in-depth understanding in the port business. Qiang Zhang is an associate professor of College of Transport and Communications at Shanghai Maritime University (SMU). Dr. Zhang obtained his PhD degree from SMU. He was once a visiting scholar at Erasmus University Rotterdam (EUR) (2016–2017) and Transport & Environment (T&E) (2018). His research interests include shipping emission control, port governance, and port geography. Dimitris Zissis (Senior Member, IEEE) is currently an associate professor with the University of the Aegean. His published scientific work includes more than 70 publications, which have received more than 2200 citations to date. His research interests and areas of expertise include several aspects of architecting and developing complex Information Systems (IS), including distributed and cloud-based big data deployments. He is a member of the editorial boards of Future Generation Computer Systems

Contributors  xxv

(FGCS) (Elsevier) and the International Journal of Internet of Things and Cyber-Assurance ­ (Inderscience) and a PC member for numerous conferences, including CLOUDCOM, GECON, SerCO, and others. He is also a member of the IEEE Computer Society, the IEEE Oceanic Engineering, the IEEE Intelligent Transportation Systems Societies, and the Young Researchers Committee of the World Federation on Soft Computing. His professional experience includes senior consulting and researcher positions in a number of private and public institutions.

Foreword Jacques Charlier

Be they from the academic world or from the port industry, readers d iscovering this collection of essays about port systems will be highly impressed by their quality and diversity of approaches. Port systems were a kind of UFO in the previous scientific and professional literature, wherein two broad categories co-exist: spatial port systems (at various scales) and functional port systems (according to various criteria). This edited book capitalizes upon the expertise in several complementary fields of the impressive list of international contributors gathered by its two scientific editors, whose own input can be felt in several chapters in the book. The latter is conveniently organized into four parts, for which keywords could be concept(s), dynamics, collaboration, and shipping networks, with port systems as the common link. The “global” port system (singular) and the shipping networks (plural) interconnecting its components can be seen as a series of subsystems imbricated according to several dimensions. In the general picture, there are overlaps, but also gaps, in the “global”, multidimensional matrix being considered, and a given port can be a component of several of these subsystems, explaining why no simple (simplistic) typology or taxonomy of port systems is possible. Among the main axes of the multidimensional matrix mentioned above, the geographical scale (from local to regional, national, and international) and the geographical context (physical, but also urban) are often key components, together with the public governance of the subsystems being considered (from strong to weak) and the implication of private players (be they from the port or from the shipping industry). Some port (sub)systems (present and past) are de jure systems (often at a national or regional scale), while others are de facto systems (often in a more f luid situation and context). The time factor and the weight (or not) of history should also be taken into consideration. At one extremity of the spectrum, South Africa is probably the best example of a rather closed port system, with a tradition of top-down public management of the national port system (from the South African Railways and Ports administration, established in 1910, to today’s Transnet National Port Authority), with a limited involvement from the private sector in the port handling industry (as the public-owned Transnet Port Terminals owns and operates most port terminals). At the other end of the spectrum of

xxviii  Jacques Charlier

the diversity of observed situations, the United Kingdom is probably the best example, with a series of independent private ports or port groups, and many international terminal operators for which their UK facilities are components of a wider portfolio. An implicit conclusion of this challenging edited book might be that, in the current globalized world, terminal systems (or subsystems) are key components of the port (sub)systems, especially in the container industry but also more generally (as can already be observed in the cruise industry). Jacques Charlier Emeritus Professor of Geography Louvain-la-Neuve (Belgium)

Introduction César Ducruet and Theo Notteboom

Motivation of this book In a world where most international trade is carried by sea, each port can be seen as a unique value-adding node competing to attract traffic and economic activities. However, ports can also be seen as parts of a wider system, which can be defined as a system of two or more ports located in proximity within a given area. The geographical and functional scales of such port systems vary from entire continents and coastlines to the notions of a “range” and a “multi-port gateway region”. The systemic approach of ports advocates the existence of specific development mechanisms by offering a relational perspective on port development. Ports are linked with each other through the networks of logistics operators along supply and value chains. Their fate and governance are jointly inf luenced when belonging to the same region, country, or transnational space. Investments, shocks, innovations, and delays occurring in one port often affect other ports within a certain spatial range and time lapse. Further understanding of such co-developments in port systems is necessary to go beyond local specificities, through a multidisciplinary and multilevel contribution. From the local to the global, various forms of port systems have emerged and are still evolving, to such an extent that a synthesis and update is much needed. As a complex system, the port system is more than the sum of its parts, fostering complementarity and synergy. This book is an answer to the strong and urgent need for reviewing the relevant theories, concepts, methods, and sources that can be mobilized for the analysis of port systems. With contributions from reputable scholars coming from 12 countries in Europe, Asia, Australia, and the Americas, this book delves into the analysis of port systems from diverse disciplinary angles (geography, regional science, economics, management, engineering, and mathematics/computer sciences), and covering innovative empirical approaches to various port systems in the world. The theoretical and empirical knowledge can support and enhance decision-making in relation with the development of ports, supply chains, and transport networks in general.

DOI: 10.4324/9781003316657-1

2  César Ducruet and Theo Notteboom

Excluding this introductory section, the book consists of 15 chapters, evenly spread over four parts. The first part focuses on fundamental research, from the general systems theory to city-systems and port system evolutionary models. The remaining parts are more empirical, with the second one focusing on the changing distribution of f lows and facilities among ports, and the third part providing a qualitative view on how ports collaborate with each other. The last part is original by providing a unique look at the situation of ports in shipping networks functioning under contrasted operational logics.

Part I: the concept of port system The first part of this book follows a fundamental research approach, focusing on the general systems theory to city-systems and port system evolutionary models. In Chapter 1, book editors César Ducruet and Theo Notteboom analyze a corpus of more than 260 articles about port systems, published between the 1960s and 2022: How has the literature on port systems evolved internally and in relation with wider academic research? Is the concept of port system still relevant today in geography and elsewhere to study ports? They demonstrate that the analysis of port systems has a long history going back to the seminal work of Taaffe et al. (1963), later updated by Rimmer and Robinson. Subsequent generalizations consisted in updating these early models, notably to illustrate the inf luence of containerization on the concentration or deconcentration of port systems. The analysis of port systems remained virtually unchanged since the 1980s, until Notteboom and Rodrigue (2005) added the phases of port regionalization and offshore hub development. Research trends in port system studies are unraveled by focusing on various relevant themes such as concentration and deconcentration factors; the emergence of new concepts (complementarity and substitutability; proximity; regional resilience; regional integration; etc.); maritime connectivity; the hinterland dimension of port system development; and cooperation and integration in port systems. The chapter also reports on an analysis of the collaboration network consisting of 390 authors of port system studies. The analysis of their collaborations is another way to estimate the cohesiveness of the field. To conclude, the book editors propose a general multilevel framework to study the various aspects of port systems. It is argued that further research on port systems may follow several pathways. Chapter 2 bears the title “Evolutionary models of port system development: an application to the Latin American and Caribbean port system”. Jason Monios and Gordon Wilmsmeier trace the development of models that conceptualize the spatial transformations from simple multifunctional port installations to specialized terminals and relocated sites to more complex systems in which various terminals or ports compete with each other over fully or partly overlapping hinterlands. Temporal-spatial changes in the position of ports within port systems at different scales emerge from the interplay

Introduction  3

of a series of factors, principally demand and geography, but also operational factors such as maritime access limitations, port congestion, and proactive business strategies including new port locations, as well as container liner shipping network strategies. The authors illustrate these trends with examples from different ranges of the Latin American and Caribbean (LAC) port system, ref lecting concentration and deconcentration, changes in transshipment markets and the impact of changes in port governance, the effects of vertical integration, and port competition. The analysis provides the background to the discussion on how the drivers of port system evolution have changed over time due to technological change and increased industry consolidation, and how more recent challenges such as the COVID-19 pandemic may trigger further changes in the sector. In Chapter 3, Benjamin Preis elaborates on “Winding paths through urban systems and urban networks”. He observes that the modern city is one largely defined by what f lows through the city and between cities, rather than what remains permanently within or outside of it. His contribution provides a rough overview of the different scales through which urban systems can be understood. The author first defines elements of an urban system and an urban network to introduce the reader to common terms used to refer to urban network principles. He then offers a range of perspectives on the methods, assumptions, and theories underlying the differing studies of urban systems. Within the literatures that study urban systems, there is a dearth of agreement on what is being studied, and how best to study it. The author sketches the areas of overlap and disagreement among these literatures, which also provides relevant insights to the study of port systems. Although non-transport geographers pay attention to the external, space-shrinking effect of transport systems while non-geography transport researchers focus mainly on transport systems per se for their internal improvement, there are occasions where they meet. In Chapter 4 “The implications of duality of trans(port) systems: Evidence from Wusongkou International Cruise Port”, James Wang, Adolf Ng, and Joseph Lau address the question why societal benefits are considered in the first place by measuring the potential increase in accessibility and connectivity of a place and the benefits to be brought by them, noting that these measures are of geographical meaning in essence? The authors not only elaborate on the concept and theoretical importance of the duality of the transport system, but also present a case study to demonstrate the concept. Furthermore, the key research gaps and agenda toward the duality of transport systems are identified from a transport geographer’s perspective.

Part II: the dynamics of port systems The second part of the book deals with the changing distribution of f lows and facilities among ports. Rania Tassadit Dial, Gabriel Figueiredo De Oliveira, and Alexandra Schaffar open the discussion on this theme in a

4  César Ducruet and Theo Notteboom

piece on “The European ports’ size dynamics and hierarchies”. This chapter aims to study the trends of ports’ hierarchies in Europe where the spatial distribution of port’s traffic is uneven. Most studies on port concentration use traditional economic tools such as the Gini exponent, the shift-share analysis, or the Hirschman-Herfindahl coefficients. To study the European ports’ system hierarchies and dynamics, the authors follow an original approach based on rank-size and the Markov chain models. In particular, Markov chains’ models are used to study the relative growth of ports within the rank-size distribution. Markov matrices allow to observe possible changes in the hierarchy of ports and permutations of rank between certain ports. The empirical work extends to liquid and dry bulk traffics in addition to containers, to examine whether the concentration trends are similar or not. This chapter delivers two main series of results: first, traffic concentration is observed over time at an aggregate level, but this is not the case for all types of traffic. Second, both large and small ports’ hierarchical positions in the rank size distribution are persistent over time, while medium-size ports often change their position, especially for the solid bulk traffic. Chapter 6 on “Port migration patterns in the global port system since the 1950s” by César Ducruet, Theo Notteboom, and Brian Slack also combines conceptual approaches with empirical analysis. Port migration can be defined as the shift of port infrastructure and/or maritime traffic from one location to one or multiple other locations within a given period of time. The discussion on port migration is embedded in research on port and port system development, as port migration processes have an impact on cargo concentration patterns in port systems. Still, the measurement of the extent of port migration processes is underdeveloped, while also a comprehensive and structured discussion on port migration factors is lacking. The chapter focuses on two goals. The first objective is to consider the forces that are shaping port migration, based upon a review of the literature. The second goal is to provide a global quantitative assessment of port migration using three large datasets on vessel calls and tonnage. The latter analysis demonstrates the ineluctable migration process globally, with variations depending on time periods, cargo types, world regions, and locations. A cartography of port migration trajectories is proposed across port cities of the world. The conclusions discuss the difficulty to match perfectly quantitative and qualitative approaches, although they remain complementary. Chapter 7 follows a more hinterland-oriented approach to the dynamics in port systems. In the contribution “Port competition in contestable hinterlands: The case of preferential relationships and barrier effects in Central Europe”, authors David Guerrero and Jean-Claude Thill analyze port competition from a hinterland perspective. They focus on a set of Central European countries for which there is not a clear geographical advantage of one port over another. Such contestable hinterlands seem particularly relevant for an appreciation of factors that can tip the balance in favor of certain port alternatives, minimizing the statistical noise induced by distance effects. With the expansion

Introduction  5

of the European Union toward the East and the subsequent development of East-West transport links, such as the Rhein-Main-Danube canal, increased competition between ports can be expected. Chapter 7 tests this idea for different industries, by using a spatial interaction model on data on container shipments to the US. Sailing frequency is used as a measure of port attractiveness and truck drive time as geographical separation. The authors also identify preferential ties between source countries and ports and barrier effects in the organization of hinterlands. Against expectations, the results highlight the path dependence in the North-South organization of hinterlands, with a persistent split between Switzerland, mostly oriented toward Rotterdam and Antwerp, and the other countries of Central Europe, historically tied to German ports, while Mediterranean ports are largely disregarded. The last chapter of part II brings the reader back to the urban and regional context in which many ports operate by focusing on “Global cities and global logistics”. Kevin O’Connor analyses the global geography of sea and air cargo activity as it is expressed in the urban and regional context of global cities. It utilizes a spatial unit derived from the global city region, a large area surrounding a global city, where activity is shaped by production f lexibility and networks of firms, along with the role of advanced services. The spatial unit used here, the Global Logistics Region (GLR), incorporates the container terminals and airports within the global city region. The chapter revisits a 2010 study on the identification of the GLRs and analyzes their role in global logistics activity in 2006 and 2020. To what extent has the significance of GLRs changed in the past few decades as the number of containers handled at the world’s ports and the weight of air cargo carried increased significantly? Did that growth lead to dispersal toward different cities, or did the GLRs maintain their role in global activity?

Part III: collaborative port systems The analysis of collaboration and cooperation in port systems starts with a discussion on port authority mergers. From a spatial and functional perspective, cooperation and collaboration with other ports might be considered as a way for a seaport authority to achieve results and meet targets. In recent years, quite a few port systems around the world have witnessed a transition from the management of individual ports to the management of multi-port systems. Port authorities are thus regionally integrated or even merged. In Chapter 9 on “Port authority mergers in port systems: the path to North Sea Port and Port of Antwerp-Bruges in Flanders”, Theo Notteboom discusses the path toward large-scale port mergers involving port authorities of the RhineScheldt Delta port system, the most important multi-port gateway region in Europe including ports in the Netherlands and the region of Flanders in Belgium. Past attempts and concrete steps are analyzed that eventually led to the creation of North Sea Port in 2018 (the merger between Zeeland Seaport in the Netherlands and port of Ghent in Flanders) and Port of Antwerp-Bruges

6  César Ducruet and Theo Notteboom

in 2022. The author demonstrates that both mergers did not materialize out of the blue, but should be regarded as critical junctures in a decades-long path creation process marked by matchmaking attempts, and governmental and port industry initiatives involving various actors and ports in proximity. Wei Yim Yap brings the cooperation theme to Asia. In Chapter 10, he examines “Co-operation and competition between container shipping networks and their impact on container hub ports in Southeast Asia”. This contribution focuses on the port system of transshipment ports in Southeast Asia which comprises the three major container hubs of Singapore, Port Klang, and Tanjung Pelepas. The author shows the previous strategy of establishing good shipping connectivity to particular geographical regions and trade routes by anchoring specific shipping lines in the selected ports to create sizeable cargo volumes gave way to the strategy of targeting container shipping alliances and their networks. Developments in the form of the most recent spate of mergers and acquisitions and alliance reshuff le in the container shipping industry provide the context for investigating and understanding impact of inter-port dynamics created by the evolution of cooperation and competition between container shipping networks calling at the three selected ports in Southeast Asia. The chapter also reports on the impact of the COVID-19 pandemic on container shipping network dynamics and how these affected inter-port relations in the selected port system. Chapter 11 also brings insights from port cooperation and competition in Asia. In Chapter 11 titled “Port collaboration in the Great Bay Area: reality, challenge and opportunity”, authors Dong Yang, Yang Chen, and Qiang Zhang examine port collaboration/integration as a tool for strategic port governance to facilitate regional port development in the Greater Bay Area (GBA) in China. At the end of 2014, the Ministry of Transport in China initiated the port industry recentralization reform by publishing an official document outlining its intent to optimize port resource utilization through port integration. By the end of 2021, most coastal provinces in Mainland China have made a substantial response toward port collaboration/integration, while the process of port collaboration in the GBA is proceeding slowly despite being considered as a port system facing terminal overcapacity. This chapter first discusses port collaboration in the GBA by investigating its collaboration scheme, business mode, and operational pattern. Supported by extensive data analysis, the authors further explore the evolution, challenges, and future opportunities within the GBA. In Chapter 12, Karel Van den Berghe, Antoine Peris, and Wouter Jacobs present a multiple case study on “Analyzing the US- Gulf Coast petrochemical port polycentric region, and its connectivity with the European Amsterdam-Rotterdam-Antwerp (ARA) region”. The chapter builds further on the results of an earlier study examining the emergence of the polycentric port system of the Dutch-Belgian Amsterdam-Rotterdam-Antwerp (ARA) region. The authors argue that the concept of a polycentric port region allows for the revelation of more hidden interrelationships in the spatial-economic (co-)

Introduction  7

development of geographically proximate and distant port regions. A polycentric port region is analyzed through three dimensions of polycentric systems: morphological, functional, and institutional. In the contribution, the analysis is extended to one of the few comparable cases in globalized oil trade: the US-Gulf Coast (USGC) in the US. USGC, like ARA, is an important price reference point in the financial energy markets. The central question is twofold. Firstly, to what extent does the USGC port system inhibit polycentricity the way ARA does? Secondly, as the oil product space is a global market, to what extent can we observe an interrelationship between ARA and the USGC?

Part IV: port systems as shipping networks The analysis of shipping networks is a major step toward unveiling the characteristics of port connections and connectivity. Because of the dynamic nature of shipping, the most reliable method of extracting such insights is by studying the movement of vessels. The first chapter of part IV relies on large amounts of mobility data based on Automatic Identification System (AIS). In the contribution “Discovering shipping networks from raw vessel movement”, Alexandros Troupiotis-Kapeliaris, Giannis Spiliopoulos, Marios Vodas, and Dimitris Zissis describe a method that processes AIS datasets, at a regional level, and produces density maps through a novel metric that considers the shipping attributes of the traveling vessels and captures the capacity and importance of shipping routes. The proposed approach, published as a configurable open-source tool, is capable of producing data-driven insights regarding vessel movement and shipping trends solely using the raw vessel positions. In Chapter 14, Christopher Dirzka and Michele Acciaro analyze “Ocean container network dynamics during the Covid-19 pandemic”. This chapter sheds light on the economic crisis imposed by the pandemic. Examining ocean containerized network dynamics enables to assess the broader economic health and operator behavior. Building upon earlier literature, the study splits between macro- and operator view on the pandemic’s implications. The macro perspective reviews industry sources and a ship status information set. The operator perspective introduces a unique sample that includes liner operator public notices, congestion information, and freight market indicators. Overall, this chapter serves to inform the reader about disruptions in transport systems alongside the broader economic system and how stakeholders responded to distress, with a particular focus on dynamics between January 2020 and December 2021. The last chapter by the hand of César Ducruet provides a thorough analysis of shipping network literature with relevance to port system research. Under the theme “Shipping network analysis: state-of-the-art and application to the global financial crisis”, the author shows that the empirical analysis of shipping networks, for long a negligible subpart of several disciplines, is increasingly seen as a new emerging field on its own. Increased shipping

8  César Ducruet and Theo Notteboom

data availability and computational power permitted novel analyses in the domain of big data, such as in the area of port and maritime disruptions, bilateral connectivity in liner shipping networks or evaluation methods of port dominance and the visualization of vessel trajectory data. The increasing interest in shipping network analysis resulted in a multiplication of publications. This trend is reviewed in this chapter through a systematic bibliometric analysis of more than 200 journal articles published between 2007 and 2022. Although half of the corpus deals with network vulnerability and disruptions, the global financial crisis (2008–2009) has not been analyzed yet from a shipping network perspective. A new case study is thus proposed based on untapped shipping data, to verify the existence of regularities in the spread of this shock across port hierarchies and geographic scales. Main results are provided from diverse angles, documenting centrality shifts across the port hierarchy and geographical scales.

Advancing the study of port systems This book bridges classic theories and state-of-the-art methodologies to measure and reveal port system dynamics in ways previously unavailable, thanks to untapped data on port traffic and shipping f lows. It pushes current knowledge on port systems one step forward, for the first time assembling different disciplines, methods, and study areas within a single opus. The reader is offered a synthesis about the multiple forms taken by port systems. Port development can be thought beyond the local through spatial and temporal mechanisms of co-development and coevolution shaped by competitive and/or cooperative forces. Thus, development mechanisms and practices transcend individual port trajectories. This book is an ideal companion to anyone interested in the analysis of transport and economic systems in general, as well as the effective ways to answer complex issues in transportation and socio-economic development. To ensure the academic quality and consistency throughout the book, the editors have ensured that the contributions are complementary to each other in terms of theoretical background, applicability for practice, geographic scope, time periods, methodologies, and disciplinary specializations. The editors and all contributors have a substantial research track record on the book’s topic, in the field of network analysis, transport systems, port development, and maritime network analysis. However, the book brings together novel works which have never been published. To complement this book, readers might also be interested in reading our earlier edited books on the topics of port systems, shipping networks, and the regional interactions between seaports: •

Ducruet C. (Ed.) (2017) Advances in Shipping Data Analysis and Modeling: Tracking and Mapping Maritime Flows in the Age of Big Data, London & New York: Routledge.

Introduction  9

• • •

Ducruet C. (Ed.) (2015) Maritime Networks: Spatial Structures and Time Dynamics, London & New York: Routledge. Notteboom T.E., Ducruet C., de Langen P.W. (Eds) (2009) Ports in Proximity: Competition and Coordination among Adjacent Seaports, Aldershot: Ashgate. ­ Wang J.J., Notteboom T.E., Olivier D., Slack B. (Eds) (2007) Ports, Cities, and Global Supply Chains, Aldershot: Ashgate.

Acknowledgments The contributors to this book and the case studies presented in this book are from different parts of the world and with different disciplinary backgrounds. The book gives readers multiple perspectives and fresh ideas on how to examine port systems around the world. We are confident this book can contribute to the ongoing policy discussion in many parts of the world on inter-port relations. We also hope this book will provide an interesting read for researchers, practitioners, and students interested in learning more about this fascinating topic. We certainly encourage and will continue supporting more academic interest in these and other issues relevant to seaports. We would like to express our thanks to all scholars who have contributed to this book. Special thanks go to Prof. Em. Jacques Charlier who has enthusiastically accepted our invitation to write the book’s preface. Finally, we also would like to express our gratitude to Mrs. Alexandra Atkinson, Routledge editor for Business, Operations & Marketing Research, for her support throughout the whole editing process. Paris and Antwerp, November 2022

Part I

The concept of port system

1

A systematic and critical review of port system research César Ducruet and Theo Notteboom

Introduction: emergence and diffusion of the port system concept The concept of system is very popular in both natural and social sciences, but its meaning differs greatly across the academic spectrum. Yet, it is commonly accepted that systems are sets of interactions forming a totality, while being more or less resilient, balanced, and self-organized (Sanders, 2005). In geography, a spatial system is open, through relations with its environment (Pumain, 2004). Related issues concern the conditions of its emergence (systemogenesis), its evolution (resilience, bifurcations), interactions among subsystems (self-organization) as well as top-down or multilevel interactions. Systems may emerge at all geographic scales, such as the world system, urban or regional system, and local productive systems. The general systems theory developed by Bertalanffy (1950) considered that a system is greater than the sum of its parts – an idea already developed by eminent scholars such as Aristotle, Confucius, Avicenna, and Pascal. This is due to synergy effects and emergent behavior among the parts, which foster new qualities or properties. Each part can also be a system on its own, as proposed by Berry (1964) when considering “cities as systems within systems of cities”. Port studies, by their principal focus on single ports, developed in relative isolation from mainstream research on systems. Systemic thinking emerged relatively soon in port geography, however. In Northwest Europe, Pounds (1947) identified a “normal pattern of relationship” between upstream (port) and downstream (outport). He summarized the emergence of outports by four main factors: (1) silting of estuaries and lower courses of rivers, (2) increasing size of vessels, (3) ports de vitesse (passenger transshipment at optimal locations), and (4) time and cost of navigating the estuary. The spatial model Anyport was proposed by Bird (1963) based on similar dynamics taking place at British estuaries. The shift of port facilities from upstream port city to downstream deep-sea locations was verified in various parts of the world by other geographers. Other spatial models or “chorotypes” of the estuary were proposed, highlighting the complementarity between upstream and downstream port cities (see Brunet, 1990; Brocard et al., 1995).

DOI: 10.4324/9781003316657-3

14  César Ducruet and Theo Notteboom

Two major contributions enlarged the focus from two ports to multiple ports. An ideal-typical sequence of transport system development was proposed by Taaffe et al. (1963) based on the cases of Ghana and Nigeria. In addition to the emergence of inland transport corridors, the spatial model and related cartographies particularly insisted on the process of port concentration within each country over successive phases. The transnational perspective was provided by Vigarié (1964) in his book on the major trading ports from the Seine to the Rhine. His historical analysis at the level of the whole maritime range also pointed to a concentration process, at the advantage of the Benelux, in a context of expanding hinterlands and ongoing European integration. One originality of Vigarié’s work had been to consider forelands in addition to the port itself and to hinterlands as a factor of port competitiveness. The importance of maritime linkages motivated Rimmer (1967a, 1967b) to update the ideal-typical sequence of Taaffe et al. (1963) and apply his new spatial model to Australia and New Zealand, although the added maritime linkages did not account for the whole foreland of ports. It is only one year later that the concept of port system emerged explicitly in the literature, from a maritime perspective. Robinson’s (1968) PhD on British Columbia ports constitutes the first-ever empirical analysis of a shipping network using graph theory, with multiple references to port geography and regional science. He proposed that the system of ports operating interdependently may now be regarded in abstract form as a set of points or nodes in a network, a transportation network in which the lines or links are in fact ‘imaginary’ routes traced out by foreign trade shipping. For the first time, ports were compared by their connectivity (degree centrality) and nodal importance (hub function, or focal ports) rather than their tonnage. A port network development model was also added. The next work of Robinson (1976) was thought as a bridge between operations research and regional science. The author proposed a new spatial model including five geographical scales: (1) intra-port or single element system (e.g., inter-berth movements), (2) port-hinterland system, (3) two-element port system (shipping network), (4) regional port system (several ports along a coastline), and (5) N-port system (the “total” system with its hinterlands and overseas linkages). Ship turnaround time was considered to be the core indicator of the system’s wealth, as “Inefficiency within the port (…) will eventually restructure the basic spatial patterns (…) of existing facilities [and] lead, of course, to spatial restructuring at an interport scale”. The works of Robinson did not have much followers, as subsequent studies primarily focused on the coastline and its hinterland. The lack of shipping data and the fact that his PhD had never been published partly explain this state of affairs. A few years later, Vigarié (1979) proposed a parent model, the port triptych, whereby ports connect forelands and hinterlands, but it was not accompanied by a proper methodology. Other generalizations mainly

A systematic and critical review of port system research  15

consisted in updating the landward-looking model of Taaffe et al. (1963), by documenting the inf luence of containerization on the concentration (Hayuth, 1981) or deconcentration (Hayuth, 1988) of port systems, especially in North America. Examples from all over the world were proposed in the seminal book of Hoyle and Hilling (1984) on Seaport Systems and Spatial Change. As a consequence, “only a few theories exist on the development of a container port system in relation to forelands, hinterlands and the technological environment” (Notteboom, 1997). Despite the relatively sheer number of publications on the spatial analysis of ports (Pallis et al., 2011) and port systems (Ducruet et al., 2009), “the analysis of port systems remained virtually unchanged since the 1980s” according to Pallis et al. (2011). Yet, Woo et al. (2011) considered spatial analysis and especially the “Hayuth model” to provide “relatively stronger theoretic bases than other research theme categories”, although being “purely geography-based”. However, Vallega (1986) questioned the systemic dimension of ports relatively early, arguing that spatial proximity ceased to be relevant to define port systems since the emergence of hubs. New spatial configurations modified traditional port systems, which were based on their integration with local industries and closer portcity relationships (Hoyle, 1989). Other reviews of port research were proposed, giving more or less importance to the port system theme. When visualizing title words’ co-occurrences within a corpus of papers about container shipping, Lau et al. (2017) only found a tiny link between “port system” and “ports” for the period 1999–2013. Among keywords associated with “port”, the term “system” did not appear among the top ranked ones in Chen et al. (2018), compared with port performance, state control, governance, choice, competition, and privatization. Munim and Saeed (2019) did not even identify port system as a co-occurrence, compared with port choice, hierarchy, cooperation, among others. It was also absent from the review of Wang and Peng (2021), which better underlined the importance of port simulation, location, hinterland, service, allocation, policy, and integration. Important questions thus arise: how has the literature on port systems evolved internally and in relation with wider academic research? Is the concept of port system still relevant today in geography and elsewhere to study ports? The literature on port systems has never been systematically reviewed. Previous attempts looked at one facet only, such as concentration and deconcentration dynamics (Ducruet et al., 2009), the evolution of port geography since the 1950s (Ng and Ducruet, 2014), and the delineation of port systems (Ducruet and Notteboom, 2022). Interestingly, Ng and Ducruet (2014) as well as Ducruet et al. (2019) classified port system papers in a category distinct from “port connectedness”, “port choice, competition and cooperation”, and “port’s place in shipping strategies and networks”, while the delineations proposed by Ducruet and Notteboom (2022) were only based on the design of shipping networks. This echoes in some way the old debate

16  César Ducruet and Theo Notteboom

between hinterland and foreland, and is partly caused by the absence of a commonly accepted definition of the port system. In this chapter, we analyze a corpus of more than 260 articles about port systems, published between the 1960s and 2022. The methodology of corpus construction is presented in the next section, together with an analysis of the main publication trends. It is followed by a classification of main research subthemes. In particular, we attempt to verify the extent to which port system research constitutes a relatively unified research field, and whether it innovates compared with its anchoring in spatial models. Further research pathways are discussed in conclusion.

