Global Dynamics: Approaches from Complexity Science

Notes on Contributors

Peter Baudains is a Research Associate at the UCL Department of Security and Crime Science. He obtained his PhD in Mathematics from UCL in 2015 and worked for five years on the EPSRC-funded ENFOLDing project, contributing to a wide range of research projects. His research interests are in the development and application of novel analytical techniques for studying complex social systems, with a particular attention on crime, rioting and terrorism. He has authored research articles appearing in journals such as Criminology, Applied Geography, Policing and the European Journal of Applied Mathematics.

Janina Beiser obtained her PhD in the department of Political Science at University College London. During her PhD, she was part of the security workstream of the ENFOLDing project at the UCL's Centre for Advanced Spatial Analysis for three years. Her research is concerned with the contagion of armed civil conflict as well as with government repression. She is now a Research Fellow in the Department of Government at the University of Essex.

Jyoti Belur is a Senior Research Associate and Senior Teaching Fellow. She served as a senior officer of the Indian Police Service for several years. Her experience and understanding of policing has contributed to her research interests in various aspects of policing, counter-terrorism, crime prevention in the United Kingdom and overseas. She has conducted research on a wide variety of topics including police use of deadly force, police investigations, police misconduct, policing left-wing extremism and crimes against women and has published a book titled Permission to Shoot? Police Use of Deadly Force in Democracies, as well as a number of journal articles and book chapters.

Steven Bishop is a Professor of Mathematics at UCL where he has been since arriving in 1984 as a post-doctoral researcher. He published over 150 academic papers, edited books and has had appearances on television and radio. Historically, his research investigated topics such as chaos theory, reducing vibrations of engineering structures and how sand dunes are formed, but has more recently worked on ‘big data’ and the modelling of social systems. Steven held a prestigious, ‘Dream’ Fellowship funded by the UK Research Council (EPSRC) until December 2013 allowing him to consider creative ways to arrive at scientific narratives. He was influential in the formation of a European network of physical and social scientists in order to investigate how decision-support systems can be developed to assist policy-makers and, to drive this, has organised conferences in the United Kingdom and European Parliaments. He has been involved in several European Commission funded projects and has helped to forge a research agenda which looks at behaviour of systems that cross policy domains and country borders.

Alex Braithwaite is an Associate Professor in the School of Government and Public Policy at the University of Arizona, as well as a Senior Research Associate in the School of Public Policy at University College London. He obtained a PhD in Political Science from the Pennsylvania State University in 2006 and has since held academic positions at Colorado State University, UCL, and the University of Arizona. He was a Co-Investigator on the EPSRC-funded ENFOLDing project between 2010 and 2013, contributing to a wide range of projects under the ‘security’ umbrella. His research interests lie in the causes and geography of violent and nonviolent forms of political conflict and have been published in journals such as Journal of Politics, International Studies Quarterly, British Journal of Political Science, Journal of Peace Research, Criminology and Journal of Quantitative Criminology.

Simone Caschili has a PhD in Land Engineering and Urban Planning, and after being a Research Associate at Centre for Advanced Spatial Analysis (UCL) and Senior Fellow of the UCL QASER Lab, he is currently an Associate at LaSalle Investment Management, London. His research interest covers the modelling of urban and regional systems, property market, spatio-temporal and economic networks and policy evaluation for planning in both transport and environmental governance.

Adam Dennett is a Lecturer in Urban Analytics in the Centre for Advanced Spatial Analysis at University College London. He is a geographer and fellow of the Royal Geographical Society and has worked for a number of years in the broad area of population geography, applying quantitative techniques to the understanding of human populations; much of this involving the use of spatial interaction models to understand the migration flows of people around the United Kingdom, Europe and the world. A former secondary school teacher, Adam arrived at UCL in 2010 after completing a PhD at the University of Leeds.

Robert J. Downes is a MacArthur Fellow in Nuclear Security working at the Centre for Science and Security Studies at the Department of War Studies, King's College London. Trained as a mathematician, Rob received his PhD in mathematics from UCL in 2014; he studied the interplay between geometry and spectral theory with applications to physical systems and gravitation. He also holds an MSc in Mathematics with Theoretical Physics awarded by UCL. As a Postdoctoral Research Associate on the ENFOLDing project at The Bartlett Centre for Advanced Spatial Analysis, Rob studied the structure and dynamics of global socio-economic systems using ideas from complexity science, with particular emphasis on national economic structure and development aid.