The corpus of port system studies A search in Google Scholar dated 15 August 2022 was performed in two successive steps. First, we selected any journal article explicitly mentioning “port system” in its title and/or abstract, including variants such as seaport system, inter-port system, multi-port system, or seaport-dry port system. The pioneering works on port systems, which did not use these expressions (Taaffe et al., 1963; Rimmer, 1967a, 1967b; Hayuth, 1981, 1988), were included in the corpus. We also included all articles not explicitly mentioning “port system” but citing these models, and dealing with ports. A total of 268 articles have been identified, comprising of 169 articles explicitly mentioning the notion of “port system” and 99 articles referring to port system models (see Appendix 1.1). Several interesting trends can be observed in Figure 1.1. The total number of articles remained relatively modest during the early period, followed by rapid growth from the early or mid-2000s. Port system research is thus a relatively buoyant field today. Considering each sub-category, articles explicitly mentioning “port system” have always been dominant in number, except for a few years. This means that the concept of port system gained popularity over time, but also that the original models never ceased to inf luence new research. Figure 1.2 underlines that the share of geography journals declined continuously over time, compared with transport and other journals. Such an evolution recalls the findings of Ng and Ducruet (2014) about port geography in general. They observed that geographers themselves increasingly published in transport journals: “port geography has gradually drifted away from actively participating in the philosophical discussions within human geography”. Between 2020 and 2022, only seven papers were published in geography journals, against 34 in transport journals and ten in other journals. However, the ratio between the share of authors being geographers and the share of geography journals has always been over 1, for the reasons mentioned above, in the 1980s (1.86), 1990s (1.56), 2000s (1.59), 2010s (1.58), and 2020s (2.00). Figure 1.2 also reveals a diffusion of the concept of port system from geography to transport and other fields, in parallel with the growing absolute number of publications.

A systematic and critical review of port system research  17

Table 1.1 presents the journals having contributed the most to port system research. While Journal of Transport Geography is both a geography and a transport journal, five “general geography” journals belong to this list: GeoJournal, TESG, Environment and Planning A, Economic Geography, and Chinese Geographical Science. Among the top transport journals are maritime journals, with Maritime Policy and Management and Maritime Economics & Logistics claiming a dominating role, followed by Marine Policy and Journal of Maritime Research. The rest of the journals are “general transport” journals.

20

Number of articles

18 16 14 12 10 8 6 4 2

Core corpus

2020

Extended corpus

Figure 1.1 Publication trend of port system articles, 1963–2022

100% 90%

Share of articles

80% 70% 60% 50% 40% 30% 20% 10% 0%

1960s

1970s

1980s

Geography

1990s Other

2000s

2010s

Transport

Figure 1.2 Distribution of articles by journal background, 1963–2022

2020s

2022

2018

2016

2014

2012

2010

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1985

1982

1977

1969

1963

0

18  César Ducruet and Theo Notteboom Table 1.1 Top 20 journals by the number of publications about port systems Journal

No. of articles

Maritime Policy and Management Journal of Transport Geography GeoJournal Transport Policy Maritime Economics and Logistics The Asian Journal of Shipping and Logistics Tijdschrift voor Economische en Sociale Geografie Journal of International Logistics and Trade Marine Policy Research in Transportation Economics Transport Reviews Environment and Planning A International Journal of Shipping and Transport Logistics Journal of Digital Convergence Sustainability Transportation Research Part A: Policy and Practice Economic Geography Journal of Maritime Research Chinese Geographical Science

38 37 13 11 8 6 6 5 5 5 5 4 4 4 4 4 3 3 3

Research trends in port system studies Concentration and deconcentration factors Three main approaches to hierarchical phenomena can be distinguished in Table 1.2, which condenses no less than one-third of the whole corpus. The first one, essentially quantitative, investigates whether port traffic distribution ref lects the successive phases of concentration and deconcentration of port system evolutionary models. Such studies apply a variety of measures to traffic time series, such as concentration ratio, Gini coefficient, and Herfindahl–Hirschman index for the most popular. Hilling (1977) has been the first scholar, to our knowledge, to employ concentration ratios. The work of Notteboom (1997) and its updates (Notteboom, 2006, 2010, 2016) on the European and other port systems worldwide reinvigorated such an approach. The early adoption of new technologies such as steam, containerization, and intermodalism (see the review by Ducruet and Itoh, 2022), and a large market size (city, hinterland) were often recognized to benefit port concentration, and the development of the “periphery” (e.g., new port development, traffic shifts) at the expense of the “core” (e.g., lack of space in large load centers, handling costs, congestion) as a factor of deconcentration, in line with the models. The second approach, also quantitative, is a search for “laws” or “mechanics” ruling the structure and/or evolution of a large sample of ports. For instance, it was demonstrated that the largest ports are also the most diversified (Charlier, 1988; Ducruet, 2017, 2022), although smaller ports may also have a wide commodity portfolio, when hosted by large cities, benefiting

A systematic and critical review of port system research  19

from a captive hinterland, and locating at the periphery (Ducruet et al., 2010). In the same vein, Ducruet (2020) verified the applicability of scaling laws to ports, showing that the largest cities concentrate the most valuable traffic (containers, general cargo). The rank-size distribution (or Zipf ’s law) of port throughput served to scrutinize hierarchical tendencies, as in the case of China (Xu et al., 2021; Li et al., 2022) and at the global scale (de Oliveira et al., 2021). The originality of the latter work is to complement the ranksize analysis by a look at the intra-distributional mobility of ports: “Markov chains aim to explore the extent to which ports move between groups of large, medium or small size units”. The authors observed a strong stability globally and regionally, due to the low probability for ports to shift across traffic size groups over time (see also Chapter 5). The third approach is more qualitative. Scholars discussed the impact of European integration on port concentration as a whole (Gilb, 1992), within UK, and in relation to its Benelux hub (Hilling, 1989; Wilmsmeier and Monios, 2013). Overman and Winters (2005) offered a New Economic Geography approach to UK port concentration dynamics, but without mentioning the port system literature. When it comes to governance, there is a consensus that countries with regulated competition and centralized administration have a more concentrated port system (Miyajima and Kwak, 1989; Le and Ieda, 2010; Wang and Ducruet, 2012), while those experiencing liberalization, devolution, integration (merger), and strategic planning exhibit deconcentration (Chaudhury and De, 1999; Wilmsmeier et al., 2014; Nguyen et al., 2018; Feng et al., 2020, 2021). Critiques to port system evolutionary models were proposed in light of actors’ strategies notably in Asia (Slack and Wang, 2002), where the “challenge of peripheral ports” was better explained by governance issues and operational strategies of global terminal operators than by diseconomies of scale in large load centers. In Europe, Frémont and Soppé (2007) demonstrated that the Le Havre – Hamburg range exhibited a stable container port throughput concentration since the 1970s, but a growing shipping line concentration at dedicated hubs, in a context of alliance formation and horizontal integration. With very few exceptions, the listed analyses were conducted in a context of uninterrupted port growth. It is only recently that scholars analyzed the effects of shocks and crises on port systems (see Chapter 14 on COVID-19 and Chapter 15 on the global financial crisis). One early exception is Walker (1989), discussing the impacts of the economic recession on Gulf States ports in the 1980s. The North Korean crisis provoked a drastic reduction of domestic port traffic, which shifted westward and concentrated near Pyongyang (Ducruet, 2008). Looking at exogenous shocks on Kobe, New York, and New Orleans, Rousset and Ducruet (2020) found that traffic rerouting occurred at the periphery of respective regional port systems. Cariou and Notteboom (2022) studied the inf luence of the COVID-19 on import cargo routing by Walmart and Nike through U.S. ports, also analyzing overall concentration dynamics in the U.S. container port system.

Author(s)

Taaffe et al. Rimmer Rimmer

Hilling Hilling Hayuth

Westerholm Charlier Hayuth Marti Hilling

Warf Miyajima & Kwak Hayuth Gilb Kuby & Reid

Hoyle & Charlier Notteboom

Wang

McCalla

Year

1963 1967 1967

1969 1977 1981

1986 1988 1988 1988 1989

1989 1989 1991 1992 1992

1995 1997

1998

1999

North America

Hong Kong & China

East Africa Europe

USA Japan EU & North Am. Europe USA

Denmark Belgium USA Pacific England & Wales

West Africa Ghana USA

Ghana, Nigeria Australia New Zealand

Area

–

Load center

Petroleum products Regulated competition Intermodal gateways European integration Containerization, larger ships & trains, computers Transport system integration –

Trade & economic growth Deep-water ports Local market size, inland accessibility, site & situation, early adoption (container) – Multifunctional ports – Load center development Containerization, larger bulk ships

Inland corridors Railway and road connections Internal accessibility

Concentration

Table 1.2 Por t system concent rat ion a nd deconcent rat ion factor s, 1963 –2022

– Inter-range competition, Med hub development, traffic shift to medsize ports Cross-border investments (HK) to the periphery (China) Shift from New York to middle ranked ports

Evening out – division of labor Specialized ports Challenge of secondary ports – Hub-feeder system between southeastern small ports and Benelux Crude petroleum Container growth in the periphery Multiple itineraries – –

– Specialization, outport development Local & national policies, commodity specialization New port development – –

Deconcentration

20  César Ducruet and Theo Notteboom

Author(s)

Hoyle Chaudhury & De Brunt

Slack & Wang

Memon et al. Notteboom & Rodrigue Notteboom Notteboom Notteboom

Ducruet

Ducruet et al.

Jaja Notteboom Ducruet et al. Le & Ieda

Veenstra Wang & Ng Wang et al. Wang & Ducruet Ducruet & Notteboom

Year

1999 1999 2000

2002

2004 2005

2006 2006 2007

2008

2009

2009 2010 2010 2010

2011 2011 2012 2012 2012

Yangtze River China Pearl River Delta Yangtze River World

Nigeria Europe Northeast Asia Northeast Asia

North Korea

USA & Europe East Asia Le Havre – Hamburg range Northeast Asia

New Zealand Europe & USA

Asia

Kenya India Ireland

Area

External hub dependence, crisis & isolation Metropolization, trade reorientation, political crisis Colonization, consolidation Strong concentration Hub centralization Centralized governance (Korea), concentrated development (China, Korea) Port regionalization Hub-and-spoke network – Metropolization, national policies Throughput growth

Inter-range structure (USA) – Stability

Import trade –

–

Mombasa – Metropolization

Concentration

Early development stage Multilayered network Peripheral challenge – – (Continued)

– Multi-port gateway regions – Decentralized governance (China, Japan)

–

–

Second deep-water port Liberalization, increased competition National Development Plans, structural funds, maritime corridors Management & operational strategies of private/public bodies Export trade Port regionalization & offshore hub development Balance (Europe) Chinese growth, Japan stagnation –

Deconcentration

A systematic and critical review of port system research  21

Fraser & Notteboom

Li et al. Liu et al. Wang & Ducruet Wilmsmeier & Monios Reina & Villena Pan et al.

Wilmsmeier et al.

Grushevska & Notteboom Mohamed-Chérif & Ducruet Pham et al. Yang et al. Wang et al. Castillo & Valdaliso Wang et al.

2012

2012 2013 2013 2013

2014

2014

Nguyen et al.

Grifoll et al. Krivoshchekov et al. He et al. Svindland et al.

2018

2018 2018 2019 2019

2016 2017 2017 2017 2017

2016

2013 2014

Author(s)

Year

Table 1.2  (Continued)

Mediterranean Russian Arctic Yangtze River Norway

Vietnam

Vietnam Yangtze River China Spain Europe-Asia

North Africa

Latin America & Caribbean Black Sea

Spain China

China & USA Pearl River Delta China UK

Southern Africa

Area

– – Hinterland rivalry, primacy Traffic growth

–

– Iron ore concentration – Maintained top 10 Multi-hub spatial agglomeration

Hub dependence (Tangier)

Low concentration

Traffic growth, hub-and-spokes systems

General cargo, containers –

Concentration 1995–2005 – Regional resilience High concentration

Political & economic stability

Concentration

Location (upstream ports) – Functional differentiation Deconcentration 1880Functional differentiation, division of labor, polycentricity Location, administration, strategic planning Eastward shift No competition Overcapacity Sustained no. of ports

Low integration, shift of transshipment to Med –

Containerization at new ports, corridors, expansion projects, regional Integration Overall deconcentration Multi-gateway port network Increased no. of ports Hub-feeder system between southeastern small ports and Benelux Bulks Challenge of the periphery, multigateway port regions Port devolution

Deconcentration

22  César Ducruet and Theo Notteboom

Monios et al.

Chlomoudis & Styliadis Michael Nguyen et al.

2019

2019

Huang et al. Feng et al.

Xu et al. Iyer & Nanyam Ducruet & Itoh Ducruet Zheng & Shao Esteve-Perez & Gutierrez-Romero Li et al. Zhang et al.

2021 2021

2021 2021 2022 2022 2022 2022

2022 2022

2020 2020 2021

Rousset & Ducruet Liu & Ping Liu & Yeo Lopez-Bermudez et al. Feng et al. Zhao et al. De Oliveira et al.

2020 2020 2020 2020

2019 2020

Author(s)

Year

Pearl River Delta Maritime Silk Road

China India World World China Baltic Sea

Yangtze River Delta Yangtze River Delta

Yangtze River Delta China World

Comparative Pearl River Delta Mediterranean Spain

Nigeria Southeast Asia

USA

Comparative

Area

– –

Container – Innovation diffusion (steam, container) Traffic diversity – Stability

– Core-periphery structures Low mobility, higher impact of top ports vs. midsize ports – –

– – Westward shift (cruise) Vulnerability of small & specialized ports

Location, hinterland coverage, port size –

Ship size, alliances, global operators

–

Concentration

Overcapacity, wasteful competition Growth of small & medium-sized ports, bifractal structure

Multi-core development Decentralization, regionalization, water depth, competition General cargo Perfect competition – Traffic specialization Chaotic competition German ports growth

Port integration – –

– Market share growth at inefficient ports Peripheral growth, rerouting Spatial spillovers, cooperation – –

Minimal sailing distance, suitable channel & berth depth, high capacity inland links –

Deconcentration

A systematic and critical review of port system research  23

24  César Ducruet and Theo Notteboom

The emergence of new concepts Taking some distance from the relative linearity of spatial models, economic geographers proposed to analyze port systems in the light of wider concepts and theories. For instance, Notteboom (2009) used the concepts of complementarity and substitutability to analyze the calling patterns of shipping lines among adjacent ports. This work particularly emphasized that large load centers increasingly act as substitutes while smaller ports are more and more complements of the latter. Hall and Jacobs (2010) investigated the various dimensions of proximity between seaports and among port users, from a cognitive and organizational point of view. Asking “how do seaports evolve in relation to each other?”, Jacobs and Notteboom (2011) claimed that previous studies on ports “lack theoretical foundations”. These authors proposed an evolutionary framework to study regional port competition. They particularly developed the concepts of windows of opportunity and the associated “critical junctures” to analyze the respective roles of the institutional context and of strategic agency in the changing roles and status of ports in the Rhine-Scheldt delta. Another concept, regional resilience, was used in the analysis of the Chinese port system between 220BC and 2010AD (Wang and Ducruet, 2013). It helped to demonstrate a certain stability in the spatial distribution of main gateways, despite numerous periods of closure and openness. Another example is Monios and Wilmsmeier (2016) in their analysis of port governance and port traffic distribution through the concepts of path dependency and contingency (see also Wilmsmeier et al., 2014). They particularly showed how port reform (devolution) in Latin America depends on initial institutional settings and what is the outcoming balance of power with the private sector. Path dependence was also used by Castillo and Valdaliso (2017) for the analysis of economic, technological, and institutional factors shaping the evolution of the Spanish port system between 1880 and 2014. The concept of regional integration was explored by Mohamed-Chérif and Ducruet (2016) in the case of the Maghreb based on inter-port container f lows, referring to the work of Lemarchand and Joly (2009) on “regional integration and maritime range”. The latter authors used statistical modeling techniques to measure the level of integration of a port system. They demonstrated that well-integrated port systems should have a negative, significant correlation between the average traffic size of ports and the standard deviation of their yearly growth rates (North America, North Europe). Low integration occurred at port systems welcoming new ports or transshipment hubs (Mediterranean, Asia). In North Africa, however, the multiplication of cross-border maritime container linkages was better explained by hub development (Tangier Med) than regional integration. New types of spatial patterns have been identified, such as multi-port gateway regions (Notteboom, 2010; Feng and Notteboom, 2013; Liu et al., 2013; Pan et al., 2014), bifractal structure (Zhang et al., 2022), multicore port

A systematic and critical review of port system research  25

development (Huang et al., 2021), spatial spillovers (Liu and Ping, 2020), and polycentricity (Wang et al., 2017), sometimes considered as new phases extending port system evolutionary models. Explicit reference has been made to peripherality in a number of publications, either as a weakness or counterforce to concentration (Feng and Notteboom, 2013; Feng, 2013; Wilmsmeier and Monios, 2013; Fraser et al., 2016; Wiradanti et al., 2018, 2020). Maritime connectivity Most studies of port systems are based on spatial proximity between ports (see Notteboom et al., 2009). Echoing the study by Notteboom et al. (2013) on port research, Figure 1.3 classifies port system articles according to their geographic reach. It confirms the strong dominance of national-level analyses, except in the 1990s and 2000s. Subnational studies emerged only in the 1990s. There is a clear tendency for larger scales to gain popularity, since the 1980s. This spatial expansion of the scope of analysis may imply changes in the definition and functioning of port systems. It goes along with the fact that a maritime network can be a port system, as the ocean space is relatively borderless: “every port node can theoretically be linked to every other port node” (Bird, 1984). Very few works studied maritime networks after the work of Robinson (1968). Among them, Marti and Krausse (1983) searched for evidence of load center development in the pattern of transatlantic container f lows. Two unpublished PhD dissertations used graph theory to analyze network rationalization (Helmick, 1994) and connectivity ( Joly, 1999), respectively. It is only in the late 2000s and early 2010s that increased computational power and shipping data availability made possible the analysis of large maritime 100% 90%

Share of articles

80% 70% 60% 50% 40% 30% 20% 10% 0%

1960s Subnational

1970s National

1980s

1990s

Transnational

2000s Continental

Figure 1.3 Distribution of articles by level of analysis, 1963–2022

2010s Global

2020s

26  César Ducruet and Theo Notteboom

networks (see Chapter 15 for an extensive review). For instance, Ducruet et al. (2010) demonstrated that Chinese ports were the largest by traffic volume and the fastest by traffic growth rates, but remained minor nodes in terms of centrality, polarized by Busan and Hong Kong hubs. The most important contribution of network analysis to port system research has been to identify tightly connected clusters of ports within maritime networks (see a review by Ducruet and Notteboom, 2022). This is because it confers a functional character to the obtained group as opposed to a purely administrative or geographic definition. Depending on the clustering algorithm employed, the nature of the obtained groups and thus of the research question may vary extensively. Single or multiple linkage analysis, focusing on the largest f low(s) of each port node (Nystuen and Dacey, 1961), reveals subsystems or “nodal regions”, each polarized by a dominant node in the form of hub-and-spokes structures (Wang and Wang, 2011; Cullinane and Wang, 2012; Ducruet, 2013; Wang and Cullinane, 2014). This structure may be more or less complex (monocentric, polycentric) and its geographic coverage can extend from a given port range to several continents, depending on spatial frictions or “barrier effects” (e.g., distance, borders, coastal morphology). Another method has been to isolate horizontal linkages between nodes of similar connectivity level (topological decomposition). Ducruet and Zaidi (2012) showed that small ports were usually clustered based on geographic proximity, except between Europe and its former African and Latin American colonies. This was considered as a marker of language proximity, trade preferences, and/or historical legacy (e.g., ports of Portugal/Brazil/Angola, Spain/Latin America, and France/North Africa/Caribbean). At the top of the port hierarchy, a cluster of major hub ports formed what is called a “rich-club” in network science. Other methods of graph partitioning and cluster identification were also proposed, including modularity, stochastic blockmodeling, bisecting k-means, and the Louvain algorithm, among many other. Port system research reaching out to hinterlands Next to developing a seaward focus through advances in the analysis of nodal-based maritime connectivity and shipping networks, research on port systems has also captured dynamics in hinterland networks. The early models on port system development as presented in the seminal works of Taaffe et al. (1963), Rimmer (1967a, 1967b), and Hayuth (1981) already explicitly considered the role of hinterland corridor infrastructure and connectivity, and the emergence of shared or overlapping hinterlands as important factors shaping concentration dynamics and inter-port competition in seaport systems. Hinterland connections have been recognized as critical to port competitiveness and development in most port systems around the world, requiring coordination between multiple actors often

A systematic and critical review of port system research  27

with conf licting mandates (Merk and Notteboom, 2015). In the past 15 years, the conceptualization and empirical analysis of spatial and functional aspects of port-hinterland relations has seen strong growth, particularly in two research domains. First, there has been a great interest in modeling inter-port competition in port systems by considering the hinterland dimension. The cargo routing choices of shippers and shipping lines depend on a lot of factors such as costs (port, inland transport, sea voyage, feeder/shortsea), time (at sea, in port, and to the hinterland), efficiency, inland connectivity of ports of call, frequency, reliability, infrastructure and equipment availability, cargo balance (import/export), and ease of business. Quite a few port choice models have been developed and tested explicitly incorporating hinterland aspects. For example, Biermann and Wedemeier (2016) determined the contestable economic potential of the hinterland of Hamburg and of its possible emerging competitors using simple travel time matrices for different transport modes. The world container model in Tavasszy et al. (2011) models the movement of containers on a global scale. Veldman and Bückmann (2003) and Veldman et al. (2005) applied a container port competition model based on logit analysis to the Le Havre-Hamburg range ports. Also, Zondag et al. (2010) built a container port competition model based on generalized cost for North West Europe. More recently, Mueller et al. (2020) developed a model which includes 31 European mainland container ports, 231 NUTS-2 regions in Europe, and three hinterland transport modes (road, rail, and barge) as well as deep-sea and feeder shipping, to analyze port choice. Other examples of spatial-economic analyses of port system-hinterland relations can be found in Guerrero and Montes (2021) for the French hinterland; Garcia-Alonso et al. (2016) and Moura et al. (2017) for the Spanish hinterland; Yang et al. (2016) for Shanghai’s inland reach; and Wan et al. (2020) on the hinterlands of Chinese major container ports. All these models show that the outcomes are very sensitive to changes in input variables, such as costs, time variables, or frequency/reliability. Second, the discussion on the role of hinterlands in port systems’ spatial development saw a new impetus with the introduction of the port regionalization concept by Notteboom and Rodrigue (2005), implying a gradual process where efficiency is derived from higher levels of integration with inland freight distribution systems. Market forces and political inf luences gradually shape “regional load center networks” with varying degrees of formal linkages between seaports and inland nodes of these networks. The port regionalization thesis sparked a wave of research on inland port/terminals, dry ports, and dry port systems. In broad terms, an inland/dry port can be defined as an inland facility with or without an intermodal terminal and logistics companies, which is directly connected to seaport(s) with high

28  César Ducruet and Theo Notteboom

capacity transport mean(s) either via rail, road or inland waterways, where customers can leave/pick up their standardized units as if directly to a seaport. (Witte et al., 2019 based on Roso et al., 2009 and Wiegmans et al., 2015) The development of intermodal transport corridors connecting seaports to inland/dry ports supported the development of large logistics zones and logistics polarization in the hinterland, while making port hinterlands becoming more discontinuous in nature, in some cases even leading to the establishment of “islands”, i.e., a discontinuous hinterland served by the port but located in the continuous hinterland of another port (Notteboom and Rodrigue, 2005). Witte et al. (2019) identify three phases in the evolution of inland port research and point to a shift in attention to governance and management of inland ports, as well as their spatial and economic impacts to surrounding regions. Some level of conceptualization has taken place with respect to the directional development of inland ports within their specific local, regional, and (inter)national contexts. For example, based on the type of vertical control of the development process, Wilmsmeier et al. (2011) distinguished between inside-out development, whereby inland intermodal terminals seek greater integration with their seaports (often driven by public body intervention) and outside-in development whereby inland terminals are used by seaport actors to expand their hinterland. This conceptualization initially applied to Scotland, Sweden, and the USA has been extended, refined, and applied in later works, e.g., the outside-in port hinterland integration in Veracruz as analyzed in Wilmsmeier et al. (2015). Witte et al. (2019) rightly observe that most work has followed an outside-in approach. Empirical work has mostly relied on country-based or individual terminal case studies, with only few exceptions such as the global dry port dataset in Nguyen and Notteboom (2019) used to examine the relations between dry port characteristics and regional port-hinterland settings. Several scholars have also embarked on analyzing the spatial development of inland port systems, particularly focusing on major inland waterways such as the Rhine and Yangtze. For example, Veenstra and Notteboom (2011) analyze the structure and development of the Yangtze River port system using statistical measures that have not been used much by transport geographers in ports, and adapt port system development models to river ports. Notteboom et al. (2020) present a comprehensive comparative empirical analysis focusing on the container shipping networks in the Yangtze and the Rhine and show a large diversity in how inland port systems and related gateway seaports are dealing with cargo f lows and supply chains. Yang et al. (2021) added an additional layer to the study of inland port systems by examining entry strategies of actors in inland container terminals on the Rhine and Yangtze in terms of their spatial, temporal, and institutional characteristics.

A systematic and critical review of port system research  29

Cooperation and integration in port systems: an emerging research field? There is a multiplication of port integration schemes and mergers at port system level. Recent examples include the 2022 merger between Antwerp and Bruges in Belgium to form the port of Antwerp-Bruges, the creation of HAROPA (Le Havre, Rouen, and Paris) in France in 2021, cross-border port mergers such as Copenhagen-Malmo (Denmark/Sweden) and North Sea Port (see Chapter 9 for an in-depth case study), the creation of port system authorities in Italy, or the formation of provincial port groups in China. While port cooperation is a hot topic in public and business circles, bibliometric studies on port-related academic research reveal that port cooperation/integration is an emerging theme, but that the number of published peer-reviewed papers is still quite limited in the maritime economic literature (see the content analysis in Pallis et al., 2011 and Woo et al., 2011) and port geography literature (Ng et al., 2014). Still, there is a growing scholarly interest in analyzing the spatial implications of port system cooperation and integration on the triptych foreland-port-hinterland. The edited book on “Ports in proximity” (Notteboom et al., 2009) as well as several volumes on port governance reform and devolution (Brooks and Cullinane, 2006; Brooks et al., 2017) provide a wealth of in-depth case studies on the changing spatial and institutional context within (often national) port systems. More recently, port authority integration and mergers have received more explicit interest as exemplified in a special journal volume by Notteboom et al. (2018). Ducruet et al. (2023) reviewed port cooperation among European ports in the realm of environment and sustainability. The analysis showed that such networks mainly concerned southern ports (see Table 1.3), with Valencia as a recurrent leader, while individual projects are more concentrated in northwestern ports. The listed projects are often financed by the European Commission to organize workshops or to test prototypes of new technologies. Overall, existing publications are mostly case-based descriptive studies. There is room for a more methodological approach in the study of port integration and its spatial implications on port systems. Port governance integration in China through the formation of port groups at provincial level has gained most interest in recent scholarly work (see e.g., Wang et al., 2015; Notteboom et al., 2017; Huo et al., 2018; Yang et al., 2019; Ma et al., 2021). The observed port integration in China has led Zhang et al. (2021) to investigate the evolution of the Chinese port system by treating 11 coastal provinces as the units of analysis. Li et al. (2022) investigated the evolution of the integrated port system in the Pearl River Delta (Guangdong-Hong Kong-Macao Greater Bay Area) by applying the evolutionary model of Notteboom and Rodrigue (2005). As the Pearl River Delta is the only major port system in China that has not been subjected to port integration at provincial level, the authors call for a more system-wide coordination and collaboration among ports, aiming to avoid overcapacity (see Chapter 11 by Yang et al.).

Hamburg Trieste _

2014–2020 2014–2020 2015–2021

2016–2020 2018–2020 2018–2020

Genoa

2012–2015

PORT-Cities: Integrating Sustainability (PORTIS) LOOP-Ports project (Circular Economy Network of Ports) SUstainable Ports in the Adriatic–Ionian Region (SUPAIR)

Valencia

2012–2014

Antwerp Valencia Trieste

Antwerp

Bremen

2012–2014

2016–2020

Zaanstad Venice

Leader port

2010–2013 2010–2013

Duration

CIVITAS PORTIS

E-Harbours Electric Common Mediterranean strategy and local practical Actions for the mitigation of Port, Industries and Cities Emissions (APICE) Green and Effective Operations at Terminals and in Ports (Green EFFORTS) Green Technologies and Eco-efficient Alternatives for Cranes and Operations at Port Container Terminals (GREENCRANES) Managing the Environmental Sustainability of Ports for a durable development (MESP) Smooth Ports Sustainable Urban Mobility in MED PORT cities (SUMPORT) Developing Low carbon Utilities, Abilities and potential of regional entrepreneurial Ports (DUAL Ports)

Project

Table 1.3  Cooperation projects on sustainability among European ports, 2010–2023

Aqaba, La Spezia, Patras, Tripoli (Lebanon) Livorno, Monfalcone, Nantes, Varna Durrës, Igoumenitsa, Koper, Kotor, Limassol, Valencia, Thessaloniki Emden, Hvide Sande, Niedersachsen Ports, Ostend, Skagen, Vordingborg, Zwolle Aberdeen, Constanța, Klaipėda, Trieste 33 European Union (EU) port cities 32 EU port cities Thessaloniki

Delft, Hamburg, Rotterdam, Sines, Trelleborg Koper, Livorno

Amsterdam, Antwerp, Malmö Barcelona, Genoa, Marseille, Thessaloniki

Partner ports

30  César Ducruet and Theo Notteboom

Valencia Kołobrzeg

Trieste

2020–2022

2019–2023 2020–

Valencia

2019–2022

Assessment of Climate Change in Ports of Southwestern Europe (ECCLIPSE) SUStainable PORTs (SUSPORT)

Green C Ports Operational Platform managing a f leet of semi-autonomous drones exploiting GNSS [Global Navigation Satellite System] high Accuracy and Authentication to improve Security & Safety in port areas (PASSport)

– –

2016–2022 2018–2022

CLean INland SHipping (CLINSH) Towards a Green and Sustainable Ecosystem for the EU Port of the Future (PortForward)

Valencia

2018–2021

Port IoT for Environmental Leverage (PIXEL)

Ancona, Central Adriatic, Chioggia, Dubrovnik, Ploče, Ravenna, Rijeka, Southern Adriatic, Split, Venice, Zadar Chioggia, Piraeus, Venice Hamburg, Le Havre, Ravenna, Valencia

Bordeaux, Gorizia, Monfalcone, Piraeus, Rijeka, Thessaloniki, Trieste Antwerp, North Sea Port Kristiansand, Magdeburger Hafen, Northern Tyrrhenian Sea Port Authority System, Ports de Balears, Vigo Aveiro, Bordeaux

A systematic and critical review of port system research  31

32  César Ducruet and Theo Notteboom

The collaboration network of port system scholars

80

4.5

70

4

60

3.5 3

50

2.5

40

2

30

1.5

20

1

10

0.5

0 1960

1970

1980

1990

2000

Number

Figure 1.4 Collaboration dynamics, 1963–2022

2010 Mean

2020

0 2030

Mean authors by article

Number of authors

A total of 390 authors base the corpus of port system studies. The analysis of their collaborations is another way to estimate the cohesiveness of the field. Along with the number of publications, the number of authors has increased tremendously over the study period (Figure 1.4). What is more relevant to the analysis of collaborations is the parallel increase of the average number of authors by article, from 1.5–2 in the early 2000s to 3–4 in the last years. The extent to which such collaborations form an interconnected network is verified in Figure 1.5. We observe that the majority of collaborations are the product of isolated groups of scholars, on the basis of one single article. In the center of the graph, however, five larger components emerge. Four of them are marked by a strong geographic and/or cultural proximity among scholars, centered on Yeo (South Korea and Vietnam), Ferrari and Parola (Italy), Laxe and Castillo-Manzano (Spain), and Grifoll (Spain and China). The largest component is the “core” of the system by the number of authors and publications. Notteboom appears as the polarizing “hub” connected to most of the different parts of the system directly or indirectly, regardless of a specific geographic logic. Secondary hubs constitute the rest of the backbone, each of them being in an intermediary position between Notteboom and a subset of authors, often based on specific proximities: Ducruet and Europe/ Asia, Ng and Europe/Asia, Rodrigue and Slack for Europe/North America, and Wilmsmeier-Monios with a European/Latin American coverage. Four authors play a strategic role at the edge of the network, connecting intermediate hubs with Asia, of which mostly China: Wang C., Wang J.J., Wang L., and Lau, together with Yang and Xu H., to some extent. It is also important to note that hub authors have more publications than other authors. However, the bridge role of certain authors is not ref lected

A systematic and critical review of port system research  33

Figure 1.5 The co-authorship network of port system studies

in their publication output. Ng is the third largest author by betweenness centrality after Notteboom and Ducruet, which is the amount of inf luence a node has over the f low of information in a graph. He is followed by Wang C. and Rodrigue. Without these five hub authors, the network would be nearly disconnected and reduced to small groups, as in the rest of the network.

Conclusion The review proposed in this chapter allows us to conclude that port system research is a buoyant field of analysis. Publications about port systems in peer-reviewed international journals are on the rise. Evolutionary models of the 1960s and 1980s, notwithstanding adjustments, updates, and critiques, continue to be used and diffused today. In parallel, port system research is constantly evolving, by exploring new concepts and methodologies, and strengthening its ties with the wider academic sphere. Yet, several questions must be asked when taking some distance from this analysis. Given that constructing a corpus is necessarily pragmatic, one may wonder whether the selected articles truly represent port system research. One first reason is that any analysis of port traffic distribution or maritime/hinterland connectivity could potentially belong to the corpus, without necessarily mentioning port systems. The works of De and Park (2003, 2004) and Ding and Teo (2010) on world container throughput concentration are typical examples of the kind. Conversely, numerous works use the concept of port system without referring to any model, in the ordinary sense of the term. The relational perspective is not present in many of the cited works. As such, the boundary between port system and non-port system research is relatively fuzzy.

34  César Ducruet and Theo Notteboom

Another reason is that entire strands of port research were left aside from our analysis despite their relevance and closeness with the topic. Port choice, port selection, shipping network design, and cargo routing dynamics imply that several ports are in competition and thus can be considered as a system. The same applies to port groups within a certain area, like port or maritime regions, ranges, façades, seaboards, basins, and seas, as geographic proximity is in itself a relational factor (see Rodrigue, 2022). Since the majority of port system studies are done at the national level, any analysis of port development in a given country could be considered as a port system study. To avoid the risk of overestimating the importance of port system research however, we restricted the corpus to explicit references to a port system and to models. As a synthesis, we propose in Figure 1.6 a general framework to study the various aspects of port systems. This multilevel perspective, with reference to Robinson (1976), implies that certain structures and dynamics at one level can have an inf luence on other levels. As a spatial model, it does not include elements of policy and governance, trade and economic wealth. Still, it assembles dispersed conceptions of port systems, between land and sea, local and global, hierarchy and specialization, morphology and urbanization. Multilevel empirical applications remain scarce, however, in port studies. The analysis of port choice and competition may include the various elements of the port triptych, but empirical applications usually include a limited number of ports (see Zondag et al., 2010). An analysis of ship turnaround time at about 2,300 container ports showed that city size, voyage delays at sea, maximum ship size, and upstream location increase it, while degree centrality, national GDP per capita, the number of vessel calls, and island location have the opposite effect (Ducruet and Itoh, 2022). Further research on port systems may follow several pathways. While ports have been seen as complex adaptative systems already (Vonck and Notteboom, 2016), the complex systems framework has been well applied to cities but not to ports. This would allow, for instance, the analysis of coevolution, which has been little explored in a port context (Gerritts, 2011). Simulation models, despite earlier attempts (Kuby et al., 1991; Parola and Sciomachen, 2005; Tavasszy et al., 2011), remain rarely used, and most quantitative analyses are fairly descriptive. If well calibrated, such models have a great potential to forecast past and future changes in traffic distribution, network formation, and hinterland coverage, with the possibility to investigate the impact of changes in one port on other ports (e.g., policies, investments, and strategies). Applications to cities and other transport systems have been proposed already, using generative network models, gravity models, and complex networks among other methods. Very recently, simulation experiments were proposed in the realm of maritime networks to understand their evolutionary processes (Kosowska-Stamirowska, 2020; Kanrak and Nguyen, 2021). Long-term analyses of maritime networks and port systems are becoming more frequent, thanks to the digitization of historical shipping records (see

A systematic and critical review of port system research  35

Figure 1.6 A multilevel modeling of ports and port systems Source: Adapted from Ducruet (2005)

Ducruet, 2017; Olukoju and Castillo, 2020), but more research is needed to check whether similarities emerge among dispersed case studies. Since ports are more than cargo transshipment nodes but elements in value-driven supply chains (Robinson, 2015) and port cities in production networks (Jacobs et al., 2010), “The ultimate goal is to be able to analyze the network of networks – sea-land, air and telecommunications – in an allencompassing study of logistics systems to address the vulnerability of gateways serving multimodal corridors” (Rimmer, 2015). This would imply that the concept of port system should be extended to a multilayered system where ports are only one element. Research in such ways is already active, looking at air-sea and sea-land connectivity for instance (see a recent review in Ducruet and Guerrero, 2022). This perspective supports the idea that port systems are more than ports, just like “ports are more than piers” (Notteboom, 2006).

36  César Ducruet and Theo Notteboom

References Berry B.J.L. (1964) Cities as systems within systems of cities. Papers in Regional Science, 13(1): 146–163. Bertalanffy L. von (1950) An outline of general system theory. British Journal for the Philosophy of Science, 1: 114–129. Biermann F., Wedemeier J. (2016) Hamburg’s port position: Hinterland competition in Central Europe from TEN-T corridor ports. HWWI Research Paper, no. 175. Bird J.H. (1963) The Major Seaports of the United Kingdom. London: Hutchinson. Bird J.H. (1984) Seaport development: Some questions of scale. In: Hoyle B.S., Hilling D. (Eds.), Seaport Systems and Spatial Change, Chichester: Wiley, pp. 21–41. Brooks M.R., Cullinane K.P.B. (2006) Devolution, Port Governance and Port Performance. Elsevier. Brooks M.R., Cullinane K.P.B., Pallis A.A. (2017) Revisiting port governance and port reform: A multi-country examination. Research in Transportation Business and Management, 100(22): 1–10. Chen L., Xu X., Zhang P., Zhang X. (2018) Analysis on port and maritime transport system researches. Journal of Advanced Transportation, https://doi.org/10.1155/ 2018/6471625 Ding D., Teo C.P. (2010) World container port throughput follows lognormal distribution. Maritime Policy and Management, 37(4): 401–426. Ducruet C. (2005) Structures et dynamiques spatiales des villes portuaires: Du local au Mondial. Mappemonde, 77(1), https://hal.archives-ouvertes.fr/halshs00459475/ Ducruet C. (2017) Advances in Shipping Data Analysis and Modeling. Tracking and Mapping Maritime Flows in the Age of Big Data. London & New York: Routledge. Ducruet C. (2020) The geography of maritime networks: A critical review. Journal of Transport Geography, 88: 102824. Ducruet C., Guerrero D. (2022) Inland cities, maritime gateways, and international trade. Journal of Transport Geography, 104: 103433. Ducruet C., Itoh H. (2022) Spatial network analysis of container port operations: The case of ship turnaround times. Networks and Spatial Economics, https://doi. org/10.1007/s11067-022-09570-z Ducruet C., Notteboom T.E., de Langen P.W. (2009) Revisiting inter-port relationships under the New Economic Geography research framework. In: Notteboom T.E., Ducruet C., de Langen P.W. (Eds.), Ports in Proximity: Competition and Coordination among Adjacent Seaports, Ashgate, pp. 11–27. Ducruet C., Panahi R., Ng A.K.Y., Jiang C., Afenyo M. (2019) Between geography and transport: A scientometric analysis of port studies in Journal of Transport Geography. Journal of Transport Geography, 81: 102527. Ducruet C., Lo Prete M., Graef V., Wabnitz K., Ben Hakoun E. (2023) Environment and health in European seaport cities. World Health Organisation Office for Europe, Bonn (forthcoming). Frémont A., Soppé M. (2007) Northern European range: Shipping line concentration and port hierarchy. In: Wang J.J., Olivier D., Notteboom T.E., Slack B. (Eds.), Ports, Cities, and Global Supply Chains, Ashgate, Aldershot, pp. 105–120. Garcia-Alonso L., Martinez-Pardo A., Vallejo-Pinto J.A. (2016) Analysis of the spatial development of the hinterland of ports: A case study. International Journal of Shipping and Transport Logistics, 8(2): 111–128.

A systematic and critical review of port system research  37 Gerritts L. (2011) A coevolutionary revision of decision making processes: An ­a nalysis of port extensions in Germany, Belgium and the Netherlands. Public Administration Quarterly, 35(3): 309–341. Guerrero D., Montes C.P. (2021) Hinterland. In: The International Encyclopedia of Geography: People, the Earth, Environment, and Technology, Wiley. Helmick J.S. (1994) Concentration and Connectivity in the North Atlantic Liner Port Network, 1970–1990. Unpublished PhD Dissertation, Miami: University of Miami. Hoyle B.S. (1989) The port-city interface: Trends, problems and examples. Geoforum, 20(4): 429–435. Hoyle B., Hilling D. (1984) Seaport Systems and Spatial Change. Technology, Industry, and Development Strategies. Wiley. Huo W., Zhang W., Chen P.S.L. (2018) Recent development of Chinese port cooperation strategies. Research in Transportation Business and Management, 26: 67–75. Jacobs W., Ducruet C., de Langen P.W. (2010) Integrating world cities into production networks: The case of port cities. Global Networks, 10(1): 92–113. Joly O. (1999) La Structuration des Réseaux de Circulation Maritime. Unpublished PhD dissertation in Territorial Planning, Le Havre University: CIRTAI. Kanrak M., Nguyen H.O. (2021) Structure, characteristics and connectivity analysis of the Asian-Australasian cruise shipping network. Maritime Policy and Management, https://doi.org/10.1080/03088839.2021.1914876 Kosowska-Stamirowska S. (2020) Network effects govern the evolution of maritime trade. Proceedings of the National Academy of Sciences, 117(23): 12719–12728. Lau Y.Y., Ducruet C., Ng A.K.Y., Fu X. (2017) Across the waves: A bibliometric analysis of container shipping research since the 1960s. Maritime Policy and Management, 44(6): 667–684. Lemarchand A., Joly O. (2009) Regional integration and maritime range. In: ­Notteboom T.E., Ducruet C., de Langen P.W. (Eds.), Ports in Proximity: Competition and Coordination among Adjacent Seaports, Ashgate, pp. 87–99. Ma Q., Jia P., She X., Haralambides H., Kuang H. (2021) Port integration and regional economic development: Lessons from China. Transport Policy, 110: 430–439. Merk O., Notteboom T.E. (2015). Port hinterland connectivity. Discussion paper no. 2015-13, International Transport Forum, OECD: Paris, May 2015. Moura T.G.Z., Garcia-Alonso L., Salas-Olmedo M.H. (2017) Delimiting the scope of the hinterland of ports: Proposal and case study. Journal of Transport Geography, 65: 35–43. Mueller M.A., Wiegmans B., Van Duin J.H.R. (2020) The geography of container port choice: Modelling the impact of hinterland changes on port choice. Maritime Economics & Logistics, 22(1): 26–52. Munim Z.H., Saeed N. (2019) Seaport competitiveness research: The past, present and future. International Journal of Shipping and Transport Logistics, 11(6): 533–557. Ng A.K.Y., Ducruet C. (2014) The changing tides of port geography (1950–2012). Progress in Human Geography, 38(6): 785–823. Ng A.K.Y., Ducruet C., Jacobs W., Monios J., Notteboom T.E., Rodrigue J.P., Slack B., Tam K.C., Wilmsmeier G. (2014) Port geography at the crossroads with human geography: Between f lows and spaces. Journal of Transport Geography, 41: 84–96. Notteboom T.E. (2006) Ports are More Than Piers: Liber Amicorum Presented to Prof. Dr. Willy Winkelmans. Antwerpen: De Lloyd.

38  César Ducruet and Theo Notteboom Notteboom T.E., Ducruet C., de Langen P.W. (2009) Ports in Proximity: Competition and Coordination among Adjacent Seaports, Ashgate. Notteboom T.E., Knatz G., Parola F. (2018) Port co-operation: Types, drivers and impediments. Research in Transportation Business and Management, 26: 1–4. Notteboom T.E., Pallis A.A., de Langen P.W., Papachristou A.A. (2013) Advances in port studies: The contribution of 40 years Maritime Policy & Management. Maritime Policy and Management, 40(7): 636–653. Notteboom T.E., Yang Z. (2017) Port governance in China since 2004: Institutional layering and the growing impact of broader policies. Research in Transportation Business and Management, 22: 184–200. Notteboom T.E., Yang D., Xu H. (2020) Container barge network development in inland rivers: A comparison between the Yangtze River and the Rhine River. Transportation Research Part A, 132: 587–605. Nystuen J.D., Dacey M.F. (1961) A graph theory interpretation of nodal regions. Papers in Regional Science, 7(1): 29–42.Olukoju A., Castillo D.H. (2020) African Seaports and Maritime Economics in Historical Perspective. Palgrave Studies in Maritime Economics. ­ Overman H.G., Winters L.A. (2005) The port geography of UK international trade. Environment and Planning A, 37(10): 1751–1768. Pallis A.A., Vitsounis T.K., de Langen P.W., Notteboom T.E. (2011) Port economics, policy and management: Content classification and survey. Transport Reviews, 31(4): 445–471. Pounds N.J.G. (1947) Port and outport in North-West Europe. The Geographical ­Journal, 109(4–6): 216–228. Pumain D. (2004) Système. Hypergeo, 29 April, https://hypergeo.eu/systeme/ Rimmer P.J. (2015) Foreword. In: Ducruet C. (Ed.), Maritime Networks. Spatial Structures and Time Dynamics, London and New York: Routledge, pp. xxi–xxii. Robinson R. (1968) Spatial Structuring of Port-Linked Flows; The Port of Vancouver, Canada, 1965. Unpublished PhD dissertation in Geography, University of British Columbia. Robinson R. (1976) Modelling the port as an operational system: A perspective for research. Economic Geography, 52(1): 71–86. Robinson R. (2015) Afterword. In: Ducruet C. (Ed.), Maritime Networks. Spatial Structures and Time Dynamics, London and New York: Routledge, pp. 374–377. Rodrigue J.P. (2022) The geography of maritime ranges: Interfacing global maritime shipping networks with hinterlands. Geojournal, 87: 1231–1244. Roso V., Woxenius J., Lumsden K. (2009) The dry port concept: connecting container seaports with the hinterland. Journal of Transport Geography, 17(5): 338–345. Sanders L. (2005) Spatial system. Hypergeo, 12 January, https://hypergeo.eu/ spatial-system/?lang=en Vallega A. (1986) From ports to the port system: A geographical explanation. In: Charlier J. (Eds.), Ports et Mers, Mélanges Maritimistes Offerts à André Vigarié, Caen: Paradigme, pp. 147–155. Veldman S.J., Bückmann E.H. (2003) A model on container port competition: An application for the West European container hub-ports. Maritime Economics & Logistics, 5(1): 3–22. Veldman S.J., Bückmann E.H., Saitua R.N. (2005) River depth and container port market shares: The impact of deepening the Scheldt river on the west European container hub-port market shares. Maritime Economics & Logistics, 7(4): 336–355.

A systematic and critical review of port system research  39 Vigarié A. (1964) Les Grands Ports de Commerce de la Seine au Rhin: Leur Evolution Devant l’Industrialisation des Arrière-Pays. Paris: S.A.B.R.I. Vigarié A. (1979) Ports de Commerce et Vie Littorale. Paris: Hachette. Wan S., Luan W., Ma Y., Haralambides H. (2020) On determining the hinterlands of China’s foreign trade container ports. Journal of Transport Geography, 85: 102725. Wang C., Ducruet C., Wang W. (2015) Port integration in China: Temporal pathways, spatial patterns and dynamics. Chinese Geographical Science, 25(5): 612–628. Wang S.B., Peng X.H. (2021) Knowledge mapping of port logistics in the recent 20 Years: A bibliometric analysis via CiteSpace. Maritime Policy and Management, https://doi.org/10.1080/03088839.2021.1990429 Wilmsmeier G., Monios J., Lambert B. (2011) The directional development of intermodal freight corridors in relation to inland terminals. Journal of Transport Geography, 19: 1379–1386. Woo S.H., Pettit S.J., Kwak D.W., Beresford A.K.C. (2011) Seaport research: A structured literature review on methodological issues since the 1980s. Transportation Research Part A, 45: 667–685. Xu X., Wang H., WU Y., Yi W. (2021) Bibliometric analysis on port and shipping researches in scope of management science. Asia-Pacific Journal of Operational Research, 38(3): 2140027. Yang Z., Guo L., Lian F. (2019) Port integration in a region with multiport gateways in the context of industrial transformation and upgrading of the port. Transportation Research Part E, 122: 231–246. Yang J., Luo M., Ji A. (2016) Analyzing the spatial–temporal evolution of a gateway’s hinterland: A case study of Shanghai, China. Transportation Research Part E, 95: 355–367. Yang D., Notteboom T.E., Zhou X. (2021) Spatial, temporal and institutional characteristics of entry strategies in inland container terminals: A comparison between Yangtze River and Rhine River. Journal of Transport Geography, 90: 102928. Zondag B., Bucci P., Gützkow P., de Jong G. (2010) Port competition modeling including maritime, port, and hinterland characteristics. Maritime Policy and Management, 37(3): 179–194.

Appendix 1.1

Corpus of port system articles, 1963–2022

1 Acciaro M., Bardi A., Cusano M.I. (2017) Contested port hinterlands: An empirical survey on Adriatic seaports. Case Studies on Transport Policy, 5(2): 342–350. 2 Airriess C. (1995) Port-centered transport development in colonial North Sumatra. Indonesia, 59: 65–91. 3 Bacon J.R., Halbrendt C.K., Gempesaw C.M. (1990) Food trade f lows of the Delaware River port system. Journal of Food Distribution Research, 21(2): 39–40. 4 Balic K., Žgaljic D., Ukic Boljat H. (2022) The port system in addressing sustainability issues—A systematic review of research. Journal of Marine Science and Engineering, 10(8): 1048. 5 Beškovnik B., Twrdy E. (2015) Developing regional approach for transport industry: The role of port system in the Balkans. Transport, 30(4): 437–447. 6 Blanco B., Sánchez L., Gutierrez C. (2016) Port system of the Spanish South Peninsular coast side. Journal of Maritime Research, 13(1): 15–24. 7 Blanco B., Sánchez L., Pellón M. (2015) Port system of the Spanish Mediterranean coast side. Journal of Maritime Research, 12(2): 35–44. 8 Bobrovitch D. (1982) Decentralised planning and competition in a national multi-port system. Journal of Transport Economics and Policy, 16(1): 31–42. 9 Bonilla M., Casarús S.R., Medal A. (2004) An efficiency analysis with tolerance of the Spanish port system. International Journal of Transport Economics, 31(3): 1000–1022. 10 Broeze F.J.A. (1996) The ports and port system of the Asian Seas: An overview with historical perspective from c1750. The Great Circle: Journal of the Australian Association for Maritime History, 18(2): 73–96. 11 Brunt B. (2000) Ireland’s seaport system. Tijdschrift voor Economische en Sociale Geografie, 91(2): 159–175. 12 Cariou P., Notteboom T.E. (2022) Implications of COVID-19 on the US container port distribution system: Import cargo routing by Walmart and Nike. International Journal of Logistics Research and Applications, https://doi.org/10.1080/ 13675567.2022.2088708 13 Castillo D., Valdaliso J.M. (2017) Path dependence and change in the Spanish port system in the long run (1880–2014): An historical perspective. International Journal of Maritime History, 29(3): 569–596. 14 Castillo-Manzano J.I., Castro-Nuño M., González-Laxe F. (2009) Low-cost port competitiveness index: Implementation in the Spanish port system. Marine Policy, 33(4): 591–598.

A systematic and critical review of port system research  41 15 Castillo-Manzano J.I., Castro-Nuño M., González-Laxe F. (2014) An analysis of the determinants of cruise traffic: An empirical application to the Spanish port system. Transportation Research Part E, 66: 115–125. 16 Castillo-Manzano J.I., Castro-Nuño M., González-Laxe F. (2018) Legal reform and the devolution of the Spanish Port System: An econometric assessment. Utilities Policy, 50: 73–82. 17 Charlier J. (1988) Structural change in the Belgian port system, 1980–1986. Maritime Policy and Management, 15(4): 315–326. 18 Charlier J. (1996) The Benelux seaport system. Tijdschrift voor Economische en Sociale Geografie, 87(4): 310–321. 19 Chaudhury B., De P. (1999) Concentration of Indian Port System: 1970–1998. Foreign Trade Review, 34(3–4): 33–46. 20 Chlomoudis C., Styliadis T. (2019) Concentration of container f lows in the port phase: The case of the US West and East Coast port ranges. Issues in Business Management and Economics, 7(1): 1–21. 21 Comtois C. (1999) The integration of China’s port system into global container shipping. GeoJournal, 48(1): 35–42. 22 Comtois C., Dong J. (2007) Port competition in the Yangtze River delta. Asia Pacific Viewpoint, 48(3): 299–311. 23 Cullinane K., Wang Y. (2012) The hierarchical configuration of the container port industry: An application of multiple linkage analysis. Maritime Policy and Management, 39(2): 169–187. 24 Danielis R., Gregori T. (2013) An input-output-based methodology to estimate the economic role of a port: The case of the port system of the Friuli Venezia Giulia Region, Italy. Maritime Economics & Logistics, 15(2): 222–255. 25 de Langen P.W. (2007) Port competition and selection in contestable hinterlands: The case of Austria. European Journal of Transport and Infrastructure Research, 7(1): 1–14. 26 de Oliveira G.F., Schaffar A., Cariou P., Monios J. (2021) Convergence and growth traps in container ports. Transport Policy, 110: 170–180. 27 De P., Park R.K. (2003) Container port system concentration. Transportation Quarterly, 57(4): 69–82. 28 De P., Park R.K. (2004) International container port system concentration: What does it look like. Journal of International Logistics and Trade, 2(1): 95–118. 29 Debrie J. (2012) The West African port system: Global insertion and regional particularities. EchoGéo, (20): https://doi.org/10.4000/echogeo.13070 30 Deng Z., Li Z.F., Zhou Y.T., Chen X., Liang S.S. (2020) Measurement and spatial spillover effects of port comprehensive strength: Empirical evidence from China. Transport Policy, 99: 288–298. 31 Do N.H., Nam K.C., Le Ngoc Q.L. (2011) A consideration for developing a dry port system in Indochina area. Maritime Policy and Management, 30(1): 1–9. 32 Ducruet C. (2008) Hub dependence in constrained economies: The case of North Korea. Maritime Policy and Management, 35(4): 377–394. 33 Ducruet C. (2013) Network diversity and maritime f lows. Journal of Transport Geography, 30: 77–88. 34 Ducruet C. (2017) Multilayer dynamics of complex spatial networks: The case of global maritime f lows (1977–2008). Journal of Transport Geography, 60: 47–58. 35 Ducruet C. (2020) Revisiting urban hierarchy and specialization from a maritime perspective. Maritime Policy and Management, 47(3): 371–387.

42  César Ducruet and Theo Notteboom 36 Ducruet C. (2022) Port specialization and connectivity in the global maritime network. Maritime Policy and Management, 49(1): 1–17. 37 Ducruet C., Cuyala S., El Hosni A. (2018) Maritime networks as systems of cities: The long-term interdependencies between global shipping f lows and urban development (1890–2010). Journal of Transport Geography, 66: 340–355. 38 Ducruet C., Itoh H. (2022) The spatial determinants of innovation diffusion: Evidence from global shipping networks. Journal of Transport Geography, 101: 103358. 39 Ducruet C., Jo J.C. (2008) Coastal cities, port activities and logistic constraints in a socialist developing country: The case of North Korea. Transport Reviews, 28(1): 35–59. 40 Ducruet C., Koster H.R.A., van der Beek D.J. (2010) Commodity variety and seaport performance. Regional Studies, 44(9): 1221–1240. 41 Ducruet C., Lee S.W. (2006) Frontline soldiers of globalisation: Port–city evolution and regional competition. GeoJournal, 67(2): 107–122. 42 Ducruet C., Lee S.W., Ng A.K.Y. (2010) Centrality and vulnerability in liner shipping networks: Revisiting the Northeast Asian port hierarchy. Maritime ­Policy and Management, 37(1): 17–36. 43 Ducruet C., Notteboom T.E. (2012) The worldwide maritime network of container shipping: Spatial structure and regional dynamics. Global networks, 12(3): 395–423. 44 Ducruet C., Notteboom T.E. (2022) Revisiting port system delineation through an analysis of maritime interdependencies among seaports. GeoJournal, 87: 1831–1859. 45 Ducruet C., Roussin S., Jo J.C. (2009) Going West? Spatial polarization of the North Korean port system. Journal of Transport Geography, 17(5): 357–368. 46 Ducruet C., Wang L. (2018) China’s global shipping connectivity: Internal and external dynamics in the contemporary era (1890–2016). Chinese Geographical Science, 28(2): 202–216. 47 Ducruet C., Zaidi F. (2012) Maritime constellations: A complex network approach to shipping and ports. Maritime Policy and Management, 39(2): 151–168. 48 Escobar M.C.M. (2021) Roman ports in the lower Tiber valley: Computational approaches to reassess Rome’s port system. Papers of the British School at Rome, 1–30, doi:10.1017/S0068246221000271 49 Esparza A., Cerbán M.M., Piniella F. (2017) State-owned Spanish port system oversizing: An analysis of maximum operational capacity. Maritime Policy and Management, 44(8): 995–1011. 50 Esteve-Pérez J., Gutiérrez-Romero J.E. (2022) Assessment of the dynamics of concentration and competitive positions of the Baltic cruise port system. European Transport Research Review, 14(1): 1–13. 51 Fan F., Zhang X., Yang W. (2021) Spatiotemporal evolution of China’s ports in the international container transport network under upgraded industrial structure. Transportation Journal, 60(1): 43–69. 52 Feng H., Grifoll M., Yang Z. (2020) Visualization of container throughput evolution of the Yangtze River Delta multi-port system: The ternary diagram method. Transportation Research Part E, 142: 102039. 53 Feng H., Grifoll M., Zheng P. (2019) From a feeder port to a hub port: The evolution pathways, dynamics and perspectives of Ningbo-Zhoushan port (China). Transport Policy, 76: 21–35.

A systematic and critical review of port system research  43 54 Feng H., Grifoll M., Zheng P. (2021) Evolution and container traffic prediction of Yangtze River Delta multi-port system (2001–2017). International Journal of Shipping and Transport Logistics, 13(1): 44–57. 55 Feng L. (2013) Regression analysis on the hinterland connection of peripheral ports in the China Bohai Sea regions. International Journal of Advances in Management Science, 2(4): 159–162. 56 Feng L., Notteboom T.E. (2013) Peripheral challenge by small and medium sized ports (SMPs) in multi-port gateway regions: The case study of northeast of China. Polish Maritime Research, 20: 55–66. 57 Feng X., Jiang L., Zhang Y. (2011) Optimization of capacity of ports within a regional port system. Transportation Research Record, 2222(1): 10–16. 58 Feng X., Zhang Y., Li Y. (2013) A location-allocation model for seaport-dry port system optimization. Discrete Dynamics in Nature and Society, 2013, https:// doi.org/10.1155/2013/309585 59 Ferrari C., Parola F., Benacchio M. (2008) Network economies in liner shipping: The role of home markets. Maritime Policy and Management, 35(2): 127–143. 60 Ferrari C., Parola F., Gattorna E. (2011) Measuring the quality of port hinterland accessibility: The Ligurian case. Transport Policy, 18(2): 382–391. 61 Franc P., Van der Horst M. (2010) Understanding hinterland service integration by shipping lines and terminal operators: A theoretical and empirical analysis. Journal of Transport Geography, 18(4): 557–566. 62 Fraser D.R., Notteboom T.E. (2012) Gateway and hinterland dynamics: The case of the Southern African container port system. African Journal of Business Management, 44(6): 10807–10825. 63 Fraser D.R., Notteboom T.E., Ducruet C. (2016) Peripherality in the global container shipping network: The case of the Southern African container port system. GeoJournal, 81(1): 139–151. 64 Galvão C.B., Robles L.T., Guerise L.C. (2013) The Brazilian seaport system: A post-1990 institutional and economic review. Research in Transportation Business and Management, 8: 17–29. 65 Garcia-Alonso L., Sanchez-Soriano J. (2010) Analysis of the evolution of the inland traffic distribution and provincial hinterland share of the Spanish port system. Transport Reviews, 30(3): 275–297. 66 Ghosh B., De P. (2001) Indian ports and globalisation: Grounding economics in geography. Economic and Political Weekly, 36(34): 3271–3283. 67 Gilb C.L. (1992) Factors contributing to change in the European port system. Landscape and Urban Planning, 22(2–4): 205–217. 68 González A.R., González-Cancelas N., Serrano B.M. (2020) Smart ports: Ranking of Spanish port system. World Scientific News, 144: 1–12. 69 González-Laxe F., Bermúdez F.M., Palmero F.M. (2016) Sustainability and the Spanish port system. Analysis of the relationship between economic and environmental indicators. Marine Pollution Bulletin, 113(1–2): 232–239. 70 Gouvernal E., Debrie J., Slack B. (2005) Dynamics of change in the port system of the western Mediterranean. Maritime Policy and Management, 32(2): 107–121. 71 Grifoll M., Karlis T., Ortego M.I. (2018) Characterizing the evolution of the container traffic share in the Mediterranean Sea using hierarchical clustering. Journal of Marine Science and Engineering, 6(4): https://doi.org/10.3390/ jmse6040121

44  César Ducruet and Theo Notteboom 72 Grushevska K., Notteboom T.E. (2014) An economic and institutional analysis of multi-port gateway regions in the Black Sea Basin. Journal of International Logistics and Trade, 12(2): 3–35. 73 Grushevska K., Notteboom T.E. (2016) The development of river-based intermodal transport: the case of Ukraine. Journal of International Logistics and Trade, 14(3): 182–199. 74 Guerrero D. (2014) Deep-sea hinterlands: Some empirical evidence of the spatial impact of containerization. Journal of Transport Geography, 35: 84–94. 75 Guerrero D., Rodrigue J.P. (2014) The waves of containerization: Shifts in global maritime transportation. Journal of Transport Geography, 34: 151–164. 76 Guy E., Urli B. (2006) Port selection and multicriteria analysis: An application to the Montreal-New York alternative. Maritime Economics & Logistics, 8(2): 169–186. 77 Haezendonck E., Coeck C., Verbeke A. (2000) The competitive position of seaports: Introduction of the value added concept. International Journal of ­Maritime Economics, 2(2): 107–118. 78 Haezendonck E., Coeck C., Verbeke A. (2006) Strategic positioning analysis for seaports. Research in Transportation Economics, 16: 141–169. 79 Hall P.V. (2004) Mutual specialisation, seaports and the geography of automobile imports. Tijdschrift voor Economische en Sociale Geografie, 95(2): 135–146. 80 Hall, Peter V, Jacobs, Wouter, (2010) Shifting proximities: The maritime ports sector in an era of global supply chains. Regional studies, 44(9): 1103–1115. 81 Hayuth Y. (1981) Containerization and the load center concept. Economic Geography, 57(2): 160–176. 82 Hayuth Y. (1988) Rationalization and deconcentration of the US container port system. The Professional Geographer, 40(3): 279–288. 83 Hayuth Y. (1991) Load centering competition and modal integration. Coastal Management, 19(3): 297–311. 84 He D., Sun Z., Gao P. (2019) Spatial–temporal evolution of the port–hinterland relationship: A case study of the Midstream Yangtze River, China. Growth and Change, 50(3): 1043–1061. 85 Heilig L., Lalla-Ruiz E., Voß S. (2017) Digital transformation in maritime ports: Analysis and a game theoretic framework. Netnomics, 18(2): 227–254. 86 Hidalgo-Gallego S., Núñez-Sánchez R., Coto-Millán P. (2022) Port allocative efficiency and port devolution: A study for the Spanish port authorities (1992– 2016). Maritime Policy and Management, 49(1): 39–61. 87 Hilling D. (1969) The evolution of the major ports of West Africa. The ­G eographical Journal, 135(3): 365–378. 88 Hilling D. (1977) The evolution of a port system – the case of Ghana. Geography, 62(2): 97–105. 89 Hilling D. (1989) Technology and the changing port system of England and Wales. Geography, 74(2): 117–127. 90 Hoyle B. (1999) Port concentration, inter-port competition and revitalization: The case of Mombasa, Kenya. Maritime Policy and Management, 26(2): 161–174. 91 Hoyle B., Charlier J. (1995) Inter-port competition in developing countries: An East African case study. Journal of Transport Geography, 3(2): 87–103.

A systematic and critical review of port system research  45 92 Huang D., Grifoll M., Feng H. (2021) Characterizing the evolution of the Yangtze River Delta multi-port system using compositional data techniques. Maritime Policy and Management, https://doi.org/10.1080/03088839.2021.1972175 93 Ignaccolo M., Inturri G., Giuffrida N. (2020) Sustainability of freight transport through an integrated approach: The case of the Eastern Sicily port system. Transportation Research Procedia, 45: 177–184. 94 Iyer K.C., Nanyam V.P.S.N. (2021) Concentration analysis of container terminals in India. Maritime Transport Research, 2: 100037. 95 Jacobs W., Notteboom T.E. (2011) An evolutionary perspective on regional port systems: The role of windows of opportunity in shaping seaport competition. Environment and Planning A, 43(7): 1674–1692. 96 Jaja C. (2009) Port development in Nigeria: Trends and patterns. Journal of Transportation Security, 2(4): 107–119. 97 Jeevan J., Harun M., Abdullah W.M.W., Othman M.R., Mohd Salleh N.H., Caesar L.D. (2021) The belt and road initiative: A pragmatic constituent for the growth of Malaysian seaport system. World Review of Intermodal Transport Research, 10(3): https://www.inderscienceonline.com/doi/pdf/10.1504/ WRITR.2021.117668 98 Jiang C., Fu X., Ge Y.E. (2021) Vertical integration and capacity investment in a two-port system. Transportmetrica A: Transport Science, 17(4): 1431–1459. 99 Jiang Z., Pi C., Zhu H. (2022) Temporal and spatial evolution and inf luencing factors of the port system in Yangtze River Delta Region from the perspective of dual circulation: Comparing port domestic trade throughput with port foreign trade throughput. Transport Policy, 118: 79–90. 100 Jin L., Chen J., Sheu J.B. (2022) Impacts of national strategies on gateway ports: An empirical study in the Bohai Rim. Transport Policy, 117: 1–11. 101 John A., Yang Z., Riahi R., Wang J. (2016) A risk assessment approach to improve the resilience of a seaport system using Bayesian networks. Ocean ­Engineering, 111: 136–147. 102 Jones J.R, Qu L., Casavant K.L. (1995) A spatial equilibrium port cargo projection model. Maritime Policy and Management, 22(1): 63–80. 103 Jung P.H., Thill J.C. (2022) Sea-land interdependence and delimitation of port hinterland-foreland structures in the international transportation system. J­ ournal of Transport Geography, 99: 103297. 104 Kim S., Kang D., Dinwoodie J. (2016) Competitiveness in a multipolar port system: Striving for regional gateway status in Northeast Asia. The Asian Journal of Shipping and Logistics, 32(2): 119–125. 105 Konings R. (2007) Opportunities to improve container barge handling in the port of Rotterdam from a transport network perspective. Journal of Transport Geography, 15(6): 443–454. 106 Kozan E. (1994) Analysis of the economic effects of alternative investment decisions for seaport systems. Transportation Planning and Technology, 18(3): 239–248. 107 Krivoshchekov P., Pham T.Y., Yeo G.T. (2018) Seaport concentration and competition development in the Arctic region of Russia along the North Sea Route. Journal of Digital Convergence, 16(4): 49–54. 108 Kuby M., Ratick S., Osleeb J. (1991) Modeling US coal export planning decisions. Annals of the Association of American Geographers, 81(4): 627–649.

46  César Ducruet and Theo Notteboom 109 Kuby M., Reid N. (1992) Technological change and the concentration of the US general cargo port system: 1970–88. Economic Geography, 68(3): 272–289. 110 Kurt I., Aymelek M., Boulougouris E. (2021) Operational cost analysis for a container shipping network integrated with offshore container port system: A case study on the West Coast of North America. Marine Policy, 126: 104400. 111 Lam J.S.L., Yap W.Y. (2011) Container port competition and complementarity in supply chain systems: Evidence from the Pearl River Delta. Maritime Economics & Logistics, 13(2): 102–120. 112 Le Y., Ieda H. (2010) Evolution dynamics of container port systems with a geo-economic concentration index. Asian Transport Studies, 1(1): 46–61. 113 Le Y., Ieda H. (2010) Modified Herfindahl–Hirschman index for measuring the concentration degree of container port systems. Transportation Research Record, 2166(1): 44–53. 114 Lee J.Y., Rodrigue J.P. (2006) Trade reorientation and its effects on regional port systems: The Korea-China link along the Yellow Sea Rim. Growth and Change, 37(4): 597–619. 115 Lee T., Yeo G.T., Thai V.V. (2014) Changing concentration ratios and geographical patterns of bulk ports: The case of the Korean West coast. The Asian Journal of Shipping and Logistics, 30(2): 155–173. 116 Le-Griffin H.D., Griffin M.T., (2010) Managing empty container f lows through short sea shipping and regional port systems. International Journal of Shipping and Transport Logistics, 2(1): 59–75. 117 Li J.B., Oh Y.S. (2010) A research on competition and cooperation between Shanghai port and Ningbo-Zhoushan port. The Asian Journal of Shipping and Logistics, 26(1): 67–91. 118 Li K.X., Luo M., Yang J. (2012) Container port systems in China and the USA: A comparative study. Maritime Policy and Management, 39(5): 461–478. 119 Li S., Haralambides H., Zeng Q. (2022) Economic forces shaping the evolution of integrated port systems-The case of the container port system of China’s Pearl River Delta. Research in Transportation Economics, 94: 101183. 120 Li W., You Z., Cai Z., Sui Y. (2022) Club convergence and allometry in Chinese mainland coastal container ports. Ocean and Coastal Management, 230: 106376. 121 Liu L., Ping H. (2020) Study of the inf luencing factors on development of ports in Guangdong, Hong Kong, and Macao from the perspective of spatial economics. Mathematical Problems in Engineering, 2020: https://doi. org/10.1155/2020/2343860 122 Liu L., Wang K.Y., Yip T.L. (2013) Development of a container port system in Pearl River Delta: Path to multi-gateway ports. Journal of Transport Geography, 28: 30–38. 123 Liu S., Nguyen T.H., Yeo G.T. (2018) The evolution of container port group in Bohai Rim of China. Journal of Digital Convergence, 16(9): 107–114. 124 Liu S., Yeo G.T. (2020) A study on the concentration and deconcentration development for cruise ports in Mediterranean Sea. Journal of Digital Convergence, 18(7): 37–46. 125 López-Bermúdez B., Freire-Seoane M.J., Lesta-Casal E. (2020) Core and comprehensive ports: The new challenge for the development of the Spanish port system. Transportation Research Interdisciplinary Perspectives, 8: 100243. 126 Ma Q., Jia P., She X. (2021) Port integration and regional economic development: Lessons from China. Transport Policy, 110: 430–439.

A systematic and critical review of port system research  47 127 Marcadon J. (1999) Containerisation in the ports of Northern and Western Europe. GeoJournal, 48(1): 15–20. 128 Marti B.E. (1981) Patterns of United States–Canadian maritime container f lows. Maritime Policy and Management, 8(4): 253–259. 129 Marti B.E. (1985) Chilean ports: Commodity specialization and potential for containerization. The Professional Geographer, 37(3): 320–328. 130 Marti B.E. (1988) The evolution of Pacific Basin load centres. Maritime Policy and Management, 15(1): 57–66. 131 Marti B.E. (1990) Geography and the cruise ship port selection process. Maritime Policy and Management, 17(3): 157–164. 132 Marti B.E., Krausse G.H. (1983) Trade route 11: Methods to assess port exchanges of maritime containerized cargo f lows. Ocean Management, 8(4): 317–333. 133 Marušic E., Šoda J., Krcum M. (2020) The three-parameter classification model of seasonal f luctuations in the Croatian nautical port system. Sustainability, 12(12): 5079. 134 Maucec H., Zelenika R., Sever D. (2013) Development rate of maritime container ports: A case of transatlantic ports. Journal of Maritime Research, 10(2): 11–20. 135 McCalla R.J. (1999) From St. John’s to Miami: Containerisation at eastern seaboard ports. GeoJournal, 48(1): 21–28. 136 Medal-Bartual A., Molinos-Senante M., Sala-Garrido R. (2016) Productivity change of the Spanish Port System: Impact of the economic crisis. Maritime Policy and Management, 43(6): 683–705. 137 Medda F., Carbonaro G. (2007) Growth of container seaborne traffic in the Mediterranean basin: Outlook and policy implications for port development. Transport Reviews, 27(5): 573–587. 138 Memon A., Milne M.J., Selsky J.W. (2004) Restructuring governance of New Zealand seaports: Geographical impacts of corporatisation. New Zealand Geographer, 60(2): 15–27. 139 Michael O.O. (2019) An appraisal of containerization in ports of western and eastern Nigeria. International Journal for Traffic and Transport Engineering, 9(2): 188–197. 140 Miyajima M., Kwak K.S. (1989) Economic analysis of interport competition in container cargo: Peripheral ports versus Tokyo Bay ports. Maritime Policy and Management, 16(1): 47–55. 141 Mohamed-Chérif F., Ducruet C. (2016) Regional integration and maritime connectivity across the Maghreb seaport system. Journal of Transport Geography, 51: 280–293. 142 Mohd Salleh N.H., Abd Rasidi N.A.S., Jeevan J. (2020) Lean, agile, resilience and green (LARG) paradigm in supply chain operations: A trial in a seaport system. Australian Journal of Maritime and Ocean Affairs, 12(4): 200–216. 143 Mohd Salleh N.H., Selvaduray M., Jeevan J., Ngah A.H., Zailani S. (2021) Adaptation of industrial revolution 4.0 in a seaport system. Sustainability, 13(19): 10667. 144 Monios J. (2017) Cascading feeder vessels and the rationalisation of small container ports. Journal of Transport Geography, 59: 88–99. 145 Monios J., Wilmsmeier G. (2012) Giving a direction to port regionalisation. Transportation Research Part A, 46(10): 1551–1561.

48  César Ducruet and Theo Notteboom 146 Monios J., Wilmsmeier G. (2013) The role of intermodal transport in port regionalisation. Transport Policy, 30: 161–172. 147 Monios J., Wilmsmeier G. (2016) Between path dependency and contingency: New challenges for the geography of port system evolution. Journal of Transport Geography, 51(2): 247–251. 148 Monios J., Wilmsmeier G., Ng A.K.Y. (2019) Port system evolution–the emergence of second-tier hubs. Maritime Policy and Management, 46(1): 61–73. 149 Montes C.P., Seoane M.J.F., Laxe F.G. (2012) General cargo and containership emergent routes: A complex networks description. Transport Policy, 24: 126–140. 150 Mulder J., Dekker R. (2014) Methods for strategic liner shipping network design. European Journal of Operational Research, 235(2): 367–377. 151 Muravev D., Rakhmangulov A. (2016) Environmental factors’ consideration at industrial transportation organization in the «seaport–dry port» system. Open Engineering, 6(1): 476–484. 152 Navarro-Chávez C.L., Zamora-Tores A.I. (2014) Economic efficiency of the international port system: An analysis through data envelopment. International Business Research, 7(11): 108–116. 153 Nguyen P.N., Woo S.H., Beresford A. (2020) Competition, market concentration, and relative efficiency of major container ports in Southeast Asia. Journal of Transport Geography, 83: 102653. 154 Nguyen T.H., Pham T.Y., Yeo G.T. (2018) Deconcentration pattern of port system: Case in Southern Vietnam. Journal of Digital Convergence, 16(3): 157–167. 155 Nguyen T.V., Nguyen H.P. (2020) A study on forecasting model of container cargo throughput of Vietnam’s seaport. International Journal of e-Navigation and Maritime Economy, 14: 75–88. 156 Notteboom T.E. (1997) Concentration and load centre development in the European container port system. Journal of Transport Geography, 5(2): 99–115. 157 Notteboom T.E. (2006) Container throughput dynamics in the East Asian container port system. Journal of International Logistics and Trade, 4(1): 31–52. 158 Notteboom T.E. (2006) Traffic inequality in seaport systems revisited. Journal of Transport Geography, 14(2): 95–108. 159 Notteboom T.E. (2007) Container river services and gateway ports: Similarities between the Yangtze River and the Rhine River. Asia Pacific Viewpoint, 48(3): 330–343. 160 Notteboom T.E. (2007) Spatial dynamics in the container load centers of the Le Havre-Hamburg range. Zeitschrift für Wirtschaftsgeographie, 51(1): 108–123. 161 Notteboom T.E. (2009) Complementarity and substitutability among adjacent gateway ports. Environment and Planning A, 41(3): 743–762. 162 Notteboom T.E. (2010) Concentration and the formation of multi-port gateway regions in the European container port system: An update. Journal of Transport Geography, 18(4): 567–583. 163 Notteboom T.E. (2010) From multi-porting to a hub port configuration: The South African container port system in transition. International Journal of Shipping and Transport Logistics, 2(2): 224–245. 164 Notteboom T.E. (2011) An application of multi-criteria analysis to the location of a container hub port in South Africa. Maritime Policy and Management, 38(1): 51–79.

A systematic and critical review of port system research  49 165 Notteboom T.E., Rodrigue J.P. (2005) Port regionalization: Towards a new phase in port development. Maritime Policy and Management, 32(3): 297–313. 166 Notteboom T.E., Yang D., Xu H. (2020) Container barge network development in inland rivers: A comparison between the Yangtze River and the Rhine River. Transportation Research Part A, 132: 587–605. 167 Ogundana B. (1972) Oscillating seaport location in Nigeria. Annals of the Association of American Geographers, 62(1): 110–121. 168 Olivier D., Slack B. (2006) Rethinking the port. Environment and Planning A, 38(8): 1409–1427. 169 Padilha F., Ng A.K.Y. (2012) The spatial evolution of dry ports in developing economies: The Brazilian experience. Maritime Economics & Logistics, 14(1): 99–121. 170 Paiva T.M., Santos R.F. (2022) The Brazilian port system: A panorama of informational processes for importing and exporting containerized cargo. Revista Gestão & Tecnologia, 22(2): 175–200. 171 Pan K., Cao Y., Liang S. (2014) New tendency of Chinese container port system: 1998–2010. GeoJournal, 79(3): 373–384. 172 Parola F., Musso E. (2007) Market structures and competitive strategies: The carrier–stevedore arm-wrestling in northern European ports. Maritime Policy and Management, 34(3): 259–278. 173 Parola F., Sciomachen A. (2005) Intermodal container f lows in a port system network: Analysis of possible growths via simulation models. International Journal of Production Economics, 97(1): 75–88. 174 Pedersen P. (2003) Development of freight transport and logistics in Sub-Saharan Africa: Taaffe, Morrill and Gould revisited. Transport Reviews, 23(3): 275–297. 175 Pham T.Y., Jeon J.W., Dang V.L. (2016) A longitudinal analysis of concentration developments for container terminals in Northern Vietnam. The Asian Journal of Shipping and Logistics, 32(3): 157–164. 176 Reina Valle R., Villena Manzanares F. (2013) Concentration of goods traffic in Spanish ports during the period 2000–2009. Regional and Sectoral Economic Studies, 13(2): 59–72. 177 Ribeiro da Silva F. (2017) The slave trade and the development of the Atlantic Africa port system, 1400s–1800s. International Journal of Maritime History, 29(1): 138–154. 178 Rimmer P.J. (1967a) Changes in the ranking of Australian seaports 1951–2— 1961–2. Tijdschrift voor Economische En Sociale Geografie, 58(1): 28–38. 179 Rimmer P.J. (1967b) The changing status of New Zealand seaports, 1853–1960. Annals of the Association of American Geographers, 57(1): 88–100. 180 Rimmer P.J. (1967c) Recent changes in the status of seaports in the New Zealand coastal trade. Economic Geography, 43(3): 231–243. 181 Rimmer P.J. (2007) Port dynamics since 1965: Past patterns, current conditions and future directions. Journal of International Logistics and Trade, 5(1): 75–97. 182 Rimmer P.J., Comtois C. (2009) China’s container-related dynamics, ­1990–2005. GeoJournal, 74(1): 35–50. 183 Rodrigo González A., González-Cancelas N., Molina Serrano B. (2020) Preparation of a smart port indicator and calculation of a ranking for the Spanish port system. Logistics, 4(2): https://doi.org/10.3390/logistics4020009

50  César Ducruet and Theo Notteboom 184 Rodrigue J.P., Ashar A. (2016) Transshipment hubs in the New Panamax Era: The role of the Caribbean. Journal of Transport Geography, 51: 270–279. 185 Rodrigue J.P., Comtois C., Slack B. (1997) Transportation and spatial cycles: Evidence from maritime systems. Journal of Transport Geography, 5(2): 87–98. 186 Rodrigue J.P., Notteboom T.E. (2010) Foreland-based regionalization: Integrating intermediate hubs with port hinterlands. Research in Transportation Economics, 27(1): 19–29. 187 Rousset L., Ducruet C. (2020) Disruptions in spatial networks: A comparative study of major shocks affecting ports and shipping patterns. Networks and Spatial Economics, 20(2): 423–447. 188 Sahu P.K., Padhi A., Patil G.R. (2019) Spatial temporal analysis of freight f low through Indian major seaport system. The Asian Journal of Shipping and Logistics, 35(1): 77–85. 189 Sánchez R.J., Wilmsmeier G. (2006) The river plate basin–A comparison of port devolution processes on the East Coast of South America. Research in Transportation Economics, 17: 185–205. 190 Santos T.A., Soares C.G. (2017) Development dynamics of the Portuguese range as a multi-port gateway system. Journal of Transport Geography, 60: 178–188. 191 Santos T.A., Soares C.G. (2019) Container terminal potential hinterland delimitation in a multi-port system subject to a regionalization process. Journal of Transport Geography, 75: 132–146. 192 Sdoukopoulos E., Boile M. (2020) Port-hinterland concept evolution: A critical review. Journal of Transport Geography, 86: 102775. 193 Shneerson D. (1981) Investment in port systems: A case study of the Nigerian ports. Journal of Transport Economics and Policy, 15(3): 201–216. 194 Shneerson D. (1983) Short term planning for a port system. Maritime Policy and Management, 10(4): 217–250. 195 Slack B. (1985) Containerization, inter-port competition, and port selection. Maritime Policy and Management, 12(4): 293–303. 196 Slack B. (1994) Domestic containerization and the load centre concept. Maritime Policy and Management, 21(3): 229–236. 197 Slack B. (1999) Across the pond: Container shipping on the North Atlantic in the era of globalisation. GeoJournal, 48(1): 9–14. 198 Slack B., Starr J. (1997) Geostrategy, resources and geopolitics: Implications for the US Navy and naval port system. Marine Policy, 21(4): 309–323. 199 Slack B., Wang J.J. (2002) The challenge of peripheral ports: An Asian perspective. GeoJournal, 56(2): 159–166. 200 Smith H.D., Pinder D. (1997) Geostrategy and naval port systems: Frameworks for analysis. Marine Policy, 21(4): 291–308. 201 Smith H.D., Pinder D. (1998) Geostrategy, naval power and naval port systems. Oceanographic Literature Review, 1(45): 141. 202 Song S.C., Park S.H., Yeo G.T. (2019) Network structure analysis of a sub-huboriented port. The Asian Journal of Shipping and Logistics, 35(2): 118–125. 203 Spengler T., Tovar B. (2021) Potential of cold-ironing for the reduction of externalities from in-port shipping emissions: The state-owned Spanish port system case. Journal of Environmental Management, 279: 111807. 204 Stamatovic K., de Langen P.W., Groznik A. (2018) Port cooperation in the North Adriatic ports. Research in Transportation Business and Management, 26: 109–121.

A systematic and critical review of port system research  51 205 Starr J.T. (1994) The mid-Atlantic load centre: Baltimore or Hampton roads? Maritime Policy and Management, 21(3): 219–227. 206 Sun Y., He W. (2008) A study on port cooperation in the Tianjin and Hebei areas using factor analysis. Journal of Comparative International Management, 11(2): 23–31. 207 Svindland M., Monios J., Hjelle H.M. (2019) Port rationalization and the evolution of regional port systems: The case of Norway. Maritime Policy and Management, 46(5): 613–629. 208 Taaffe E.J., Morrill R.L., Gould P.R. (1963) Transport expansion in underdeveloped countries: A comparative analysis. Geographical Review, 53(4): 503–529. 209 Talley W.K. (1988) The role of US ocean ports in promoting an efficient ocean transport system. Maritime Policy and Management, 15(2): 147–155. 210 Tavasszy L., Minderhoud M., Perrin J.F., Notteboom T.E. (2011) A strategic network choice model for global container f lows: Specification, estimation and application. Journal of Transport Geography, 19(6): 1163–1172. 211 Todd D. (1993) The interplay of trade, regional and technical factors in the evolution of a port system: The case of Taiwan. Geografiska Annaler: Series B, Human Geography, 75(1): 3–18. 212 Todd D., Zhang L. (1993) Political and technical factors impinging on port operations: The case of Manchuria. GeoJournal, 30(4): 441–454. 213 Trepte K., Rice Jr J.B. (2014) An initial exploration of port capacity bottlenecks in the USA port system and the implications on resilience. International Journal of Shipping and Transport Logistics, 6(3): 339–355. 214 Tuljak-Suban D., Twrdy E. (2015) Fuzzy empty containers excess estimation as an economic indicator—the case of the north Adriatic port system. Maritime Policy and Management, 42(8): 759–775. 215 Uderman S., Rocha C.H., Cavalcante L.R. (2012) Modernization of the port system in Brazil: A methodological proposal. Journal of Transport Literature, 6: 221–240. 216 Van Klink H.A. (1998) The port network as a new stage in port development: The case of Rotterdam. Environment and Planning A, 30(1): 143–160. 217 Van Thai V., Grewal D. (2005) An analysis of the efficiency and competitiveness of Vietnamese port system. Asia Pacific Journal of Marketing and Logistics, 17(1): 3–31. 218 Veenstra A., Notteboom T.E. (2011) The development of the Yangtze River container port system. Journal of Transport Geography, 19(4): 772–781. 219 Vermeiren T., Macharis C. (2016) Intermodal land transportation systems and port choice, an analysis of stated choices among shippers in the Rhine–Scheldt delta. Maritime Policy and Management, 43(8): 992–1004. 220 Vonck I., Notteboom T.E. (2016) Panarchy within a port setting. Journal of Transport Geography, 51: 308–315. 221 Walker A.R. (1989) Recessional and Gulf war impacts on port development and shipping in the Gulf States in the 1980’s. Geojournal, 18(3): 273–284. 222 Wang B., Chin K.S., Su Q. (2022) Prevention and adaptation to diversified risks in the seaport–dry port system under asymmetric risk behaviors: Invest earlier or wait? Transport Policy, 125: 11–36. 223 Wang B., Chin K.S., Su Q. (2022) Risk management and market structures in seaport–dry port systems. Maritime Economics & Logistics, 24(1): 114–137.

52  César Ducruet and Theo Notteboom 224 Wang C., Ducruet C. (2012) New port development and global city making: Emergence of the Shanghai–Yangshan multilayered gateway hub. Journal of Transport Geography, 25: 58–69. 225 Wang C., Ducruet C. (2013) Regional resilience and spatial cycles: Long-term evolution of the Chinese port system (221BC–2010AD). Tijdschrift voor Economische en Sociale Geografie, 104(5): 521–538. 226 Wang C., Wang J. (2011) Spatial pattern of the global shipping network and its hub-and-spoke system. Research in Transportation Economics, 32(1): 54–63. 227 Wang C., Wang J., Ducruet C. (2012) Peripheral challenge in container port system: A case study of Pearl River Delta. Chinese Geographical Science, 22(1): 97–108. 228 Wang J.J. (1998) A container load center with a developing hinterland: A case study of Hong Kong. Journal of Transport Geography, 6(3): 187–201. 229 Wang J.J., Ng A.K.Y. (2011) The geographical connectedness of Chinese seaports with foreland markets: A new trend? Tijdschrift voor Economische en Sociale Geografie, 102(2): 188–204. 230 Wang J.J., Slack B. (2000) The evolution of a regional container port system: The Pearl River Delta. Journal of Transport Geography, 8(4): 263–275. 231 Wang L., Lau Y.Y., Su H. (2022) Dynamics of the Asian shipping network in adjacent ports: Comparative case studies of Shanghai-Ningbo and Hong Kong-Shenzhen. Ocean and Coastal Management, 221: 106127. 232 Wang L., Notteboom T.E., Lau Y.Y. (2017) Functional differentiation and sustainability: A new stage of development in the Chinese container port system. Sustainability, 9(3): 328. 233 Wang L., Zhu Y., Ducruet C., Bunel M., Lau Y.Y. (2018) From hierarchy to networking: The evolution of the “twenty-first-century Maritime Silk Road” container shipping system. Transport Reviews, 38(4): 416–435. 234 Wang R., Yu G. (1997) The open port system in northeast China. Chinese Geographical Science, 7(3): 270–277. 235 Wang W., Wang C., Jin F. (2017) The functional structure convergence of China’s coastal ports. Sustainability, 9(12): 2185. 236 Wang Y., Cullinane K.P.B. (2014) Traffic consolidation in East Asian container ports: A network f low analysis. Transportation Research Part A, 61: 152–163. 237 Wang Y., Cullinane K.P.B., Hu Y. (2014) The role of feeder shipping in Chinese container port development. Transportation Journal, 53(2): 253–267. 238 Wang Y., Yeo G.T. (2019) Transshipment hub port selection for shipping carriers in a dual hub-port system. Maritime Policy and Management, 46(6): 701–714. 239 Warf B. (1989) US imports of petroleum and related products in the 1980s. Maritime Policy and Management, 16(4): 265–275. 240 Warf B., Kleyn L. (1989) Competitive status of US ports in the mid-1980s. Maritime Policy and Management, 16(2): 157–172. 241 Westerholm J. (1986) The development of a national port system-Denmark 1950–1980. Fennia-International Journal of Geography, 164(2): 211–290. 242 Wiegmans B.W., Konings R., Priemus H. (2009) Critical mass for the development of a new container port in Vlissingen. Maritime Economics & Logistics, 11(4): 399–417. 243 Wiegmans B.W., Witte P., Spit T. (2015) Characteristics of European inland ports: A statistical analysis of inland waterway port development in Dutch municipalities. Transportation Research Part A, 78: 566–577.

A systematic and critical review of port system research  53 244 Wilmsmeier G., Monios J. (2013) Counterbalancing peripherality and concentration: An analysis of the UK container port system. Maritime Policy and ­Management, 40(2): 116–132. 245 Wilmsmeier G., Monios J. (2016) Institutional structure and agency in the governance of spatial diversification of port system evolution in Latin America. Journal of Transport Geography, 51: 294–307. 246 Wilmsmeier G., Monios J., Farfán A.F.B. (2021) Port system evolution in Ecuador–Migration, location splitting or specialisation? Journal of Transport Geography, 93: 103042. 247 Wilmsmeier G., Monios J., Pérez-Salas G. (2014) Port system evolution–the case of Latin America and the Caribbean. Journal of Transport Geography, 39: 208–221. 248 Wilmsmeier G., Monios J., Rodrigue J.P. (2015) Drivers for outside-in port hinterland integration in Latin America: The case of Veracruz, Mexico. Research in Transportation Business and Management, 14: 34–43. 249 Wilmsmeier G., Notteboom T.E. (2011) Determinants of liner shippping network configuration: A two-region comparison. GeoJournal, 76(3): 213–228. 250 Wilmsmeier G., Sanchez R.J. (2017) Evolution of national port governance and interport competition in Chile. Research in Transportation Business and Management, 22: 171–183. 251 Wiradanti B., Pettit S., Potter A. (2018) Ports, peripherality and concentration– deconcentration factors: A review. Maritime Business Review, 3(4): 375–393. 252 Wiradanti B., Pettit S., Potter A. (2020) Willingness to invest in peripheral ports: Perceptions of Indonesian port and maritime industry stakeholders. Maritime Economics & Logistics, 22(4): 699–714. 253 Witte P., Wiegmans B., Ng A.K.Y. (2019) A critical review on the evolution and development of inland port research. Journal of Transport Geography, 74: 53–61. 254 Witters R.L., Ivy R.L. (2002) Port competition for cargo tonnage in the US South. Southeastern Geographer, 42(1): 65–80. 255 Wu X., Luo M., Zhang A. (2017) Do larger ships visit fewer regions/ports? An empirical analysis on global liners serving China. Journal of Transport Economics and Policy, 51(3): 173–192. 256 Xu M., Li Z., Shi Y. (2015) Evolution of regional inequality in the global shipping network. Journal of Transport Geography, 44: 1–12. 257 Xu Q., Huang T., Chen J. (2021) Port rank-size rule evolution: Case study of Chinese coastal ports. Ocean and Coastal Management, 211: 105803. 258 Yang D., Wang K.Y., Xu H. (2017) Path to a multilayered transshipment port system: How the Yangtze River bulk port system has evolved. Journal of Transport Geography, 64: 54–64. 259 Yang Z., Xiu Q., Chen D. (2019) Historical changes in the port and shipping industry in Hong Kong and the underlying policies. Transport Policy, 82: 138–147. 260 Ye S., Qi X., Xu Y. (2020) Analyzing the relative efficiency of China’s Yangtze River port system. Maritime Economics & Logistics, 22(4): 640–660. 261 Yu M., Lee C.Y., Wang J.J. (2017) The regional port competition with different terminal competition intensity. Flexible Services and Manufacturing Journal, 29(3): 659–688. 262 Yu M., Shan J., Ma L. (2016) Regional container port competition in a dual gateway-port system. Journal of Systems Science and Systems Engineering, 25(4): 491–514.

54  César Ducruet and Theo Notteboom 263 Žgaljic D., Tijan E., Jugovic A. (2019) Implementation of sustainable Motorways of the Sea services multi-criteria analysis of a Croatian port system. Sustainability, 11(23): 6827. 264 Zhang Q., Yan K., Yang D. (2021) Port system evolution in Chinese coastal regions: A provincial perspective. Journal of Transport Geography, 92: 103031. 265 Zhang X., Lu J., Peng Y. (2022) Spatio-temporal evolution of the container port system along the 21st-century Maritime Silk Road. Maritime Policy and Management, https://doi.org/10.1080/03088839.2021.2017038 266 Zhao C., Wang Y., Gong Y. (2021) The evolution of the port network along the Maritime Silk Road: From a sustainable development perspective. Marine Policy, 126: 104426. 267 Zheng J., Shao B. (2022) Throughput properties and clustering analysis of coastal ports in mainland China: An analysis method by time series complex network. DYNA-Ingeniería e Industria, 97(4): 398–405. 268 Zografos K.G., Martinez W. (1990) Improving the performance of a port system through service demand reallocation. Transportation Research 002EPart B, 24(2): 79–97.

2

Evolutionary models of port system development – an application to the Latin American and Caribbean port system Gordon Wilmsmeier and Jason Monios

Introduction Some of the first works of transport geography were based on the evolution of port systems in the 1960s. This chapter traces the development of these models that conceptualize the spatial transformations from simple multifunctional port installations to specialized terminals and relocated sites to more complex systems in which various terminals or ports compete with each other over fully or partly overlapping hinterlands. As the advent of containerization reduced the uniqueness of each port through standardizing freight transport and allowing economies of scale in shipping and hinterland transport, intermodal transport corridors allowed inland locations to be served from a choice of ports. As a result, containerized cargo traffic began to consolidate at larger ports which was then transshipped to smaller ports via hub-and-spoke networks. Evolutionary models have revealed concentration, deconcentration, and hierarchization processes in the global port system over the last decades. These temporal-spatial changes in the position of ports within port systems at different scales emerge from the interplay of a series of factors, principally demand and geography, but also operational factors such as maritime access limitations, port congestion, and proactive business strategies including new port locations, as well as container liner shipping network strategies. This chapter illustrates these trends with examples from different ranges of the Latin American and Caribbean (LAC) port system, ref lecting concentration and deconcentration, changes in transshipment markets and the impact of changes in port governance, the effects of vertical integration and port competition. The analysis provides the background to the discussion on how the drivers of port system evolution have changed over time due to technological change and increased industry consolidation, and how more recent challenges such as the COVID-19 pandemic may trigger further changes in the sector.

The geography of port systems Port system evolution is a function of both individual port developments and wider system dynamics coming from the sea and land interactions.

DOI: 10.4324/9781003316657-4

56  Gordon Wilmsmeier and Jason Monios

Port systems are compositions of physical geography, infrastructure, and governance frameworks, located within an economic system, where the shipping sector responds to and shapes trade relations. The economic system and the shipping sector in their interplay generate pressure on the port system in the form of ever-evolving specific requirements with respect to infrastructure, superstructure, equipment, efficiency, institutional structure, and agency. This prompts a process of time-lagged reaction within the port system to satisfy this changing demand and it is this reactive progression that actually constitutes the port development process, determined by and ref lected in its physical (infrastructure and superstructure), economic, social/ environmental, and institutional arrangements. Conceptualizations of individual port developments go back to the 1960s and Bird’s “anyport” model (Bird, 1963; Hoyle, 1968). Bird’s (1963) original model of the spatial evolution of port infrastructure included six phases: the original port, quay extension, quay elaboration, dock elaboration, simple lined quays, and specialized quays. Rodrigue et al. (2009) argued that these can be condensed into three phases: setting, expansion, and specialization. Expansion involves an enlargement of port activities as traffic grows, often constructing purpose-built berths with greater water depth. This building of new facilities then leads naturally into a division of activities via the phase of specialization, for example, specialized timber, grain, oil, or container terminals. As more products were containerized throughout the 1980s and 1990s, the need for traditional break bulk and general cargo handling activities was reduced. With growing port activity, in some cases, ports migrated away from their original city locations, either entirely or just certain specialized terminals, which then secured their hinterland through high capacity inland transport corridors (Taaffe et al., 1963; Rimmer, 1967; Hayuth, 1981; Barke, 1986; Hoyle, 1989). As an individual port expands its original installations and develops specialized terminals, these inf luence the wider system. A specialized container terminal with a deep-water location and good hinterland links attracts more container ships, gaining economies of scale and offering lower port costs to shipping lines, leading to a concentration of traffic at a handful of key ports. Other ports that cannot compete, whether due to geographical or institutional disadvantages, see their traffic decline (Ducruet et al., 2009; Wiradanti et al., 2018). In the last half century, port ranges around the world have seen traffic concentrate in a handful of key nodes, with smaller ports either becoming feeder ports in hub-and-spoke networks or even losing their freight traffic entirely and shifting their focus toward passengers or leisure uses. Over time, however, diseconomies of scale and port congestion can lead to a move toward deconcentration (Slack and Wang 2002; Frémont and Soppé, 2007). Various authors have conceptualized the temporal dimension of port development, such as the UNCTAD generational model (UNCTAD, 1992), the more operationally focused WORKPORT model (Beresford et al., 2004), and the port life cycle as it moves through phases of development,

Evolutionary models of port system development  57

introduction, growth, maturity, and decline (Charlier, 1992; Cullinane and Wilmsmeier, 2011). Due to the fact that the port and port system development cycles advance in a discrete manner, their adjustment to the continuous evolution of freight transport demand will inevitably lead to alternating situations of either infrastructural insufficiency or scarcity of supply on the one hand (i.e., excess demand), or to a surfeit of port infrastructure (i.e., surplus supply). Further, port development is dependent upon and determined by the degree to which a specific port system in question is embedded within local, regional, or national institutional structure. This is critically important not only to the port but also to the economy it serves as the governance model creates the match with the environment and ultimately defines the degree of connectivity enjoyed by the economic system that prevails within a port’s hinterland. This ref lexive and contextual embeddedness of ports led Wilmsmeier and Sánchez (2010) to define the port system as an “autopoietic” system (cf. Maturana and Varela, 1980), meaning that it changes its state with each new input emphasizing the importance of the first mover advantage meaning that a delayed action may no longer be suitable to the new state of the system and consequently time-lagged investments and strategy replication carry significant risks. Temporal empirical analyses of port ranges around the world confirmed trends toward first concentration and then deconcentration (e.g., Notteboom, 2006; Notteboom, 2010; Wilmsmeier and Monios, 2013; Wilmsmeier et al., 2014; Svindland et al., 2019). A meta-analysis performed by Ducruet et al. (2009) on 31 studies from 1963 to 2008 found that concentration is mostly driven by the development of economies of scale, technological change, and the appearance of inland corridors. Deconcentration comes as a result of port congestion and constraints in hinterland connections as well as strategic and institutional factors such as port governance and investment strategies. However, it is important to differentiate between deconcentration that emerges upon failure of a system in a reactive manner, deconcentration that materializes from proactive port development strategies, and deconcentration that emerges from new economic and industrial development. Deconcentration provides a new opportunity for previously peripheral ports to take on “the challenge of the periphery” (Hayuth, 1981; Barke, 1986; Slack and Wang, 2002) and recapture some lost traffic; this may particularly be the case in developing economies, and often driven by active government policies (Li et al., 2012; Wiradanti et al., 2018; Monios et al., 2019). While almost all empirical studies analyze traffic concentration at the level of the port range, Guerrero and Rodrigue (2014) studied container port concentration between 1970 and 2010 at the global level. Their analysis showed an increase in concentration from 0.57 to 0.7 up to 1990, which remained stable until 2010, with 21% of ports handling 75% of global TEU throughput. This result was based on the growth of container ports in Asia and the emergence of transshipment hubs. Ducruet (2017) found that, despite the

58  Gordon Wilmsmeier and Jason Monios

volatility of shipping networks, the global port hierarchy remained rather stable and strongly path dependent. De Oliveira et al. (2021) analyzed global port hierarchies and concluded that trends in concentration make it very difficult for medium ports to maintain competition with the largest ports. As the container shipping market becomes more mature, port ranges become more stable in terms of small, medium, and large ports, notwithstanding competition within each group.

Liner shipping networks and port system evolution The evolution of port systems cannot be understood in isolation from an understanding of liner shipping networks, both of which have continued to exhibit strong concentration processes as well as increasing vertical integration. During the financial crisis of 2008/9, the top ten container carriers held 70.8% of total f leet capacity but a wave of mergers and acquisitions followed, due to drops in profitability after the overcapacity crisis, as well as the bankruptcy of Hanjin, at that time the seventh-largest container carrier. As a result, this figure had risen to 91.5% by 2020. A similar situation can be observed in container ports, where the top ten operators have increased their share of global throughput from 41% in 2001 to 70% in 2018 and many of these leading port operators are vertically integrated with carriers (Notteboom et al., 2021). These trends have made port development dependent on the network strategies of global players. The development of liner shipping networks is primarily driven by the demand for containerized transport, depending on the strategies of shipping companies and the expectations of shippers for specific service characteristics. As such, the location of a port or a region within the global liner shipping network is determined by the density of trade f lows to and from a specific port or region. These factors then become the determinants of the service characteristics (i.e., frequency, loading capacity, number of port calls per roundtrip, and transshipment or relay strategies) as well the level of competition (number of shipping lines offering services). From the carrier’s perspective, the economies of scale, scope and density in shipping, port operations, and inland operations would favor a very limited number of load centers in a region, which however might not correlate directly with the distribution of demand. This work refers to the Wilmsmeier and Notteboom (2011) generic fourphase model of the evolution of liner shipping networks, which is applied as a qualitative holistic model to place the evolving role of the port within the port system context (Figure 2.1). The model tracks the geographical development of port systems, from point-to-point services with a more regional focus, to the widespread devolution and liberalization of the sector to allow entry of larger private operators in both ports and shipping in tandem with economic growth that drives the increase in demand. Continued technology evolution and industry consolidation leads to larger ports and ships, which

Evolutionary models of port system development  59

First phase.

Second phase.

Third phase.

Fourth phase.

• • • •

liner shipping network determined by point-to-point direct services with a strong local or regional orientation. liner service network highly regional in orientation and interconnectivity to the overseas markets is poor. Government involvement in the port sector is high international market players (shipping lines and terminal operators) face limited possibilities.

• actors seek a higher connectivity to overseas markets by consolidating cargo in an intermediate hub. • first tendencies towards a hub-and-spoke network emerge. The evolving liner service network configuration increases the dependency of the port system on indirect services via the hub, while direct regional services start to lose their importance. • growing connectivity of the port system to overseas markets increases the region’s a ttractiveness to shipping lines and international port operators. • local/regional/national government seek port devolution process to face the mounting infrastructural and operational port challenges. • changes in the port governance and policy framework enable international stevedoring groups and shipping lines to access key assets in the local ports and to seek control over terminal operations.

• • • •

Port traffic growth leads to a further outreach of the hub-and-spoke network and the inclusion of new ports in this pa ttern. ITOs strengthen market presence state intervention in ports is strongly reduced. main lines are growing and smaller regional services start to develop again in a secondary network.

• market size of specific ports has grown for shipping lines offer direct services from these ports to overseas regions. • hubs see functional positions undermined. To maintain role in the network, hubs seek liner service connections to smaller ports in the region which still lack connectivity to overseas market. • terminal activity in hubs shifts in geographical terms and a new secondary hub-and-spoke network emerges involving other gateway ports. • state intervention in ports limited and institutional reform discussions emerge

Figure 2.1 Generic four-phase model of the evolution of liner shipping networks and ports Source: Adapted from Wilmsmeier and Notteboom (2011)

underpin a hub-and-spoke model, leading to concentration of traffic at larger ports. Finally, diseconomies of scale and port congestion can lead to a move toward deconcentration with some secondary ports able to recapture some traffic and establish point-to-point services. Wilmsmeier et al. (2014) demonstrated that the LAC port system was in transition from phase three to four of the model, and identified five key drivers that impact on port system development: economic development, technological change, port devolution, port function, and shipping line strategy. Based on these earlier studies, this chapter revisits the evolution of maritime networks and the autopoietic nature of port development and the results of proactive and reactive strategies of secondary ports in their endeavor to (re) position themselves within the container shipping market over the last quarter century.

Evolution of drivers and outcomes in port systems: the case of Latin America and the Caribbean Setting the scene The evolution of container port systems, measured by their activity, is measured in two dimensions: geographical and industry (terminal operator). The geographical dimension in relation to a port or port system´s activity relates to the dichotomies of peripherality and centrality, and concentration and

60  Gordon Wilmsmeier and Jason Monios

deconcentration. The former being related to the insertion and connectivity of a port within the global liner shipping network, the latter relates to the geographical container volume distribution within a port system in pure volume terms but also considering the relationship of the activity to industry groups (e.g., ITO). This chapter describes the changes in the evolution of the LAC port system and discusses the drivers of these changes as ports are situated in the interplay between supply (liner shipping services as well as port capacity) and demand (container traffic). While this discussion cannot be conclusive, the aim is to place arguments for a more systemic view of port system development and to reflect on geographical changes in port system hierarchy, taking into account main, secondary, and tertiary container port activity ports. Such discussion of port system evolution cannot happen without considering the institutional framework conditions that shape and accompany the specific development strategies. A previous study of the LAC system by Wilmsmeier et al. (2014) found evidence on most coastal ranges of initial spatial activity concentration followed by spatial deconcentration, which was driven by the conversion of a point-to-point to hub-and-spoke liner shipping network during the first decade of this century. Such development was facilitated by the port reforms that occurred in LAC in the 1990s and 2000s, instituting the devolution of port operations to the private sector and has allowed the inf lux of shipping lines and terminal operators in the regional market. In many Latin American countries, the result of the reforms were quick advances to create operational improvements but a loss of long-term agency that became more evident since the 2010s (Wilmsmeier and Monios, 2016) when the operational improvements reached global standards and the port systems were being pressured by global terminal operators and carriers for new investments and expansions beyond the traditional port sites. The lack of institutional capacity and agency has led many LAC countries to become once again reactive in port and logistics policy development and strategies. The following analysis will chart and update the progress of the LAC port system and examine whether the drivers previously identified have changed. The methodology includes both quantitative and qualitative analysis. The former is based on a database covering 180 container ports across the LAC port system from the years 1997 to 2019 (in some cases, the work will refer to more recent developments). This data allows an examination of processes of spatial concentration and deconcentration over time, transshipment location selection, emergence of new ports, and vertical and horizontal integration impacts, from both a macro system perspective (subregional LAC port systems) and a national and local port system perspective. This quantitative descriptive analysis is paired with qualitative data relating to port development and infrastructure investment strategies at primary and secondary ports. The selected samples are chosen to exemplify the evolutionary processes identified in the literature and to demonstrate the ports’ ability to (re)act, in the context of the interplay between the economy, the environment, institutional structures, and industry strategies.

Evolutionary models of port system development  61

Evolution of container vessel size: West Coast of South America (WCSA) LAC has not been an exception to increased container vessel sizes over the last decades (Figure 2.2). The WCSA can stand as an example of how a market that was dominated by vessels of a maximum of 3,000 TEU in 2008 has converted to a region where in 2022 over 80% of all deployed container vessels are between 8,000 and 17,000 TEU. The inf lux of Post-Panamax and Neo-Panamax vessels started in 2010 on WCSA – Asia services (around 10% of deployed vessels) and accelerated between 2014 and 2017, reaching around 80% of deployed vessel shares (68% 8,000–11,999 TEU, 13% 12,000–16,999 TEU). The dominance of this vessel type stabilized with the opening of the expanded Panama Canal. It should also be noted that vessels smaller than 3,000 TEU are basically absent in the international liner services with calls on the WCSA. However, the effects of the hub-and-spoke network on the port hierarchy can best be exemplified by comparing the evolution of vessel sizes of a primary port, Callao, and a secondary port, Paita, both are located in Peru (Figure 2.3). While Callao shows dynamics in vessel size development in line with the tendencies on the whole WCSA, vessel sizes and composition of the f leet calling in Paita remain with little changes and maximum sizes are not larger than 7,999 TEU.

100% 90% 80% 70% 60% 50% 40% 30% 20%

0%

Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2 Qtr3 Qtr4 Qtr1 Qtr2

10%

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

Feeder 100-1,000

Feeder 1-1,999 TEU

Feeder 2-2,999 TEU

Intermediate 3-7,999 TEU

Neo-Panamax 12-16,999 TEU

Neo-Panamax 8-11,999 TEU

2020

2021

2022

Pure Car Carriers 10-19,999 dwt

Figure 2.2 Comparison of container vessel size evolution on the WCSA, 2012–2021 Source: Own source based on processed data from Clarksons Research (2022) Note: *Always referring to the last observation in April of that year

62  Gordon Wilmsmeier and Jason Monios 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

2012

2013

2014

2015

2016

2017

2018

2019

Feeder 100-1,000

Feeder 1-1,999 TEU

Feeder 2-2,999 TEU

MPP 500-999 TEU

Neo-Panamax 8-11,999 TEU

Neo-Panamax 12-16,999 TEU

2020

2021

2022

Intermediate 3-7,999 TEU

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

2012

2013

2014

Feeder 100-1,000

2015

2016

Feeder 1-1,999 TEU

2017

2018

Feeder 2-2,999 TEU

2019

2020

2021

2022

Intermediate 3-7,999 TEU

Figure 2.3 Comparison of vessel size evolution in primary and secondary ports on the WCSA, 2012–2021 (left: Callao; right: Paita) Source: Own source based on processed data from Clarksons Research (2022) Notes *Always referring to the last observation in April of that year

Evolutionary models of port system development  63

Thus, this example illustrates the literature discussion on issues of infrastructure demand in a hierarchical system and exemplifies the different pressures of infrastructure development depending on the role of a port within a national port system. Ports that intend to be or become principal ports in a system have faced significant exposure to cascading dynamics over the last decade and had to respond in accommodating larger vessels within a short time span. Port saturation leads to deconcentration: East Coast of South America (ECSA) Traditionally, container traffic on the ECSA range has been concentrated at three major ports: Santos (Brazil), Buenos Aires (Argentina), and Montevideo (Uruguay) (Figure 2.4). While the three countries on the ECSA are significantly different in size and coastline, it can be observed that the port activity in Buenos Aires and Montevideo has seen little dynamics over the last decade, and in either country, no major secondary ports have emerged. Overall the port system of the River Plate Basin has been stagnant over the last decade. The development of Uruguay´s port system is restrained by the sheer size and population of the country and decades-long discussions about existing 600

50,00,000

45,00,000

500 40,00,000

35,00,000

30,00,000

300

25,00,000

TEU

Index (1997= 100)

400

20,00,000

200 15,00,000

10,00,000

100 5,00,000

0

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Santos

Buenos Aires

Montevideo

Rio Grande

Santos

Buenos Aires

Montevideo

Rio Grande

0

Figure 2.4 Growth dynamics comparison of main ports along the ECSA port range, 1997–2021 Source: Authors, based on UN-ECLAC and individual port information Note: Index 1997=100

64  Gordon Wilmsmeier and Jason Monios

physical restrictions in the port of Montevideo and the possible dominant market position of the Terminal Cuenca de la Plata in the country have not led to the development of a new container port location. With the recovery after the pandemic, discussions of the Port of Buenos Aires to be downgraded to a feeder port have re-emerged. This possible “threat” prevails since 2004, when Hamburg Süd highlighted the still-prevailing draught restrictions in the navigation channel and access channels to the port, the high cost of transiting the River Plate and restrictive customs practices as limitations to the port´s competitive position, a situation that has not improved significantly since then. The case of Brazil shows significantly different dynamics, which can be observed by the continued growth of Santos as the main container port over the last two decades (Figure 2.5). The dynamics in the Brazilian container port system include the spatial diversification of container activity along the coastline, which can on the one hand be observed by the activity growth in traditional port locations, e.g., Rio de Janeiro, Rio Grande, or Paranagua. On the other hand, the port system has experienced the emergence

6000000

6000000

5000000

5000000

4000000

4000000

3000000

3000000

2000000

2000000

1000000

1000000

TEU

7000000

TEU

7000000

0

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Paranaguá

Rio Grande

Itapoa

Navegantes

Chibatao (Manaus)

Itajai

Suape

Pecem

Rio de Janeiro

Salvador

other ports

Santos

Buenos Aires

Montevideo

0

Figure 2.5 Container throughput (TEU) along the ECSA port range, 1997–2019 Source: Authors, based on UN-ECLAC and individual port information Note: Santos, Buenos Aires, and Montevideo are shown separately in order to see the smaller ports more clearly. Secondary ports with less than 10% market national market share have been summarized in “other ports” respectively for Argentina and Brazil for better illustration. The number in brackets represents the number of ports summarized under “other”. Some ports with very little container movements have been removed.

Evolutionary models of port system development  65

and establishing of new secondary locations, such as Navegantes, Chibatao (Manaus), Pecem, and Itapoa. Thus, Brazil is a clear example of how port systems in emerging economies spatially diversify to respond to the evolution of economic activity and consumer demands. The figure shows that the container activity within the ECSA port system displayed minor dynamics between 2012 and 2016, but since 2017 displayed significant growth dynamics, particularly in the previously mentioned secondary Brazilian ports. The case of Brazil deserves a closer look. Despite the expansion of container activity and advances in terminal operations, public port infrastructure development has been lagging behind in delivering the needed capacity due to challenges in public sector management. Public contracting and procurement regulations have been an obstacle to the efficient management of port infrastructure assets, therefore the country has benefited from private sector involvement in the past that has contributed to significant improvement of the quality of services provided and operational efficiency, as well as providing and maintaining port capacity. In order to improve port competitiveness, five port authorities, responsible for eight ports (Vitoria, Barra do Riacho, Santos, Salvador, Aratu, Ilheus, Itajaí, and São Sebastião), are in a privatization or divestiture process. These are: Companhia Docas do Espírito Santo (CODESA), the Santos Port Authority (SPA) and Companhia Docas da Bahia (CODEBA), Port Itajaí, and Port São Sebastião. This process of further involvement of the private sector is quite interesting as it expands and renews port devolution programs from the 1990s. It can be expected that the results will further alter the port system landscape in Brazil. The expected impacts of privatization and divesture read like a re-imagination of the 1990s port reforms in the region and include the following: The last point ref lects on a general challenge in Latin American countries, and relates to the lack of professionalization of public sector port management since the earlier reforms and thus a growing gap of agency over the last two decades. In this scenario, the Port of Santos container terminal has started the second stage of its expansion. The expansion strategy is the result of the concession’s early renewal process, in 2016. The first completed phase included a

66  Gordon Wilmsmeier and Jason Monios

berth expansion, increased water depth to 16 meters, and an annual capacity increase from 2 to 2.4 million TEU. The second stage includes a further annual capacity expansion to 2.6 million TEU, through the acquisition of automated cranes, a new yard construction and adjustment expansion of the railway tracks for container services. An already planned third stage will bring the annual capacity to 3 million TEUs, by 2031. The current privatization process is an opportunity to further expand operations. The ECSA range thus reveals at least two streams of development. The first being the re-emergence of the threat of new peripheralities in the case of the River Plate Basin, and particularly Buenos Aires, which is driven by a continued lack of improvement of port infrastructure, its maritime access, and lack of demand dynamics. The second being a new “wave” of privatizations in the case of Brazil as the public sector has not been able to respond to market dynamics in an increasingly geographically dispersed container port system. Thus, the geographical deconcentration process is forcing the public sector to give way to further devolution processes and restructuring the role of the public sector in the Brazilian port system. The impact of port governance and port competition: WCSA (Chile) Looking at the WCSA, specifically at the country of Chile, Figure 2.6 shows the evolution of the two major container ports – San Antonio and Valparaiso – as well as the increase of smaller ports. A slight deconcentration can be observed in the small reduction of the national market share of these two major ports, from just over 60% in 1997 to a low of just over 50% in 2009 to about 56% in 2019. What is interesting is the growing gap between these two ports, which were quite close until 2011 but have since diverged. Valparaiso has not lost traffic, staying steady since 2011, but it has lost market share, as all the national traffic growth here has gone to its competitor San Antonio and other secondary ports. At the secondary port level, fruit exports in the south of the country are handled by competitors San Vincente and Coronel. Coronel is beating San Vincente in terms of share, by entering the market in 2009 and since then taking all the new growth, while San Vincente stayed steady throughout this decade. The interesting dimension is that San Antonio and San Vincente are the same operator, thus added together, this operator alone was handling 46% of Chilean container traffic in 2019. This example reveals how a port system expands and diversifies in its geographical location as the national economy develops and different national regions increase their participation in international trade. However, given the devolved structure of port governance, this is actually being contrasted by a concentration of container activity in terms of the terminal operator. Further, the historic decade-long competition of the two main ports over a similar hinterland has been decided in favor of San Antonio. The stagnant port activity in Valparaiso can be interpreted as major changes in the pathway of

Evolutionary models of port system development  67 25,00,000

20,00,000

TEU

15,00,000

10,00,000

5,00,000

-

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019

Arica

Coronel

Iquique

Lirquen

Puerto Angamos

San Vicentte / Talcahuano

Other ports

San Antonio

Valparaiso

Figure 2.6 Container throughput (TEU) at Chilean ports, 1997–2019 Source: Authors, based on UN-ECLAC and individual port information Note: Austral-Punta Arenas includes Puerto Natales, Arturo Prat, José de los Santos Mardones

the future national port system, where the port of Valparaiso, despite its long history, can be expected to move toward a more secondary role, given not only its physical geographical limitations, but also due to the national government’s decision to focus future new port developments on San Antonio, due to the better land availability and hinterland connections (rail and road). Chile in this respect is interesting as these decisions on the future direction of the port system development have been taken at national level in a devolved port governance model. An emerging question here is, if the continued focus on the central range ports (San Antonio and Valparaiso), is not omitting the current and future shifts in production toward the southern central regions of the country. Evolution in port functions from transshipment to hybrid ports in LAC Before entering the analysis of the geography of transshipment traffic within LAC, attention should be drawn to the new equilibrium that has been reached via-à-vis the Pacific and Caribbean port systems. Figure 2.7 shows container throughput at ports in Mexico, Central America, and North Coast South

68  Gordon Wilmsmeier and Jason Monios

America (NCSA), separated by their location on the Pacific or Caribbean coast. The data shows that the “Pacific shift”, related to the increased trade with Asia and China specifically since the turn of the century reached its maximum with the financial crisis in 2008 and since then the traffic shares have stabilized. In 1997, the ratio was 81/19 in favor of the Caribbean. By 2019 it was 56/44. This “Pacific Shift” in trade activity is a main factor for the growth in container traffic, challenging traditional transshipment locations in the Caribbean and leading to new port developments on either side of the Panama Canal. The development of hub-and-spoke liner service strategies paired with vertical integration and import-export trade developments have started to challenge traditional transshipment locations in the Caribbean and have also led to new port developments on either side of the Panama Canal. Within the region traditional and new “pure” transshipment and hybrid ports have been competing and altered the port system and transshipment geography over the last two decades. Figure 2.8 shows total container throughput in key transshipment ports in LAC. The transshipment ports on the Pacific Coast and within the Caribbean basin serve two different but related markets. Containers are transshipped either side of the Panama Canal due to the size restrictions imposed 1,40,00,000

100%

90%

1,20,00,000 80%

1,00,00,000

70%

TEU

50%

60,00,000

Share

60%

80,00,000

40%

30%

40,00,000

20%

20,00,000 10%

-

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019

Caribbean (primary axis)

Pacific (primary axis)

Caribbean

0%

Pacific

Figure 2.7 Container throughput and traffic shares along Pacific and Caribbean ranges (Mexico, Central America, and NCSA), 1997–2019 Source: Authors, based on UN-ECLAC and individual port information

Evolutionary models of port system development  69

by the Canal lock system. While the expansion in 2016 moved the maximum capacity of vessels crossing the Canal from the traditional Panamax (ca. 5,000 TEU) to the Neo-Panamax (ca. 14,000 TEU), the Canal remains a bottleneck for larger vessel types such as Post-Panamax and Ultra-Large Container Vessels (ULCV). On the Caribbean side, containers are transshipped for the Caribbean, NCSA, and the US markets (due to the Jones Act which makes it difficult for US ports to feeder to other US ports), as well as to pass through the canal. Lazaro Cardenas and Manzanillo in Mexico have been attracting increasing shares of transshipment on the Pacific Coast. The transshipment here is mainly from mainline services to regional feeder services on the WCSA and West Coast of Central America (WCCA). The pure transshipment ports are located on the Pacific side (Balboa 90% transshipment) and the Caribbean side (Colón 90% transshipment) of the Panama Canal, and the Caribbean Sea (Kingston, Jamaica 80% and Freeport, Bahamas 97% transshipment). The main hybrid ports show the following transshipment percentage: Manzanillo and Lazaro Cardenas, Mexico (36% and 30%), Cartagena, Colombia (72%) and Caucedo, Dominican Republic (46%) (all percentages 2019 from CEPAL). Beyond these main ports, smaller ports have been competing for transshipment market shares with less and 50,00,000 45,00,000 40,00,000 35,00,000

TEU

30,00,000 25,00,000 20,00,000 15,00,000 10,00,000 5,00,000 0

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Colón (MIT, Evergreen, Panamá Port) PA C

Balboa PA C

Kingston JM C

Lazaro Cárdenas MX P

Cartagena Bay CO C

Manzanillo MX P

San Juan PR C

Freeport BS C

Caucedo DO C

Figure 2.8 Container throughput at transshipment and hybrid ports, 1997–2019 Source: Authors, based on UN-ECLAC and individual port information Note: Figure shows total throughput at each port, not necessarily transshipment. Solid lines represent majority transshipment ports, broken lines represent hybrid ports

70  Gordon Wilmsmeier and Jason Monios

greater success. Buenaventura on the Pacific Coast of Colombia, despite its relative proximity to the Panama Canal, has lost its transshipment traffic in 2019 after having emerged as a transshipment location in the previous years, while the new port of Posorja, Ecuador, in 2021 had a transshipment incidence of 35%. Looking first at the ports with proximity to the Panama Canal, the figure shows the dominance of Colón, Panama and Cartagena, Colombia on the Caribbean side over Balboa on the Pacific coast. The f latness earlier and recent decline of Balboa hides the fact that it is two terminals. Traffic in the traditional Panama Ports Company Balboa terminal has been declining significantly while PSA Panama International Terminal (PIT) has displayed significant growth rates since 2017, when its second expansion phase was finished. This reveals the increased inter-port competition on the Pacific side of the Canal. On the Caribbean side, three terminals compete: Manzanillo International Terminal, Panama Ports Co. Cristóbal, and Colón Container Terminal. It is interesting that the terminals losing share in both ports (Cristóbal terminal in Colón port and the original Balboa terminal in Balboa port) are operated by Panama Ports company. Actually, both terminals run by Panama Ports Company are faring poorly: on the Pacific side, Balboa is losing traffic to PSA, and on the Caribbean side, Cristobal, while still ahead of Colon Container Terminal, has been displaying no growth dynamics. One potential explanation for the loss of transshipment traffic by Balboa on the Pacific side could be explained by large growth in transshipment traffic at both Manzanillo and Lazaro Cardenas in Mexico and shipping lines shifting services between the Panamanian terminals. In the Caribbean, Kingston, Freeport, and Caucedo have not shown significant growth, due to the fact that the Caribbean market in general has not shown significant dynamics and the network strategies of shipping lines and ITOs have been on the development of hybrid port locations. Thus, the limited hinterlands of the Caribbean island ports have somewhat stalled their development. Caucedo, Dominican Republic, is a particular case, as it entered the market in 2003 as a greenfield development, operated by DP World with the aim to become a new transshipment port in the region. On the one hand, the port in the first decade of operation evolved more as a hybrid port with national cargo reaching up to 50%, but since 2011, container traffic has been oscillating around the one million TEU mark. On the other hand, DP World inaugurated the expansion of the port from 1.2 to 2.5 million TEU in 2020, which is expected to allow for new growth dynamics. As in the case of Cartagena, Caucedo and the Dominican Republic seek to position themselves a logistics and manufacturing center for the Americas. Thus, a tendency toward hybrid ports (transshipment representing Lazaro Cardenas 50%, Manzanillo 36%, Cartagena 33%) has been shaping the geography of transshipment traffic in the region. Lazaro Cardenas only started handling transshipment traffic in 2012, but as gateway cargo declined over the last decade, its share of transshipment reached 50% in 2021 (861,000 of 1.6m TEU),

Evolutionary models of port system development  71

keeping its total throughout relatively stable. Manzanillo is the real success story in Mexico, growing both gateway and transshipment volumes. Cartagena, Colombia with the terminals CONTECAR and SPR Cartagena has continued to cement its role in the transshipment market and in 2020 became the largest transshipment port in the Caribbean handling more than three million TEU. This is also ref lected in the port having the highest connectivity of liner services in the region. Cartagena has been very successful on building transshipment traffic as a business that started in 2005, when Hamburg Süd decided to make the port its strategic transshipment hub for Latin America and the Caribbean connecting to seven of the carrier’s services between North and South America, the Caribbean, the Mediterranean, and North Europe. Beyond the productivity of the port, its strategy as a logistics partner for companies such as Decathlon or Panalpina and its expansion of reefer infrastructure have shaped the competitiveness in the market. Beyond the successes, port development “fantasies” based on transshipment activities in LAC and particularly in the Caribbean basin and on the Pacific coast of Central and South America have been present in probably each country in the region and many remained white elephants. Some examples are Port La Union, El Salvador; Puerto Manta, Ecuador, and Buenaventura, Colombia. Thus, the evolution of the port system in terms of transshipment market participation has to be interpreted in the dichotomy of success and failure, which depends on multidimensional critical factors. In general, the complementarity between global and local f lows, providing transshipment from main to regional services as well as serving local and regional gateway traffic, seems to offer advantages as the hub-and-spoke liner service network expands. This example from LAC illustrates the literature discussion on the key role played by transshipment hubs in hub-and-spoke networks, how they can compete locally for this market as their service offerings are almost identical, but also how hybrid ports can grow their share of transshipment traffic to compete against pure transshipment hubs, reducing their reliance on gateway traffic. However, analysis of specific liner service network line and vertical integration strategies is required to reveal this finding in more detail, which is beyond the scope of this chapter. Port migration from cities to deep water: the Guayas port system, Ecuador While the previous sections dealt with regional and national port system evolution, this section takes a closer look at a local port system development. The city of Guayaquil is the second-largest in Ecuador and the location of the country’s major port, handling over 2m TEU in 2019, the second-largest container throughput on the WCSA. The port (CONTECON) is located on the river Guayas, approximately 85 km from the Pacific coast, and forms part of the Guayas port system, which is composed of various ports spread out along the estuary. The historic port location was in the city, but in the

72  Gordon Wilmsmeier and Jason Monios

1950s, it was moved downriver of the city, the new port opening in 1963. But this was still eight hours from the coast, and over the decades, the river passage needed to be dredged several times, at a cost of millions of dollars. In the early 2000s, a plan developed to construct an entirely new deep-water port at Posorja at the river mouth. Initially APM Terminals was involved in the plans, but they later withdrew as the project dragged on. Finally, DP World was awarded a 50-year concession on a BOOT (build, own, operate, transfer) model, with the goal that the port would eventually be transferred to the state. Many issues were raised about the negotiation of this deal directly by the Ecuadorian government, as well as the low returns being made to the state from the expected port revenues. Similarly, the lack of a coordinated port policy at the national level has meant that the existing upriver port continues to dredge the river to 12.5 m and continue operations, now being able to accept new Panamax vessels of around 14,000 TEU, although limitations in individual berths means that they cannot be fully loaded to their full draft. Lack of inland road and rail infrastructure to the new port also raises questions as to whether traffic at the new port will connect with the hinterland overland or by feedering to the original Guayaquil port. This example illustrates the literature discussion on issues of port cities facing the challenge of increasing traffic, upriver locations, and particularly increasing ship size. Even going back to the earliest literature from Bird (1963), we find a discussion of ports located in cities, usually on rivers, that need to expand their ports, build specialized quays and terminals, and often eventually migrate at least part of the activity to a new dedicated location. The discussion of this case by Wilmsmeier et al. (2021) further revealed how these decisions are complicated by issues of port governance and port competition, as well as historic legacies of city governance and local economic interests.

Conclusion: reviewing the changing drivers of port system evolution and looking ahead to potential future developments The previous sections have described the changes in different parts of the LAC port system, changes that belie a high level of complexity. The described phenomena are the result of the interaction of a large number of independent or semi-independent factors. Consequently, the described port system changes will in no case have to do with simple relationships, but rather will be symptoms of success of failure in interpreting, understanding drivers, and implementing corresponding strategies in a changing environment. Table 2.1 summarizes the key drivers of port system evolution in Latin America and the Caribbean, divided into time periods 1997–2012 (taken from Wilmsmeier et al., 2014) and 2013–2019 (based on the analysis in this chapter). The analysis in this chapter has shown examples of some of the key trends identified in the port geography literature charting the evolution of

Evolutionary models of port system development  73 Table 2.1 Drivers of container port system evolution in Latin America and the Caribbean, 1997–2019 Dimensions

Drivers 1997–2012

Economic • growth of container development volumes • diversification of geography of trade • change in the structure of cargoes, emergence of containerized reefer cargo Technological • ship size increase change (limitation on Panama Canal routes) • investment in new superstructures • logistics information systems Port devolution • port reform • inf lux and expansion of international private operators • intra-terminal competition

Port function

Shipping line strategy

Drivers 2013–2019 • slower growth in volumes • Pacific shift • Strong growth in containerized reefer trade

• continuing growth in ship size, removal of limitations by Panama Canal expansion • emerging electrification of port equipment • emerging semi-automatization • increased of digitalization in port services • re-emergence of port privatization efforts • discussion on concession extension and renewals from 1990s port reform • limitations in devolved port governance issues preventing some ports dealing with challenges, e.g., in Chile and Argentina • challenges in institutional capacities to respond to port sector development • hybrid ports taking the lead from traditional transshipment hubs

• emergence and strong growth of transshipment ports • emergence of hybrid ports • network evolution from • transshipment traffic an direct services toward important growth strategy for transshipment strategies hybrid (former gateway) ports • emergence of vertical • growth in presence of vertical integration integration • appearance of liner• expansion of presence of linerspecific transshipment specific transshipment hubs hubs

port systems. One of the major trends is how container port systems move toward concentration, but then reach saturation as diseconomies of scale drive a process of deconcentration whereby secondary ports are able to obtain some market share. This was shown clearly on the ECSA and also to some degree in Chile on the WCSA. The role of technology, primarily ship

74  Gordon Wilmsmeier and Jason Monios

size, continues to be a key driver as discussed in the case of cascading vessels to the WCSA. Similarly, the need for ports in traditional city locations to expand and even migrate toward deep water as a result of these changes was shown in the case of Ecuador. Transshipment remains essential to huband-spoke systems, but a previously mature market can be challenged by infrastructure investment which changes the dynamic and need for transshipment as shown in the Panama Canal and the Pacific shift. The analysis of the transshipment market also revealed the growth of hybrid ports which are challenging traditional transshipment specialists and capturing much of the growth in this segment. Changes in port governance and increased port competition was shown in the Chilean case, in which traditional port authorities are stagnating while a new operator is gaining much of the new growth in these locations. While the drivers of port systems in the last decade continue to be primarily the same as the previous decade, some more recent trends may inf luence how these systems continue to evolve in the future. For example, the diseconomies of scale in large ports that has already been occurring during the last decade is now being exacerbated by COVID-19 and port congestion, which may accelerate a drive toward more direct calls and less use of transshipment as well as growth at smaller ports not facing this congestion. Also a result of the COVID-19 disruption and massive price increases by carriers is increasing calls for regulation of alliances – could this also lead to less consolidation of both carriers and port operators which will also drive more direct calls and a less hierarchical port network? The challenges of COVID-19 have provided an opportunity to imagine how increased climate change impacts and the challenges of geopolitics may lead to a reduction of globalization and a shortening of some supply chains. How will these trends affect port development at both the individual port and system levels?

References Barke M. (1986) Transport and Trade. Conceptual Frameworks in Geography. Edinburgh: Oliver & Boyd. Beresford A.K.C., Gardner B.M., Pettit S.J., Naniopoulos A., Wooldridge C.F. (2004) The UNCTAD and WORKPORT models of port development: Evolution or revolution? Maritime Policy and Management, 31(2): 93–107. Bird J. (1963) The Major Seaports of the United Kingdom. London: Hutchinson. Charlier J. (1992) The regeneration of old port areas for new port uses. In: Hoyle B.S., Pinder D.A. (Eds.), European Port Cities in Transition. London: Belhaven Press, pp. 137–154. Cullinane K.P.B., Wilmsmeier G. (2011) The contribution of the dry port concept to the extension of port life cycles. In: Böse J.W. (Eds.), Handbook of Terminal Planning, New York: Springer. de Oliveira G.F., Schaffar A., Cariou P., Monios J. (2021) Convergence and growth traps in container ports. Transport Policy, 110: 170–180.

Evolutionary models of port system development  75 Ducruet C., Roussin S., Jo J.C. (2009) Going West? Spatial polarization of the North Korean port system. Journal of Transport Geography, 17(5): 357–368. Ducruet, C. (2017). Multilayer dynamics of complex spatial networks: The case of global maritime f lows (1977–2008). Journal of Transport Geography, 60: 47–58. Frémont A., Soppé M. (2007) Northern European range: Shipping line concentration and port hierarchy. In: Wang J., Olivier D., Notteboom T.E., Slack B. (Eds.), Ports, Cities and Global Supply Chains. Aldershot: Ashgate. Guerrero D., Rodrigue J.P. (2014) The waves of containerization: Shifts in global maritime transportation. Journal of Transport Geography, 35: 151–164. Hayuth Y. (1981) Containerization and the load center concept. Economic Geography, 57(2): 160–176. Hoyle B.S. (1968) East African seaports: An application of the concept of ‘anyport’. Transactions of the Institute of British Geographers, 44: 163–183. Hoyle B.S. (1989) The port-city interface: Trends, problems and examples. Geoforum, 20(4): 429–435. Li K.X., Luo M., Yang, J. (2012) Container port systems in China and the USA: A comparative study. Maritime Policy and Management, 39(5): 461–478. Maturana H., Varela F. (1980) Autopoiesis and cognition: The realization of the living. In: Cohen R.S., Wartofsky M.W. (Eds.), Boston Studies in the Philosophy of Science, 42, Reidel Publishing Co. Monios J., Wilmsmeier G., Ng A.K.Y. (2019) Port system evolution–the emergence of second-tier hubs. Maritime Policy and Management, 46(1): 61–73. Notteboom T.E. (2006) Traffic inequality in seaport systems revisited. Journal of Transport Geography, 14(2): 95–108. Notteboom T.E. (2010) Concentration and the formation of multi-port gateway regions in the European container port system: An update. Journal of Transport Geography, 18(4): 567–583. Notteboom T.E., Pallis A.A., Rodrigue J.P. (2021) Disruptions and resilience in global container shipping ports: The COVID-19 pandemic versus the 2008–2009 financial crisis. Maritime Economics & Logistics, 23: 179–210. Rimmer P.J. (1967) Changes in the ranking of Australian seaports 1951–2–1961–2. Tijdschrift voor Economische En Sociale Geografie, 58(1): 28–38. Rodrigue J.P., Comtois C., Slack B. (2009) The Geography of Transport Systems, 2nd ed. Abingdon: Routledge. Slack B., Wang J.J. (2002) The challenge of peripheral ports: An Asian perspective. GeoJournal, 56(2): 159–166. Svindland M., Monios J., Hjelle H.M. (2019) Port rationalization and the evolution of regional port systems: The case of Norway. Maritime Policy and Management, 46(5): 613–629. Taaffe E.J., Morrill R.L., Gould P.R. (1963) Transport expansion in underdeveloped countries: A comparative analysis. Geographical Review, 53(4): 503–529. UNCTAD (1992) Development and Improvement of Ports: The Principles of Modern Port Management and Organisation. Geneva: UNCTAD. Wilmsmeier G., Monios J. (2013) Counterbalancing peripherality and concentration: An analysis of the UK container port system. Maritime Policy and Management, 40(2): 116–132. Wilmsmeier G., Monios J. (2016) Institutional structure and agency in the governance of spatial diversification of port system evolution in Latin America. Journal of Transport Geography, 51: 294–307.

76  Gordon Wilmsmeier and Jason Monios Wilmsmeier G., Monios J., Ballén Farfán A.F. (2021) Port system evolution in Ecuador – migration, location splitting or specialisation? Journal of Transport Geography, 93: 103042. Wilmsmeier G., Monios J., Pérez-Salas G. (2014) Port system evolution: The case of Latin America and the Caribbean. Journal of Transport Geography, 39: 208–221. Wilmsmeier G., Notteboom T.E. (2011) Determinants of liner shipping network configuration: A two-region comparison. GeoJournal, 76(3): 213–228. Wilmsmeier G., Sánchez R.J. (2010) Evolution of shipping networks: Current challenges in emerging markets. Zeitschrift fuer Wirtschaftsgeographie, 3/4: 180–193. Wiradanti B., Pettit S.J., Potter A., Abouarghoub W. (2018) Ports, peripherality and concentration–deconcentration factors: A review. Maritime Business Review, 3(4): 375–393.

3

Winding paths through urban systems and urban networks Benjamin Preis

Introduction Gone are the days of the closed city. While the medieval town may have had high walls and strict checkpoints limiting the entry and exit of guests, the modern city is one largely defined by what f lows through the city and between cities, rather than what remains permanently within or outside of it. The modern “city” can have intra-city f lows: connections between the boroughs of London, the comunas of Santiago, the patchwork of 101 cities and town that make up metropolitan Boston. Cities can also have inter-city f lows: connections from different city regions, like the movements of goods or people or information. For the purpose of this chapter inter-city f lows are our main focus. This chapter will provide a rough overview of the different scales through which urban systems can be understood. We first define elements of an urban system and an urban network to introduce the reader to common terms used to refer to urban network principles. We then offer a range of perspectives on the methods, assumptions, and theories underlying the differing studies of urban systems. Within the literatures that study urban systems – which, importantly, ranges from sociology to economics to complexity science – there is a dearth of agreement on what is being studied, and how best to study it. We sketch the areas of overlap and disagreement among these literatures.

What is an urban system? Basic terminology An urban system can be roughly defined as a set of cities (ranging from large metropolitan areas to small villages) that are connected by linkages that are referred to as edges, links, arcs, or connections, that form a network of cities. These connections range from physical networks, notably hydrological systems of rivers and oceans, to conceptual networks such as personal relationships across cities. Between these two extremes are infrastructure: such as rail and roads, freight, commuters, migrants, and telecommunications f lows (mass media, social media, e-mail, etc.). It’s common to create a network of cities from a bipartite network where cities are “connected”, through

DOI: 10.4324/9781003316657-5

78  Benjamin Preis

common linkages to other entities, like firms with a presence in multiple cities (Derudder, 2021). As such, when codified, these connections have multiple discrete properties. Foremostly, they can be weighted or unweighted: A weighted edge can be marked with the number of entities that f low between cities, for example, the number of f lights between Tokyo and Brisbane or number of ships traveling between Accra and Monrovia. An example of an unweighted edge is the existence of a path between two villages, or a series of underwater cables where there is no information about usage. These edges also have edge properties that signify their (a) capacity, how many lanes a highway has or how many trains travel between two cities, (b) cost of traversing the edge, which can be measured in travel time, or monetary cost, and (c) other auxiliary properties such as who owns the infrastructure, whether it crosses administrative boundaries, or whether it is in operation. Edges can also be directed or undirected. An undirected edge only indicates an existence of a relationship, such as a bilateral agreement between cities. A directed edge, on the other hand, indicates that relationships may not be mutual, or may not be equal if they are mutual: migrants may move from Caracas to Bogotá, or goods from Shanghai to Los Angeles. Thus, in a network with directed edges, a pair of nodes may have zero, one, or two edges connecting them, indicating a lack of a relationship, a unidirectional relationship, or a bidirectional relationship. Weighted edges may also be combined with directed edges, leading to a weighted, directed network. Cities themselves can be defined as nodes in a network. Nodes can have properties that are intrinsic to the city: such as population size, gross domestic product, or climate, and they also can have properties that describe their “role” in the network, inf luenced by the characteristic of the node’s edges. These roles in the network can be distilled into its centrality (i.e., whether it is easily reachable and does it help conduct the overall f low), and its degree (how many connections it has?)

Why understand cities as networks? One reason to understand cities as networks is to use f lows and connections to give meaning to the boundaries of a city, a region, or an urban system. As a recent review of the urban networks literature stated, “[o]ne might even suggest that ‘city’ is simply one name we give to a spatially concentrated network of human interactions” (Neal et al., 2021, 10). Using this observation, one might draw a map of Great Britain using a network representation of human interactions – in the form of telephone calls (Ratti et al., 2010). In turn, defining cities through networks provides an opportunity to use mathematically normative, rather than sociopolitical, rules to define the bounds of an urban system. From a mechanical point of view, nodes and edges are the main components of translating an urban system/network into a mathematical, graph-based,

Winding paths through urban systems and urban networks  79

and/or visual model. The overall goal of this translation, or schematization, is to help with the analysis of the system: what cities are the key players, what cities produce vs. consume, how the system evolved, and how it might change in the future. We can consider the relationship of cities themselves, through f lows of capital or goods and services, through interconnections and relationships. As such, by conceptualizing cities as networks, we’re able to break down the components of what we mean by nodes and edges in the study of urban systems. Is a city a dense cluster of human connections, discovered through a community detection algorithm? Or is a city a node, and edges are migrants from another city? Or is a city a node in a bipartite graph, where companies are connected to cities based on where they have branches? Network analysis allows us to study urban systems as nested connections – connections within cities, connections among cities, and connections of cities. It’s incredibly important to be precise when defining the methodological approach, for there are ways to understand the same system as different network types. For example, much of the research dealing with street networks since the 1970s took them to be planar networks, with intersections as nodes and streets as links (Ducruet and Beauguitte, 2014). This makes perfect sense, as network links often represent flows between edges. However, street networks can also be reconstituted such that streets are nodes and intersections are edges, in what is known as a dual graph (Neal, 2013; Barthelemy, 2016). Dual graphs are useful to researchers studying simplest paths through cities, as well as the continuity, connectivity, and depth of the street network. Choosing which set of ideas to represent as nodes and which to represent as edges significantly changes the network measures, highlighting the importance of node definition. For example, Neal (2013) notes that a street’s planar (or primal) graph has characteristics of a random network – where each node has roughly equal number of edges – while its dual graph is often scale-free – where the distribution of the number of edges over all the nodes in the network follows a power law distribution.

Different ways of studying urban systems as networks An underlying tension within the study of urban systems through a network analysis lens is whether one ought to study the network itself or whether network methods are used toward understanding existing theories. These differences may align with the disciplinary backgrounds of those who are studying the network (Hidalgo, 2016), but it is not nearly so simple. These different approaches to the study of urban systems arise from scholars attempting to answer different questions, dealing with different data, and working in different contexts. These different ways of studying urban systems leads to not infrequent crosstalk within the urban studies literature, as what is considered to be studying urban networks appears quite different.

80  Benjamin Preis

To study the network in-and-of-itself is to take the topology of the network, and probe its structure, and perhaps its implications. For instance, defining the boundaries of a region is far from a settled issue (Duranton, 2021). Royuela and Vargas (2009) use the structure of commuting and migration networks in Spain to define the boundaries of “Housing Market Areas”. Thus, the network structure is meant to be representative of boundaries of urban systems that exist in the real world. As another example, Andersson et al. (2003) argued that if trade exhibits properties of scale-free networks, then so too should urban land costs. In this work, the authors start from the argument that trade networks exhibit the same properties as other scale-free networks: there is some level of preferential attachment (the rich-get-richer), coupled with new entrants and some degree of randomness in deciding with whom to trade. Insofar as trade takes place in space, the land prices on which that trade takes place would, in turn, be expected to be scale-free. In this way, they use network theory to understand urban systems of trade and land prices. How one incorporates networks into the study of existing theory can take many forms. Meijers (2005) and Burger and Meijers (2012) sought instead to utilize network methods to illuminate existing debates within the regional studies literature. Central place theory has been one of many competing theories that seeks to explain why, in a given region, one often finds one large, central, place, and many smaller satellite places surrounding it (see Mulligan et al., 2012 for a review of the central place theory literature). Central place theory assumes monocentricity – one central node surrounded by many smaller nodes. Polycentric urban form, by contrast, asserts that there are regions composed of multiple places of roughly the same scale. Meijers (2005) analyzed cooperation networks between cities within a region, while Burger and Meijers (2012) analyzed commuting and shopping trips. In their work, they find some degree of polycentricity in the regions they study, though they differentiate between morphological and function polycentricity. At the national level, Irwin and Kasarda (1991) examined the growth and evolution of airline networks in the US. By capturing network centrality from the airline network, they then used that information in various regression models to understand the impact of this infrastructural network on employment in US metropolitan areas. Political and economic geography are other disciplines have taken ample interest in urban systems and networks. From a political economy perspective, one might study governmental interconnections as indicators of political cooperation or antagonism. Governmental interconnections might take the form of interlocal agreements, participation on regional councils of government, or other forms of cooperation between municipalities. Global examples of urban government networks include the “C40 Cities”, a global network of local governments committed to addressing climate change through local solutions, and the interlocking connections comprised of sister cities. Interconnections between cities through firms are also a sort of “urban system”, albeit a bit of an abstraction. Cities have always been sites of commerce.

Winding paths through urban systems and urban networks  81

Local agglomeration economies are what define many of the world’s superstar cities today, from finance in London and New York, to technology in San Francisco and Shenzhen, and entertainment in Los Angeles and Mumbai. In the 1980s, urbanist Jane Jacobs studied the components of strong urban economies, and argued that economically successful cities source raw materials from their “supply regions”, and in turn produce goods for export to other cities ( Jacobs, 1985). Jacobs had a particular view of urban economies: cities have supply regions and trade partners, and cities grow through “import replacement”, that is, producing goods that it previously imported from its trade partners. While trading partners went far beyond regions, it was a view of urban economies limited to trade within the nation-state. In the late 20th and early 21st centuries, owing to globalization and increased inequality between cities, many scholars began to focus on so-called “global” (Sassen, 2004) or “world” (Taylor, 2004) cities. These world cities were connected not so much by trade among them, but by capital f lowing through them. These are the cities where the same service firms operate, where one international conglomerate may have many branches, and where capital f lows from one to another in the pursuit of returns on that capital. In studying these global cities, scholars have moved beyond examining physical trade of individuals or commodities, instead focusing on the rising importance of financialization in the modern economy. These world cities are connected not just by transportation infrastructure – like airports – but also by digital infrastructure, like internet cables.

Challenges in defining a city as a network As discussed in the chapter thus far, the different disciplinary and methodological approaches to studying urban systems as networks lend itself to various challenges. Here, we discuss four, related to node definition, edge definition, centrality measures, and data availability. Within the context of urban networks, the problems associated with node definition appear in many forms. The first challenge with most urban networks is that they are spatial networks, meaning that the nodes are assumed to have a fixed location in space, and tie formation may be inf luenced by that space. Ducruet and Beauguitte (2014, 303) refer to this as “physical embedding”, meaning that nodes are “grounded in a physical (Euclidian) space, which in turn constrains the multiplication of links and orientates [sic] the layout of the network, with the crucial importance of borders”. Where applying network methods to an urban system that is not originally a network, decision points abound. For instance, if one were to examine the daily mobility of people within a city, one might define nodes as neighborhoods, while the links between the nodes are based on movement of people between them (Candipan et al., 2021). People are thus aggregated up to neighborhoods to understand the patterns of mobility within the city. However, which people and which neighborhoods? One might limit the neighborhoods only to

82  Benjamin Preis

those block groups within central cities, eliminating all neighborhoods outside. We have node definition decisions – block groups and not people – and node set decisions – central city residences, and not suburban residences. Some cutoff is clearly necessary, daily mobility is constrained by space, but whether jurisdictional borders (and not some more organic definition) are used is a clear challenge. This, of course, is a key objective among those who use f lows to define morphological urban areas, as opposed to their administrative definitions (see Duranton, 2021). Questions of borders and aggregation are a central problem within the context of urban systems. Known as the “Modifiable Areal Unit Problem”, it plagues studies of segregation, where different definitions of a neighborhood – block, block group, census tract, etc. – result in different measures of segregation (see, for example, Barthelemy, 2016; Butts, 2009). As Barthelemy (2016, citing Arcaute et al., 2015) demonstrated, one can define a city at many different scales, using density gradients and different cut-offs of distance. The challenges of the modifiable areal unit problem are most acute when the nodes that are being studied are comprised of constituent parts – people are inside blocks inside cities inside regional administrative regions within counties. By choosing a particular level of aggregation, the researcher chooses their node definition and node set, but, in turn, leaves the possibility that another node definition would have yielded a network with a different topology. Cascading problems come from node definition. Specifically, how does one aggregate the constituent parts of a node and its interconnections? Candipan et al. (2021) choose to have weighted edges where “each weight indicates the average proportion of visits by residents from a block group to all other block groups”. Thus, by defining nodes as neighborhoods, the researchers have to then define residence of the constituent parts, and links by proportion of trips from the nodes’ constituent parts. Had they chosen a different definition of a node, then a different definition of links may have followed. Another example may demonstrate this challenge: Irwin and Kasarda (1991) examined the relationship between airline traffic and employment growth at the metropolitan statistical area (MSA). Many MSAs have more than one airport; indeed, some major cities have more than one airport. Thus, they chose to add the airline traffic among all the airports within the MSA. Had they chosen a different node definition, their aggregated airline passenger data would have yielded different results. A related problem in the study of urban systems is that many network measures were not developed with weighted networks in mind. Many centrality measures, for example, were based on simple unweighted graphs, where an edge was either present or not present. Yet many networks within urban systems are weighted: it is the number of airline connections between two cities, or the number of migrants that move between them, the amount of capital that f lows between them, or the total value of the traded goods that move between them. Those who study urban systems have three paths

Winding paths through urban systems and urban networks  83

forward. They can either choose to convert their weighted edges to binary edges, using a cut-off point to indicate whether a connection is sufficiently “there” to represent an edge. Or, alternatively, they can use adaptations of network measures that have been generalized for weighted networks, bearing in mind that they may not be directly analogous to the unweighted measures. For instance, one can modify the calculation for degree centrality to include the relative weights of a node’s edges, resulting in a similar, but not identical, measure of node centrality (Candeloro et al., 2016). Alternatively, they can create their own metrics, measures, and tests, appropriate to the peculiarities of spatial networks. Phillips et al. (2021) created an “equitable mobility index” and a “concentrated mobility index” derived from centrality and distance measures designed for weighted graphs. Similarly, Neal et al. (2022) explicitly developed methods to identify network “backbones” with weighted networks in mind. Even absent the challenges with weighted networks, centrality measures are also important sources of disagreement within the study of urban systems. As students of network science are well aware, measures of centrality within a network are legion: degree, betweenness, closeness, and eigenvector centrality are but four well known measures. Many who study urban systems, however, find the need to define niche centrality measures, owning to the nature of the urban systems under study. Irwin and Hughes (1992), for example, argue that Bonacich centrality – which has a sensitivity parameter that can take on values between −1 and 1 – can be used to correspond to different theories in the urban context, central place theory, human ecological conceptions of the city, and economic base theory. In turn, they apply Bonacich centrality to different urban system networks: airline patterns, commuting patterns, and central places. Neal (2011) similarly argues that standard centrality measures don’t make sense in some urban domains, in his case, the “world city” network as proposed by Taylor (2004). Instead, Neal proposes the use of recursive centrality, and a related concept of recursive power. Neal finds that cities in the world city network – as represented by the internet backbone – are often either powerful or central, though some, such as New York, London, and Frankfurt, are both. A word of general challenge in the urban systems literature: data are often unreliable, unavailable, or non-existent. Rare is the study in urban systems that exists with data collected for the express purpose of studying the network that researchers construct. Instead, data are adapted from existing sources, be it “big data” from cell-phone calls or GPS pings, administrative data like tax returns or Census responses, business data available online or in legal filings. These sources of data often mean that the network is incomplete or, perhaps, simply wrong. Perhaps migration data contains information about in-nation migration, but immigrants are simply absent from the data, or all lumped in together in a catch-all “other” category. Branches of a firm are identified from their website, but subsidiaries aren’t captured because they aren’t as obvious. Street networks are studied using available maps from the official

84  Benjamin Preis

state-provided files, but they’re missing new subdivisions and urban sprawl. Urban systems, being loosely defined and a collection of different interacting parts, are systems where the existing data are brought to bear on questions that the researchers hope to investigate; given this reality, researchers are often left contorting the available data into the shape they wish it took.

One concept, five approaches to the problem Thus far, we have discussed the reasons to study urban systems as a network, different ways of conceptualizing them, and the challenges therein. In doing so, we have highlighted a collection of loosely connected studies around urban systems, from a variety of fields. In this section, we highlight five studies that all purport to measure the same thing: commuting networks. Commuting networks are among the most ubiquitous in the context of urban network research today (Irwin and Hughes, 1992; Royuela and Vargas, 2009; Goetz et al., 2010; Tilahun and Levinson, 2011; Caschili and De Montis, 2013; Dash Nelson and Rae, 2016). Yet the papers that examine them agree on little: node definition, link definition, and the appropriate measures to study in the context of an urban network. Thus, examining these five studies, all in the US context, demonstrates many of the issues around node and edge definitions that are so crucial to urban networks. Before diving into the papers themselves, however, it is worth noting the disciplinary diversity among the authors. They are found in departments and centers as far apart as Sociology (Irwin), Demography (Hughes), Economics and Business (Vargas), Agricultural Economics (Goetz and De Montis), Regional Quantitative Analysis (Royuela), Public Affairs (Tilahun), Civil Engineering (Levinson), Geography (Dash Nelson), Urban Studies (Rae), and Advanced Spatial Analysis (Caschili). None of the five articles analyzed here are in the same journal, and may not be considered in the same field, ranging from Cities to Transportation Research Part A to Social Forces. While all ostensibly social scientists and all publishing in journals related to urban studies, it is perhaps no surprise that these articles differ so greatly. Turning to the papers themselves, Irwin and Hughes (1992) examined commuting patterns in the Anaheim and Oklahoma City MSAs using 1980 data. Their goal in the paper was to show how different measures of centrality – degree, betweenness, closeness, and competitiveness – mapped onto different sociological constructs. They argue that degree centrality is the most appropriate measure when examining commuting networks, since commuters move from node A to node B, but don’t have recursive effects throughout an entire network. In their analysis of the network, they make a number of key (and justifiable) definitional choices: nodes are at the level of the jurisdiction (where cities are limited to municipalities greater than 25,000 in population), and only jurisdictions within the MSA are included. When visualizing the network, they make an additional restriction, only including those f lows with more than 5% of the total proportion of commuting f lows. We see here

Winding paths through urban systems and urban networks  85

definitions of the node (cities) and node set (in the MSA, greater than 25,000 people, more than 5% of the f lows.). In demonstrating the applicability of a centrality measure to a paradigmatic set of questions, these methodological decisions may be justified, but absolute cut-offs may not work for a larger whole-of-the-nation study of commuting patterns. Tilahun and Levinson (2011), writing nearly 20 years later, had access to significantly more detailed data – but only for the state of Minnesota. Thus, they examined home and work location at the block level in the Twin Cities. While Irwin and Hughes were interested in the relationship between cities, Tilahun and Levinson were interested in the relationship between people, asking: are people who live on the same block more likely to also work on the same block? Thus, even though both sets of authors were studying commuting networks, the unit of analysis, unit of aggregation, and methodological choices were wholly different. Indeed, Tilahun and Levinson did not even calculate any centrality measure, instead they used inference methods like the quadratic assignment procedure to see if the co-incidence of working and living together was more likely than a null model. Thus, we can see here how commuting networks are a type of multilevel networks, where one can look either at individual commuters, or at aggregated f lows of commuters. Goetz et al. (2010) studied the relationship between commuting networks and economic growth. They constructed commuting networks at the county level and examined degree distribution and entropy to then use in regressions on economic outcomes. With this third commuting network paper, we see a third definition of a node of interest: counties. A challenge with county-level nodes, as noted by the authors, is that counties are not uniformly sized, nor of uniform population. Thus, aggregating up to the county level means accepting a heterogeneity of nodes that is often unaccounted for in the actual statistical models. Herein lies a key challenge with much urban system research: the data availability leads to the units of analysis that are possible to study, even if they are not the most sensible. Caschili and De Montis (2013), like Irwin and Hughes, wanted to relate network measures, like degree centrality, betweenness centrality, and strength, to measures of interest in their field – in this case, transit accessibility measures. Unlike Irwin and Hughes, however, Caschili and De Montis took the entire US commuting network as a whole, keeping strange connections like the more than 200 commuters who apparently go from Los Angeles County to the counties containing Seattle, Chicago, Dallas, Houston, among others. Although Irwin and Hughes explained why betweenness centrality is a poor metric for a commuting network, Caschili and De Montis make it central to their analysis. The disciplinary differences among the papers may shine through here: while Irwin and Hughes were seeking sociological meaning, Caschili and De Montis sought to analyze the network in terms of complexity and similarity, leading to different concerns in the measures pursued. Finally, Dash Nelson and Rae (2016) studied the US commuting network, using the same data source as Caschili and De Montis. Rather than

86  Benjamin Preis

trying to understand the accessibility of the network, however, Dash Nelson and Rae use the structure of the network to create definitions of regions. Unlike the other four papers considered here, Dash Nelson and Rae used the Census tract as the level of aggregation. A census tract is composed of multiple Census blocks, and has a population of approximately 4,000 people. A fifth paper on commuting, a fourth node definition. Dash Nelson and Rae used a cut-off of 160 kilometers as the cut-off for a reasonable commute – a long commute, to be sure, but a correction to the extreme commutes found in Caschili and De Montis’s paper. Using these data, they then partitioned the United States into 50 “megaregions”, using a partitioning algorithm, coupled with a visual heuristic approach. Interestingly, the partitioning algorithm is not an explicitly spatial algorithm – it is concerned with the network topology, absent its geographic features – thus requiring, in their opinion, some level of human input.

Concluding thoughts The literature encompassing urban systems is necessarily broad. To conceptualize cities-as-systems has been adopted across various urban-inclined disciplines for decades: as noted in the Introduction to this book, Berry suggested a cities-as-systems approach as early as 1964 (Berry, 1964). Left unmentioned in this chapter are many offshoots of the urban networks literature, such as the recent literature on scaling laws on cities (see, e.g., Bettencourt et al., 2010), which argues that differences in urban outcomes – typically bounded with an urban system comprising a nation-state – are derived from fundamental properties of cities, typically their population or physical area. Disciplines ranging from physics to geography have embraced the urban systems concept, and urban network analysis to accompany it, yet they often speak different languages, consider different nodes, assign different types of edges, and ask different questions. Demonstrating the diversity of this academic interest, Peris (2016) created a co-citation network of authors studying “urban networks”. With 396 authors, he found 10 distinct clusters, including clusters centered on authors cited here, such as Michael Batty, Peter Taylor, Carlo Ratti, and many others. Disciplinary clusters are clear, but not absolute, with co-authorship among differing traditions apparent, even if shy of a complete graph. The question, moving forward for the urban systems literatures, is how these insights can be put to use. Once a system of cities has been analyzed, what do we learn from it that we did not know before? What of these insights can be translated for the urban minister, or the urban administrator, the CEO, or the citizen? Yes, cities exist as systems and are connected through f lows, but what alters those f lows, or alters the topology of the network? How did Tokyo come to be the size that it is, and now that it is that size, how can its infrastructural network change its carbon footprint? The urban systems literature has proven that it can provide valuable insights about the structure

Winding paths through urban systems and urban networks  87

of cities and the structure of a system of cities, but it must now prove that it can provide value toward adapting these cities for the rest of the 21st century.

Acknowledgments The author would like to thank Clio Andris for her thoughtful comments, edits, and advice in the drafting of this chapter, in addition to the helpful comments and recommendations from César Ducruet.

References Andersson C., Hellervik A., Lindgren K., Hagson A., Tornberg J. (2003) Urban economy as a scale-free network. Physical Review E, 68(3): 036124. Arcaute E., Hatna E., Ferguson P., Youn H., Johansson A., Batty M. (2015) Constructing cities, deconstructing scaling laws. Journal of the Royal Society Interface, 12(102): 20140745. Barthelemy M. (2016) The Structure and Dynamics of Cities. Cambridge University Press. Berry B.J.L. (1964) Cities as systems within systems of cities. Papers in Regional Science, 13(1): 147–163. Bettencourt L.M.A., Lobo J., Strumsky D., West G.B. (2010) Urban scaling and its deviations: Revealing the structure of wealth, innovation and crime across cities. PLOS ONE, 11: e13541, https://doi.org/10.1371/journal.pone.0013541 Burger M., Meijers E. (2012) Form follows function? Linking morphological and functional polycentricity. Urban Studies, 49(5): 1127–1149. Butts C.T. (2009) Revisiting the foundations of network analysis. Science, 325(5939): 414–416. Candeloro L., Savini L., Conte A. (2016) A new weighted degree centrality measure: The application in an animal disease epidemic. PLOS ONE, 11: e0165781, https:// doi.org/10.1371/journal.pone.0165781 Candipan J., Phillips N.E., Sampson R.J., Small M. (2021) From residence to movement: The nature of racial segregation in everyday urban mobility. Urban Studies, https://doi.org/10.1177/0042098020978965 Caschili S., De Montis A. (2013) Accessibility and complex network analysis of the U.S. commuting system. Cities, 30: 4–17. Derudder B. (2021) Bipartite network projections of multi-locational corporations: Realising the potential. Geographical Analysis, 53(2): 383–393. Ducruet C., Beauguitte L. (2014) Spatial science and network science: Review and outcomes of a complex relationship. Networks and Spatial Economics, 14(3): 297–316. Duranton G. (2021) Classifying locations and delineating space: An introduction. Journal of Urban Economics, 125: 103353. Goetz S.J., Han Y., Findeis J.L., Brasier K.J. (2010) U.S. commuting networks and economic growth: Measurement and implications for spatial policy. Growth and Change, 41(2): 276–302. Hidalgo C.A. (2016) Disconnected, fragmented, or united? A trans-disciplinary review of network science. Applied Network Science, 1(1): 6. Irwin M.D., Hughes H.L. (1992) Centrality and the structure of urban interaction: Measures, concepts, and applications. Social Forces, 71(1): 17–51.

88  Benjamin Preis Irwin M.D., Kasarda J.D. (1991) Air passenger linkages and employment growth in U.S. metropolitan areas. American Sociological Review, 56(4): 524–537. Jacobs J. (1985) Cities and the Wealth of Nations: Principles of Economic Life. New York: Vintage Books. Meijers E. (2005) Polycentric urban regions and the quest for synergy: Is a network of cities more than the sum of the parts? Urban Studies, 42(4): 765–781. Mulligan G.F., Partridge M.D., Carruthers J.I. (2012) Central place theory and its reemergence in regional science. The Annals of Regional Science, 48(2): 405–431. Neal Z. (2011) Differentiating centrality and power in the World City Network. Urban Studies, 48(13): 2733–2748. Neal Z. (2013) The Connected City: How Networks Are Shaping the Modern Metropolis. 1st ed. Metropolis and Modern Life. New York, NY: Routledge. Neal Z., Derudder B., Liu X. (2021) Using urban networks to gain new insight into old questions: Community, economy, bureaucracy. Journal of Urban Affairs, 43(1): 2–15. Neal Z., Domagalski R., Sagan B. (2022) Analysis of spatial networks from bipartite projections using the R backbone package. Geographical Analysis, 54(3): 623–647. Nelson G.D., Rae A. (2016) An economic geography of the United States: From commutes to megaregions. PLOS ONE, 11: e0166083, https://doi.org/10.1371/ journal.pone.0166083 Peris A. (2016) Penser les Villes en Réseaux: Une Analyse des Théories sur les Liens Interurbains. Master Dissertation in Geography, University of Paris I Panthéon-Sorbonne. Phillips N.E., Levy B.L., Sampson R.J., Small M.L., Wang R.Q. (2021) The socialiIntegration of American cities: Network measures of connectedness based on everyday mobility across neighborhoods. Sociological Methods and Research, 50(3): 1110–1149. Ratti C., Sobolevsky S., Calabrese F., Andris C., Reades J., Martino M., Claxton R., Strogatz S.H. (2010) Redrawing the map of Great Britain from a network of human interactions. PLOS ONE, 5(12): e14248. Royuela V., Vargas M.A. (2009) Defining housing market areas using commuting and migration algorithms: Catalonia (Spain) as a case study. Urban Studies, 46(11): 2381–2398. Sassen S. (2004) The global city: Introducing a concept. The Brown Journal of World Affairs, 11(2): 27–43. Taylor P.J. (2004) World City Network: A Global Urban Analysis. London & New York: Routledge. Tilahun N., Levinson D. (2011) Work and home location: Possible role of social ­networks. Transportation Research Part A, 45(4): 323–331.

4

The implications of duality of trans(port) systems Evidence from Wusongkou International Cruise Port James J. Wang, Adolf K.Y. Ng, and Yui-yip Lau

Introduction While there have been many geographical studies on transport for the past three decades since the first issue of the Journal of Transport Geography published in 1993, transport geographers still struggle to reconnect with mainstreamed economic geographies, in order to position this sub-discipline. In general, the common methodologies and theories such as cluster analysis (de Langen, 2004), evolutionary approach (Wang and Xiao, 2015), geographical political economy approach (Pike et al., 2016), cognitive spaces (Essletzbichler, 2009), and the theory of global production network (Jones, 2018) are mainly used to analyze the general transport and port sectors. However, their ‘core business’ – transport geographies, has become unfocused. Considering two facts about a blurring trend: (1) some key ‘spatial’ characteristics of the transport system (i.e., associations between demand, nodes, and networks), such as the hub-andspoke structure, which was considered a geographical phenomenon in essence (e.g., Markusen, 1996), have been taken as basic management or engineering concepts by transport operators, which makes it not really associated with geography anymore; and (2) transport geographers tend to contribute more by applying geography-related theories, such as eco-system analysis, industrial cluster analysis, and activity-based human behavior analysis directly to transport system, which can be regarded as impact analysis of freight circulation and behavior analysis of people mobility. On the one hand, the service providers are expected to offer travelers f lexible, easy, environmentally, and reliable sustainable everyday travel. On the other hand, freight circulation refers to the process and system by which goods are gathered, transported, and distributed within environments. Freight circulation can consist of manufacturing plants, airports, seaports, distribution centers, and warehouses that are linked by a network of highways, rail yards, and railroads that foster goods to get to the final destinations. The term ‘impact’ or ‘behavior’ here may (not) have anything to do with geography. If the first fact is taking away something that may otherwise be considered geography by non-geography transport disciplines, the second fact is representing how do transport geographers explore new research areas that broaden the horizon of their own discipline, if thinking positively.

DOI: 10.4324/9781003316657-6

90  James J. Wang et al.

However, these efforts of transport geographers do not seem to have changed much about the perception and understanding of non-transport geographers (mainly economic geographers on the role of transport in geography). To most of them, the central role of transportation and its improvement is to shorten the time/cost distance, to link together the factors of production in a complicated association between consumers and producers, and to allow and facilitate the physical movement of people or goods between places. As such, transport might create significant contributions, such as attracting valueaddition (e.g., jobs, firms), creating technological progress and innovation, and developing the local socio-economic environment. However, transport operation often generates various negative externalities in environmental degradation, such as hazardous waste disposals, unrenewable fuel consumption, and air pollution (Ducruet et al., 2019). The relative connectivity and accessibility of a place in comparison with other places are more important than how transport systems are planned, financed, organized, operated, and controlled in providing transport products – the shipment of people or goods, no matter how to transport geographers claim that geography matters in these processes of planning, financing, organization, control, and operation of transportation. Besides, geographical research on transport seems unable to shape the views of non-geography transport scholars about the role of geography either. To most of them, the real core of transport research is to optimize transport systems or the provision of them, and geography or space is not in this core. At most, geography is regarded as pre-conditions or impacts, of transport systems. For instance, the world population distribution and dynamics of land use patterns may condition the demand pattern of travel, and through a dynamic process, an additional line of transit may alter the distribution or urban activity pattern, which turns out to be a geographical outcome. A fundamental reason for the ‘geography’ aspect not being seen as a transport system per se is that it cannot be altered as endogenous components or factors to improve a transport system. As long as the research aim is to improve the transport system per se, no matter financial, managerial, social, or environmental concerns, geographical aspects are often treated as external factors or environment (Ducruet et al., 2019). Although non-transport geographers pay attention to the external, space-shrinking effect of transport systems while non-geography transport researchers focus mainly on transport systems per se for their internal improvement, there are occasions where they meet. For example, when there is a new port that needs to be built or an extra berth to be expanded or added, both internal and external, social and fact geographical aspects have to be assessed at the same time. This is particularly true if huge subsidies from some non-transport budget sources are involved. Ironically, after a period of operation of this new port, what happens most often is that a large amount of initial capital investment as ‘sunk cost’ and wiped out, and ‘after-analysis’ on social or geographical gain (or loss) outside the transport system is rarely

The implications of duality of trans(port) systems  91

carried out (Ducruet et al., 2019). By “sunk cost”, it literally means that the capital invested initially would not be gained back through the operation of the transport system itself. In practice, transport economists do not need to deal with it anymore since it is out of the question as long as the transport system is concerned. However, our question is, why were the social or societal benefits considered in the first place by measuring the potential increase in accessibility and connectivity of a place and the benefits to be brought by them, noting that these measures are of geographical meaning in essence? In this chapter, we address this question with the help of the cases of ports as illustrative examples. The rest of the chapter is as follows. Next section describes the concept and keynotes of the duality of the transport system. It is followed by a case study to demonstrate the concept of the duality of transport system. After, we discuss the theoretical importance of understanding the duality of the transport system. Finally, the conclusion can be found in the last section.

The duality of transport systems There are people outside the transport system who are closely concerned with the development of transport but not because of its final product – the shipment of cargo or movement of people from one place to another, but the intermediate product – how a place, a city or even a region, becomes more connected and accessible than before. Figure 4.1 explains this point further.

Figure 4.1 The duality of transport system: an overall view Source: Authors

92  James J. Wang et al.

This figure exhibits how (non-)transport stakeholders have different perceptions, views, and interests in the provision of the intermediate and final product of transport. •

•

Transport stakeholders are concerned about: • How transport infrastructure as an intermediate product may provide; • In which way and what quality and quantity the final product of transport is provided; and • How to improve and gain from it. Non-transport stakeholders are concerned about: • How may transport system change the place value? • How to gain from this change by establishing business or property in relation to the transport provision? • How may the quality and quantity of transport services affect this gain?

In this case, transport provision is conceptualized as a dual process that crosses two domains – transport system and land-use or activity system. In Figure 4.1, the upper two quadrates illustrate that prior to the realization of transport, i.e., either cargo shipment or personal travel made, the provision of transport infrastructure, facilities, and services from construction to the service ready to serve, they are intermediate products. At this stage, the firms within the transport sector gain only when involved in the construction of infrastructure. However, to the land/property owners that are affected by these intermediate products, such as a port, the expected connectivity to be provided may be good enough to cause the rise of land value or property rent in the vicinity, if other components of accessibility are ready. For example, a cruise terminal in the downtown area can be upgraded to a multi-modal hub to connect other transport modes (e.g., the port of Vancouver, BC, Canada). As such, it could bring business opportunities and more visitors and boost the enlargement of the peripheral area of the terminal. In contrast, if a terminal is located in “a place of nowhere”, while the expected connectivity is still there, the land value would not rise due to the lack of accessibility to the activities that travelers, as well as other people, want to participate in (c.f. Webster, 2010). Another illustrative example is Jebel Ali port in Dubai. A key part of Jebel Ali port is the Jebel Ali Free Zone. The free zone has attracted nearly 4,000 businesses emerging from more than 100 countries. To strengthen Dubai’s position as vital transshipment hub for the Subcontinent, Dubai considered establishing Jebel Ali International Airport. To this end, Jebel Ali port has created comprehensive sea-air intermodal transport hub and formed the port-industrial cluster ( Jacobs and Hall, 2007). However, we have also seen cases in China where local governments decided to locate new transport facilities in a relatively new and remote township or district, expecting the growth pole effect to be brought in by such facilities, and then gain from the land rent in the vicinity of the facility. Also, it is possible for cities not

The implications of duality of trans(port) systems  93

to believe in any gains from such facilities and put it as far away from their core areas as possible. One may not appreciate such a way of thinking, but it does demonstrate that it is possible and that a major stakeholder (e.g., the local government) may push and lead a transport system to be constructed for land value gain, which has been considered as an externality in classical transport economics. In essence, the key point to our theorization is that locating new transport facilities in peripheral areas was not originally driven by cargo demands but by the willingness to take advantage of being that hub for ancillary activities which, by Fleming and Hayuth’s (1994) theory of geographical intermediacy, is of relatively high risk for being challenged by other cities. To generate a sustainable competitive advantage in the long term, Notteboom et al. (2019) addressed that ports require to provide various value-added services rather than purely transshipment activities. Otherwise, highly contestable transshipment f low poses the threats to the vulnerability of intermediate hubs. Therefore, we argue that although travel or shipment of cargoes as the final product is derived demand, investing and building transport infrastructure and facilities to create linkages or networks does not necessarily have to be a derived demand. It can be driven by the increase of connectivity which aims at gaining in agglomeration of many activities associated directly or indirectly with that transport system, and hence the place’s value (Bertolini, 1999). In this regard, some empirical works (e.g., Ducruet et al., 2015; Ducruet and Itoh, 2016) have been done to investigate the actual linkages between port f lows and spaces, revealing interdependencies but also mismatches. The two quadrates in the lower part of the figure indicate transport as a final product produced by the transport service providers/operators (as the quadrate on the lower-left corner) and consumed by the users (as the quadrate on the lower-right corner). Much fewer explanations are needed here, as these two quadrates represent the typical relationships in classical, mobility-centered transport economics: the supply of services on the left-hand side is fulfilled by the demand from the users on the right-hand side; the level, complication, and quality of transport services are improved and the operators gain from such improvement by better satisfying the door-to-door travel or transport demand from the users when they enjoy good transport connectivity to access their activities at destinations. Regarding ports, it mainly handles transit trade by sea-land (gateway) and/or sea-sea (hub). In fact, this conventional mobility-centered view can be best described when focusing on quadrates one, three, and four together in Figure 4.2. Transport suppliers provide hardware as conditions and operate services to finalize the transport product of place utility on the one hand; on the other hand, users of this system, either the travelers or the cargo consigners, consume the movement of people or cargo. It is a process of mobility gain directly for both the suppliers and consumers of the transport system. However, if we shift to another perspective to interpret the same process by looking into the other three quadrates (one, two, and four) in Figure 4.3,

94  James J. Wang et al.

Figure 4.2 The duality of transport system: typical relationships seen in classical, mobility-centered transport economics Source: Authors

Figure 4.3 The duality of transport system: concerns in urban or regional planner’s and place-stakeholders’ mind Source: Authors

we could find a different set of considerations between space/place providers for transport and other activities, and the transport users as place seeker and activity participants. From this perspective, the stakeholders of a place (such as a city government) look for the accessibility gain (see detailed discussion in Webster, 2010), and so are the users of transport or even non-transport

The implications of duality of trans(port) systems  95

users who want to gain from the agglomeration of activities in a place with improved transport connectivity and hence higher accessibility. Thus, it is a process of accessibility gain for the transport place makers and the users of the connected spaces where accessibility is created, and place value is added. In this regard, we emphasize that individuals, firms, and the government may not need to travel (or transport) anything to gain from the connectivity and accessibility so created.

Case study: Wusongkou International Cruise Port In this section, we use a study conducted by us recently on the Wusongkou International Cruise Port (WICP) in Shanghai, China as an illustrative example (Sun et al., 2020). We found that the overall accessibility of onshore tourist attractions with WICP as the starting point is acceptable and that the access time of most attractions (82%) within the total of Shanghai under the optimal path is 60–120 minutes. This illustrates our earlier discussions where the existence of transport facilities does not only lead to gain in transport but also gain in place value. However, there is still a large gap compared with excellent cruise ports, whose tourists can reach many leisure and entertainment facilities onshore by walking. Moreover, increased traffic congestion makes the schedule of shore activities face great challenges. To optimize the accessibility of onshore tourist attractions so as to maximize the place value, we argue that Shanghai needs to build more intensive rapid road network traffic systems to provide more convenient traffic channels for distant scenic spots. For the central urban areas with traffic congestion, so as to enhance connection with the WICP, there should be more efforts to strengthen other non-port transport facilities, such as the reconstruction of the rapid viaduct and the construction of the internal ring expressway. Indeed, it needs to improve the construction of urban public transport, such as the subway and public transport, to alleviate traffic congestion, including access to the WICP. For the marginal suburbs, the improvement of the regional road network structure should be taken as a priority, and efforts should be made to increase regional highway density and build a higher-level highway system. Finally, it needs to optimize the transport support network around cruise ports and increase rapid transportation lines between the cruise port and the scenic spots. Also, our study suggests that the number of tourist attractions in Shanghai is unevenly distributed. Among them, the geographical density of spots in Huangpu, Pudong, and Hongkou has advantages but far away from the WICP. This creates a discrepancy where the duality of the trans(port) system in maximizing gains for both transport and place value largely fails to materialize. As distance determines the accessibility of attractions (Li and Wang, 2012), Shanghai should focus on excavating and integrating tourism resources in the surrounding areas of the port (e.g., potential natural and cultural resources in Baoshan, Jing’an, and Yangpu), and develop shore-based

96  James J. Wang et al.

tourism activities around the port area. For example, the Baoshan District of Shanghai, where the WICP is located, should rely on some local tourist attractions (e.g., Gucun Ecological Tourism Area, Luodian Cultural Tourism Area, Forest Park of Taiwan Wetland, Meilan Lake, Yangtze River Estuary Science and Technology Museum, and some inf luential festivals, such as Shanghai Cruise Tourism Festival, Meilan Lake Music Festival, Luodian Dragon Boat Culture Festival, and Baoshan International Folk Art Festival) and include them into onshore tourism itinerary elements to enhance the attraction (and thus competitiveness) of the transport facility (i.e., the cruise port). To further illustrative this point, we make a brief comparison on the locations between the WICP and another (competing) cruise terminal in the Shanghai cruise port system, namely the Shanghai Port International Cruise Terminal (SPICT). While SPICT is very close to the city core (2.2 km away from the Bund (Waitan)) and in close proximity with Shanghai Hongqiao Airport (19 km) and Shanghai Pudong International Airports (SPIA) (46 km), the WICP is located much farther away from both the city core (26 km away from the Bund) and the airports (38 km and 61 km from SHA and SPIA, respectively). Unsurprisingly, the WICP faces much more serious challenges on to connectivity, regional competitiveness, utilization, to name but a few.

Theoretical importance of understanding the duality of the transport system The conceptualization of transport duality brings geography back to the core of transport studies. More than a century ago, Karl Marx asserted that transport was something that was produced and consumed simultaneously within the same process (Harvey, 2004). Research on this process has been focused on narrowly defined transport. Anything, if not in this process of production and consumption of the final product, is treated as externalities. Our concept of duality exhibits the existence of an intermediate product of transport. No matter before or after a transport system is put into operation, the reconstruction and new shape of accessibility and connectivity pattern or geography in essence brought by transport infrastructure are recognized and valued not only by its user but also by the non-users. Some non-users of the transport system may be involved in the decision of building the system in the first place due largely to a geographical consideration: rising the place’s/ region’s value by better transport connection and higher accessibility from other places. Second, the duality concept offers a solid foundation in asserting that travel and cargo shipment are derived from demands, but not necessarily the case for transport facilities and infrastructures. Transport and travel have long been treated as “derived demands” in major transport economics textbooks (e.g., Button, 1992; Bamford, 2001). As derived demand refers to the demand for one good or service in one sector occurring as a result of demand from another, transport demand is regarded as derived from activities that travelers

The implications of duality of trans(port) systems  97

need to participate in at destinations, or from products to be used/consumed at destinations. This statement has been seen as an issue that deserves further investigation (Preston, 2001), and challenged from various perspectives. Salomon and Mokhtarian (1998) claim that enhancing accessibility does not necessarily come from the need to increase mobility; Rodrigue (2006) argues from his observations in the freight industry and globalization (also in Hesse and Rodrigue, 2004), that (freight) transport should be no longer treated as a separate identity as it has been largely integrated into supply chains. Now, we argue with the duality concept that the development of a transport system or its components, such as ports and airports, may (not) be primarily for transport per se: it can be just a major component of investments that aims to induce the spatial shift of development foci, although this purpose may fail to materialize. If it is a successful one, travel and shipment of cargo can be induced and derived. Third, our concept of duality helps to reconnect connectivity and accessibility studies with mainstream economic geography and the current trends in transport governance, policy, and planning. Ever since the quantification of human geography in the 1950–1960s, transport geographers have been studying transport provision by measuring the connectivity and accessibility of networks and nodes. They began to be applied more often by urban geographers and planners and incorporated into some benchmarking analyses/reports (e.g., see Taylor’s world city ranking). The conceptualization of transport duality would not only associate these measures properly with the other concrete measures considered by transport professionals (as given in the fourth quadrat in Figure 4.1), but also bring in transport’s contribution as a geographical component of place and spatial change. However, hitherto, transport geographers still seem to face the difficult task to internalize the value, such as integrating accessibility and connectivity back into more quantifiable formulae to make transport an integral part of the overall nodal or regional strategy.

Conclusion In contrast to the individuals who gain from ‘individual’ accessibility, i.e., the ease of reaching the site of his/her activity to engage, the stakeholders of the place with increased transport connectivity gain from “general” accessibility, realized from a general recognition of the place value due to a mix of multiple activities taking place within the same area (Webster, 2010). If desirable, a place maker may proactively create such a place by installing transport facilities and links them as instruments to increase the absolute and relative connectivity to other places to facilitate the (potential) increase of accessibility, hence the value of land (in the case of urban scale) or that of place (in the case of regional or national level) or the competitiveness of city (in the case of global level). In this way, geography is shaped by a supply-led process: investments in transport infrastructure – increased connectivity – agglomeration of

98  James J. Wang et al.

activity within around the place of increased connectivity – increased general accessibility – increased land value (and possibly even competitiveness). Such a complex process is more likely to take place when a nodal development of transport tends to be separately financed with local support. For example, building a port, which may presumably, serves any shipping lines interested in the market it supports, needs huge investments and the related stakeholder decisions on urban planning and land from local authorities. Building critical infrastructures is vital to economic development and societal well-being, and thus urban planners, transport operators, and policymakers are serious about the use of critical infrastructures to reduce regional rivalries and conf licts (Panahi et al., 2022). As highlighted by Lau et al. (2022), port infrastructures must be designed in a way so that it becomes resilient enough to recover rapidly from disruptions and to reduce potential socio-economic loss. Of course, not every part of a transport system has a land/territory component. For example, cross-ocean shipping routes do not involve (at least not explicitly) land or accessibility gains. However, these are non-exit or bypassing components, which make sense as a transport only when the other parts of the same system are performing in a process of accessibility gain. To some parts of the world, such as a desert or an oceanic area without any human settlement, bypassing is natural. Technological advancement in transport makes long-haul non-stop transport possible, which may lead to some places being skipped and less connected regularly and accessible as before, such as Honolulu, Hawaii, which was a refueling stop for every aircraft crossing the Pacific Ocean before the Second World War. Of course, we should not forget that it is a normal practice to subsidize (or to provide huge incentives to) transport infrastructures. In maritime transport, illustrative examples include the dredging of deep-water port access channels by local governments or various ways to selectively nurture new shipping routes at their embryonic stages, such as shipping services connecting Beibuwan ports (e.g., Fangcheng and Qinzhou ports in Guangxi Province in China) to Hong Kong and Singapore. For non-maritime transports, examples include airlines serving Shenzhen to several Latin American cities, as well as the Hong Kong International Airport (HKIA)’s incentive to airlines operated in Hong Kong that have successfully opened new air routes/ services through granting them big discounts on landing fees. In the case where the development of a transport system is considered a joint initiative of the transport operator and land/territory authority, through a public-privatepartnership (PPP) project on a new port terminal, for example, we see dual perspectives from multiple stakeholders: the transport sector examines how the “final products” may benefit the service providers and their users, while place/land/territory authorities investigate how the value may be added and the accessibility increased in the place/area to be connected. We, therefore, argue that this is the duality of the transport system: the process of accessibility gain is the mirror image or shadow of the process of mobility gain. Such duality should be regarded as the core area of transport geography research.

The implications of duality of trans(port) systems  99

This chapter provides the conceptual idea and general discussions on the duality of transport systems. Also, from transport geographers’ perspectives, the key research gaps, and agenda toward the duality of transport systems have been identified. For further research, we can use complex network analysis to assess the connections between ports, network density, average path length, and individual port level. Moreover, we can conduct critical reviews relevant to the duality of transport systems. As such, it may help transport geographers to recognize trends, frameworks, models, mechanisms, tools, methodologies, and measurement approaches for the implementation of policy and strategy in the duality of transport systems. This would further enhance (trans)port operational practices and support self-assessment so as to reprioritize activity. Furthermore, this study relied on a case of a cruise terminal. To generalize the research study, we may use multiple ports serving the same immediate city (Shanghai) and discuss how to place makers on the one hand and port actors on the other adjust or not their goals. To a large extent, the case of Shanghai as the import/export center served by multiple gateways could be exemplary of such a system, where metropolitan dynamics are vital. We believe that the chapter provides useful insight in enhancing our understanding on this important topic in transport geography.

Acknowledgments The support from CCAPPTIA (www.ccapptia.com) is gratefully acknowledged. Also, Adolf K.Y. Ng acknowledges the support from Chaire de recherche Couche-Tard sur les chaînes de valeurs globales. The usual disclaimers apply.

References Bamford C.G. (2001) Transport Economics (2nd edition), Oxford: Heinemann. Bertolini L. (1999) Spatial development patterns and public transport: An application of an analytical model in the Netherlands. Planning Practice and Research, 14(2): 199–210. Button K.J. (1992) Transport Economics (2nd edition), Aldershot: Edward Elgar. Ducruet C., Itoh H. (2016) Regions and material f lows: Investigating the regional branching and industry relatedness of port traffics in a global perspective. Journal of Economic Geography, 16(4): 805–830 Ducruet C., Itoh H., Joly O. (2015) Ports and the local embedding of commodity f lows. Papers in Regional Science, 94(3): 607–627. Ducruet C., Panahi R., Ng A.K.Y., Jiang C., Afenyo M. (2019) Between geography and transport: A scientometric analysis of port studies in Journal of Transport Geography, Journal of Transport Geography, 81: 102527. Essletzbichler J. (2009) Evolutionary economic geography, institutions, and political economy. Economic Geography, 85(2): 159–165. Fleming D.K., Hayuth Y. (1994) Spatial characteristics of transportation hubs: Centrality and intermediacy. Journal of Transport Geography, 2(1): 3–18. Hesse M., Rodrigue J.P. (2004) The transport geography of logistics and freight ­distribution. Journal of Transport Geography, 12: 171–184.

100  James J. Wang et al. Jacobs W., Hall P.V. (2007) What conditions supply chain strategies of ports? The case of Dubai. GeoJournal, 68: 327–342. Jones A. (2018) Geographies of production III: Economic geographies of management and international business. Progress in Human Geography, 42(2): 275–285. Lau Y.Y., Yip T.L., Dulebenets M.A., Tang Y.M., Kawasaki T. (2022) A review of historical changes of tropical and extra-tropical cyclones: A comparative analysis of the United States, Europe, and Asia. International Journal of Environmental Research and Public Health, 19: 4499–4517. Li L., Wang D. (2012) The impact of urban low-carbon public transport to tourist attractions’ accessibility – Suzhou city area as the example. Economic Geography, 32(3): 166–172. Markusen A. (1996) Sticky places in slippery space: A typology of industrial districts. Economic Geography, 72: 293–313. Notteboom T.E., Parola F., Satta G. (2019) The relationship between transshipment incidence and throughput volatility in North European and Mediterranean container ports. Journal of Transport Geography, 74: 371–381. Panahi R., Gargari N.S., Lau Y.Y., Ng A.K.Y. (2022) Developing a resilience assessment model for critical infrastructures: The case of port in tackling the impacts posed by the COVID-19 pandemic. Ocean and Coastal Management, 226: 106240. Pike A., MacKinnon D., Cumbers A., Dawley S., McMaster R. (2016) Doing evolution in economic geography. Economic Geography, 92(2): 123–144. Preston J. (2001) Integrating transport with socio-economic activity: A research agenda for the new millennium. Journal of Transport Geography, 9: 13–24. Rodrigue J.P. (2006) Challenging the derived transport-demand thesis: Geographical issues in freight distribution. Environment and Planning A, 38(8): 1449–1462. Salomon I., Mokhtarian P.L. (1998) What happens when mobility-inclined market segment face accessibility-enhancing policies? Transportation Research Part D, 3(3): 129–140. Sun X., Xu M, Lau Y.Y., Ng A.K.Y. (2020) Onshore product planning of cruise ports based on the accessibility of scenic spots: Evidence from Shanghai International Cruise Port. Paper presented at The Cartagena Dialogue on Cruise, Ports, and Cities, Cartagena, Colombia, February 20. Wang J.J., Xiao Z. (2015) Co-evolution between etailing and parcel express industry and its geographical imprints: The case of China. Journal of Transport Geography, 46: 26–34. Webster C. (2010) Pricing accessibility: Urban morphology, design and missing ­markets. Progress In Planning, 73(2): 77–111.

Part II

The dynamics of port systems

5

The European ports’ size dynamics and hierarchies Rania Tassadit Dial Currently, Gabriel Figueiredo De Oliveira, and Alexandra Schaffar

Introduction Maritime ports are not only the boundaries but also the connecting links between land and sea transport activities. The development of intermodal transport along with the worldwide spread of containerized traffic have contributed to the growth of international maritime trade, especially for semi-finished products which are highly dependent upon the time and cost performance of the transport chains. The growth in container traffic progressively led to the appearance of specific high-density trade routes and the development of a two-tier maritime container network (Yap and Notteboom, 2011). The latter is characterized, on one hand, by the extensive use of international hubs and load centers able to host simultaneously an important number of large vessels and, on the other hand, by the development of a series of lower indirect transport services carried out by smaller vessels between these hubs and smaller regional ports (Yap and Lam, 2006). The concentration of maritime traffic in a limited number of ports has been observed and well documented in the literature (Ducruet et al., 2009). According to the traditional theories on port growth (Taaffe et al., 1963; Ogundana, 1970), the first stages of maritime trade correspond to the concentration of traffic f lows in a limited number of ports. During these early stages, the traffic from/to the ports with the best connection to the hinterland increase allowing them to build a leading position in the maritime international trade network. In more advanced stages, the benefits of concentration may be offset by negative externalities due to congestion and limited amount of land. In the long term, secondary or new ports can challenge the hegemony of major ports by capturing an increasing share of port traffic, which leads to a geographical dispersion of maritime trade (Hayuth, 1981; Barke, 1986; Soppé and Frémont, 2007). More recently, Notteboom and Rodrigue (2005, 2010) have introduced the concept of port regionalization which describes the expansion of ports toward their hinterland, by forging links to the inland freight distribution centers. In order to capture more traffic and cope with regional competition, ports are encouraged both to expand their

DOI: 10.4324/9781003316657-8

104  Rania Tassadit Dial Currently et al.

inf luence further in the hinterland (hinterland-based) and to strengthen their integration into the regional maritime network (foreland-based). In most growth models, port development appears to be relatively pathdependent since future changes depend upon past decisions, infrastructures, and processes. However, path dependency sometimes may be overruled since a few ports seem able to deviate from an existing development trend ­ (Notteboom, 2009). This chapter aims to study the trends of ports’ hierarchies in Europe where the spatial distribution of port’s traffic is uneven. Europe features strong competition between ports with almost 70% of the total European container throughput by sea being concentrated in the top 15 ports and more than half of this traffic in the Hamburg-Le Havre region (Notteboom, 2010). A strong hierarchical port system has progressively established in Northern Europe, where the ports of Rotterdam, Hamburg, and Antwerp are constantly improving their handling capacities and competitiveness and concentrate the major maritime transport activity. Most studies on port concentration use traditional economic tools such as the Gini exponent (Kurby and Reid, 1992; Notteboom, 1997; Notteboom, 2006; Li et al., 2022), the shift-share analysis (Notteboom, 2010; Li et al., 2022), or the Hirschman-Herfindahl coefficients (Notteboom, 2006; Wilmsmeier et al., 2014). To study the European ports’ system hierarchies and dynamics, this chapter delivers an original exploration, which complements the analysis by Oliveira et al. (2021) based on rank-size and the Markov chain models. First, we build a rank-size model to characterize port hierarchies and their changes over time. Rank-size models are a precious illustration of ports’ hierarchies at a fixed date. However, they fail in analyzing the rank-size distribution’s dynamics. Second, we use Markov chains’ models to study the relative growth of ports within the rank-size distribution. Markov matrices allow to observe possible changes in the hierarchy of ports and permutations of rank between certain ports. Most studies on concentration in maritime traffic and on ports’ hierarchies only focus on containerized traffic. Our work extends to other types of traffics, such as liquid and dry bulk in addition to containers, to examine whether the concentration trends are similar or not. This chapter delivers two main series of results: first, traffic concentration is observed over time at an aggregate level, but this is not the case for all types of traffic; liquid bulk and containers feature strong hierarchical trends, while dry bulk traffic is characterized by a more diffused distribution. Second, both large and small ports’ hierarchical positions in the rank-size distribution are persistent over time, while medium-size ports often change their position, especially for the solid bulk traffic.

A brief empirical literature review Empirical studies on port hierarchy focus on three main issues1 (Ducruet and Notteboom, 2020): maritime networks, port selection, and traffic

The European ports’ size dynamics and hierarchies  105

concentration. The study of maritime networks is usually based on the coverage of the liner shipping network. This leads to the establishment of a hierarchy of primary and secondary networks connecting different hubs and feeder ports (Ducruet et al., 2010; Ducruet and Notteboom, 2012). Some studies focus on the development of specific ranges of ports and identify the emergence of primary and secondary hubs in a general ports’ hierarchy (Wang and Ng, 2011; Wang and Cullinane, 2014; Ducruet, 2020). These studies provide two main findings: first, they show that the connectivity of a port should be analyzed within a large-scale network designed by shipping companies (Ducruet et al., 2010). Second, they show the first movers’ advantage in establishing hub ports, which allows them to dominate the global trade network. The literature on port selection provides key elements for understanding the levels of concentration and the establishment of steady hierarchies in port systems. When studying the determinants of port efficiency and performance, Tongzon (1995) highlights the importance of port services and their frequency in a port’s growth and position in maritime networks. According to Ugboma et al. (2006), shipping companies and shippers attach great importance to efficiency in terms of the frequency of ship visits and the existence of adequate port infrastructure. This assumption is confirmed by the empirical work of Chang et al. (2008) and Tongzon (2009). The latter show that seaport efficiency, adequate infrastructure, transshipment volumes, and the frequency of maritime services are the main factors in seaport selection by shipping companies. Notteboom et al. (2017) summarize the port selection process, by identifying four distinct groups of selection factors for shipping companies: the demand profile, the supply profile, the market profile, and the carriers’ strategies. Most studies on port competition use data featuring TEU or tonnage throughput to investigate traffic concentration trends. From a methodological point of view, Kurby and Reid (1992) use the Gini index to study the concentration trends in the U.S. general cargo port system that occurred following the technological changes in containerization between 1970 and 1988. Notteboom (1997) applied the Gini index to the European port system and then added to his analysis the Gini decomposition technique to compare between the European and the American system (Notteboom, 2006). He delivers evidence of higher inequalities in the US system compared to the European one. Notteboom (2007) used the “Net Shift Analysis” to explain changes in the port hierarchy in the Hamburg-Le Havre range. While studying the European port system, Notteboom (2010) also considers the port concentration process in the light of structural changes in the logistical, economic, and institutional framework of maritime activities. Following this line, Wilmsmeier et  al. (2014) use a time series analysis of container movements to study the institutional transition of Latin American and Caribbean countries’ ports toward deconcentration. Li et al. (2022) combine several tools to study the trends of a port system, such as the concentration ratios, the Herfindahl-Hirschman

106  Rania Tassadit Dial Currently et al.

Index, and the Dynamic Shift-Share Analysis (DSSA). They apply these tools to the container port system of the Pearl River Delta (Shenzhen, Hong Kong et Guangzhou), with evidence on a certain path dependency. Some recent studies have forwarded original methodological tools to study the dynamics of port hierarchies. By using rank-size models, a Markov chain approach, and transition modeling applied on a dataset of 222 container ports worldwide from 2000 to 2015, Oliveira et al. (2021) show the existence of a middle-ranking growth trap in the container port market. They explain that medium ports basically aim to maintain their existing position and not necessarily to compete with larger ports on the top of the port system hierarchies. Following Oliveira and al.’s work, we apply, in this chapter, rank-size and Markov chains’ models to the European port system for different types of traffic (containers, dry, and liquid bulk).

Methodology and data In this work, first we build a rank-size model for the European port system. By ranging (R) the ports according to their size measured in 1,000 tons (S), the rank-size model admits that the distribution of ports follows a power law such as R (S ) = k ln (S )− β =>  ln R = ln  K +  β ln R (S ) with K a parameter indicating the size of the largest port and β a concentration exponent (Pareto exponent). A weak Pareto exponent ( β < 1) indicates a strong traffic concentration within the larger ports; on the opposite, a strong Pareto exponent ( β > 1) means that the maritime traffic is more equally distributed among the different ports. A distribution with a β parameter equal to 1 is a Zipf distribution, where the largest port represents twice the size of the second largest port, three times the size of the third port, etc. To avoid the statistical bias when using the ordinary least square (OLS) method on small samples (Gabaix and Ibragimov, 2011), the rank-size relation is better estimated with the following equation, where R is the rank of a given port and S is the size in 1,000 tons of a given port: 1 ln  R −  = α + β ln (S )  2 An extensive literature developed over the rank-size model, with applications to different statistical distributions such as the cities’ sizes, the firms’ sizes but also the individual income and GDP per capita (Dimou and Schaffar, 2009; Schaffar and Dimou, 2012). In recent years, some studies have questioned the assumption of the linear distribution of the cities’ sizes especially when one considers the entire distribution and not only the upper truncated part. Rosen and Resnick (1980), Eeckhout (2004), and Anderson and Ge (2005) focused on a possible deviation from strict linearity between the logarithms of the rank and the sizes. They bring evidence that in most cases the relation

The European ports’ size dynamics and hierarchies  107

between the rank and the size best fits a parabolic type of curve (if not a log-normal) that can be studied by adding a quadratic term to the equation (Schaffar and Dimou, 2012): In( R ) = α + β In S + δ InS 2 When δ > 0 , the rank-size distribution is convex, meaning that the average-sized ports are less predominant while when δ < 0, the distribution is concave, meaning that a large proportion of average-sized ports counterbalances the weight of small and large ports. As the goal of this analysis is not simply a static reading of port concentration but a contribution in understanding the dynamics of port size groupings, combining these two methods allows us to observe the role of medium-sized ports in relation to large or small ports, which is not provided by the traditional port concentration measures. Second, in this work, we use Markov chains models on state transitions in order to provide a detailed analysis on the intra-distributional dynamics of the port system (Oliveira et al., 2021). This requires the discretization of the distribution by assigning each port to one of a predetermined number of groups, based on its relative size. In this chapter, the discretization process has been performed by building five homogeneous groups (quintile) of ports with cut-off points exogenously defined 2 (see Table 5.4). Let f be the vector of the distributional shares for each group. By assuming that this distribution follows a homogeneous first order stationary Markov process, f evolves over time according to f t+1 = f   ×  M where M is the transition matrix of ports from time t to t+1. Under this stochastic setting, the maritime traffic at a port follows a Markov chain if, when knowing its volume St in t, we can predict the future throughput (St+1) in t+1, regardless of its previous traffic trends. In this process, the probability pij, t that a port belonging to a size category Si in t, moves to a size category Sj in period t+1 is given by Pr (St +1 ) = j|S0 = i0 , S1 = i1, …, St = it ) = Pr(St +1 = j |St = it ) The transition probabilities are estimated with the maximum likelihood method. The transition matrix measures the speed and extent of the change in the ports’ position in the rank-size distribution. The matrix of the first passage shows the minimum number of years required for a port to reach state j for the first time, when starting from state i. The Markov chain becomes ergodic when pij, t = 0 for any i or j, which means there are no more intra-distributional movements. To compare the mobility of different Markov chains, we use two mobility measures. The first one, suggested by Shorrocks (1978), measures the average

108  Rania Tassadit Dial Currently et al.

probability across all classes that an individual will leave its initial class in the following period: M S (P ) =

k−

∑

k

pi , i

i=1

k −1

where k denotes the number of states of the Markov system and pi , i is the diagonal element of the ith row. The M S ( P ) index is bounded between [0, 1]. A value of zero implies that there is no probability of leaving the initial state while a value of one implies full mobility. The second measure is the Prais (1955) conditional mobility measure, which ref lects the mobility trends conditioned by the chain in state i at the beginning of a transition period. M P ( P ) = 1 − pi , i We apply rank-size and Markov chains models on a dataset, extracted from Eurostat databases (mar_go_qmc), which features information on annual throughput of different types of cargo (total, container, dry, and liquid bulk). In this work, we use a panel dataset of 2243 ports from 13 countries4 covering a 20-year period dating from 2000 to 2019, with 4,480 observations (224 × 20). Descriptive statistics (Table 5.1) show a general increase in maritime traffic over time: the European ports’ traffic met a 21% increase over the entire period which represents an annual growth rate of 1% during the last 19 years. The trend is not linear: maritime traffic in European ports features an early fast growth period that goes from 2000 to 2008, with an average annual growth rate of 2.2%, followed by a short recession due to the subprime crisis (between 2008 and 2010, maritime traffic in Europe has decreased by 4% per year) and finally a slower steady growth period with an annual growth rate of 1.1% between 2010 and 2019. Table 5.1 also shows the changes in specific traffic f lows: while throughputs of liquid and dry bulk commodities are Table 5.1 Gross weight (106) of commodities traffic in European ports Total Period 2000

Liquid bulk 2019

2000

Dry bulk 2019

2000

Container 2019

Total 2,738,464 3,326,457 1,259,604 1,187,058 678,908 627,838 Mean 12,225.29 14,850.25 7,065.8 7,497.6 3,428.8 3,170.9 Std. 25,041.95 36,135.07 14,935.1 18,842.1 7,894.2 6,924,00 Dev. Max 302,484 439,631 145,254 207,366 87,395 70,612 Min 98 96 3 2 11 7 Obs 224 224 168 168 198 198 Source: Authors’ calculations based on Eurostat databases

2000

2019

367,557 823,260 3,869 8,665.9 8,243.9 21,032 51,335 127,901 13 5 95 95

The European ports’ size dynamics and hierarchies  109

dominating the European maritime traffic, their part steadily decreases while containerized traffic experiences a strong annual growth of 4.3%. The dynamic of port traffic is also characterized by significant regional disparities. Figure 5.1 illustrates the relative traffic trends for six different port ranges (Notteboom, 1997). It shows the changes in traffic volume in each port range comparatively to the European average level. A score less than unity indicates that the level of the port range is below the panel average, while a relative connectivity above unity indicates that the level of the port range is above the panel average. The coefficient of the North European port range increased from 2.8 in 2000 to 3.2 in 2019. The value of this coefficient indicates that maritime traffic in the ports of North European is much more important than the average traffic in European ports. West Mediterranean ports stagnate around the average value (from 1.11 to 1.15), while East Mediterranean meet a clear increase only after 2011. This movement can be illustrated by the rise of the port of Piraeus, whose container throughput increased by almost 50% between 2016 and 20195. Finally, while the Baltic and the Atlantic ranges stagnate under the average value, the UK range features a strong decrease of 27% over the whole period. Indeed, large vessels are deployed to call at hub ports in Hamburg-Le Havre Range and then serve the UK ports by feeder (Wilmsmeier and Monios, 2013).

Figure 5.1 Relative transition path by port range (total traffic) Source: Authors’ calculations based on Eurostat databases

110  Rania Tassadit Dial Currently et al.

Results: rank-size and Markov chain models Rank-size analysis of port hierarchy Table 5.2 reports the results of the rank-size equation applied to our European ports sample for 2000, 2005, 2010, and 2019 for each type of cargo. The Pareto exponent β is relatively low for the entire sample (between 0.74 and 0.65) and even more by cargo level (between 0.59 and 0.39). Liquid and container traffic show a higher level of concentration than dry bulk. In fact, large ports often concentrate both liquid bulk and containers (petrochemical complexes like the ports of Rotterdam, Marseille, Le Havre, Antwerp), while dry bulk is more spread out as it generates less economies of scale. There is a decrease in the coefficients over the entire period, regardless of the type of cargo considered. This means that there is a tendency toward traffic concentration and higher hierarchy in the European port system, which goes against the conclusions of previous studies. Next, we apply the quadratic model to explore the relative importance of medium-sized ports to large or small ports. Table 5.3 delivers the results. The coefficient δ is negative and decreasing (in absolute value) over time which means that the distribution is characterized by many medium-size ports but with a decreasing inf luence, while the weight of top ports increases over time. This result is consistent with the previous findings of the rank-size model and confirm the concentration trend in the European port traffic over time. When studying the hierarchical distribution of ports according to the total traffic (Figure 5.2a), there is evidence of important positional changes in the Table 5.2 Rank-size model for container port systems 2000 Ln(S) Constant Observations Ln(S) Constant Observations Ln(S) Constant Observations Ln(S) Constant Observations

Total traffic −0.726*** 10.656*** 224 Liquid bulk −0.439*** 7.380*** 168 Dry bulk −0.592*** 8.524*** 198 Container −0.439*** 6.503*** 95

2005

2010

−0.744*** −0.720*** 10.898*** 10.633*** 224 224

2015

2019

−0.649*** −0.658*** 9.991*** 10.098*** 224 224

−0.441*** −0.401*** −0.393*** −0.393*** 7.428*** 7.075*** 6.967*** 6.989*** 168 168 168 168 −0.573*** −0.541*** −0.503*** −0.513*** 8.421*** 8.085*** 7.799*** 7.899*** 198 198 198 198 −0.472*** 6.925*** 95

−0.487*** 7.080*** 95

−0.450*** −0.424*** 6.851*** 6.674*** 95 95

Standard errors in parentheses and *** p