Shane Johnson is a Professor in the Department of Security and Crime Science at University College London. He has worked within the fields of criminology and forensic psychology for over 15 years and has particular interests in complex systems, patterns of crime and insurgent activity, event forecasting and design against crime. He has published over 100 articles and book chapters.

Rob Levy is a researcher at the Centre for Advanced Spatial Analysis at University College London. He has a background in quantitative economics, database administration, coding and visualisation. His first love was Visual Basic but now writes Python and Javascript, with some R when there is no way to avoid it.

Elio Marchione is a Consultant for Ab Initio Software Corporation. Elio was Research Associate at the Centre for Advanced Spatial Analysis at University College London (UK). He obtained his PhD at the University of Surrey (UK) at the Centre for Research in Social Simulation; MSc in Applied Mathematics at the University of Essex (UK); MEng at the University of Naples (ITA). His current role consists, among others, in designing and building scalable architectures addressing parallelism, data integration, data repositories and analytics while developing heavily parallel CPU-bound applications in a dynamic, high-volume environment. Elio's academic interests are in designing and/or modelling artificial societies or distributed intelligent systems enabled to produce novelty or emergent behaviour.

Francesca Romana Medda is a Professor in Applied Economics and Finance at University College London (UCL). She is the Director of the UCL QASER (Quantitative and Applied Spatial Economics Research) Laboratory. Her research focuses on project finance, financial engineering and risk evaluation in different infrastructure sectors such as the maritime industry, energy innovation and new technologies, urban investments (smart cities), supply chain provision and optimisation and airport efficiency.

Pablo Mateos is Associate Professor at the Centre for Research and Advanced Studies in Social Anthropology (CIESAS) in Guadalajara, México. He is honorary lecturer in the Department of Geography, University College London (UCL), in the United Kingdom where he was Lecturer in Human Geography from 2008 to 2012. At UCL, he was a member of the Migration Research Unit (MRU) and Research Fellow of the Centre for Research and Analysis of Migration (CReAM). His research focuses on ethnicity, migration and citizenship in the United Kingdom, Spain, the United States and Mexico. He has published over 40 articles and book chapters, and a book monograph titled Names, Ethnicity and Populations: Tracing Identity in Space published by Springer in 2014.

Thomas Oléron Evans is a Research Associate in the Centre for Advanced Spatial Analysis at University College London, where he has been working on the ENFOLDing project since 2011. In 2015, he completed a PhD in Mathematics, on the subject of individual-based modelling and game theory. He attained a Masters degree in Mathematics from Imperial College London in 2007, including one year studying at the École Normale Supérieure in Lyon, France. He is also an ambassador for the educational charity, Teach First, having spent two years teaching mathematics at Bow School in East London, gaining a Postgraduate Certificate in Education from Canterbury Christ Church University in 2010.

Alan Wilson FBA, FAcSS, FRS is Professor of Urban and Regional Systems in the Centre for Advanced Spatial Analysis at University College London. He is Chair of the Home Office Science Advisory Council and of the Lead Expert Group for the GO-Science Foresight Project on the Future of Cities. He was responsible for the introduction of a number of model building techniques which are now in common use – including ‘entropy’ in building spatial interaction models. His current research, supported by ESRC and EPSRC grants, is on the evolution of cities and global dynamics. He was one of two founding directors of GMAP Ltd. in the 1990s – a successful university spin-out company. He was Vice Chancellor of the University of Leeds from 1991 to 2004 when he became Director-General for Higher Education in the then DfES. From 2007 to 2013, he was Chair of the Arts and Humanities Research Council. He is a Fellow of the British Academy, the Academy of Social Sciences and the Royal Society. He was knighted in 2001 for services to higher education. His recent books include Knowledge Power (2010), The Science of Cities and Regions and his five volume (edited) Urban Modelling (both 2013) and (with Joel Dearden) Explorations in Urban and Regional Dynamics (2015).

Belinda Wu holds a PhD in Geography and has a broad interest in modelling and simulating complex systems in socio-spatial dimensions, with a focus on quantitative strategic decision-support systems. Currently, she works as a Research Associate on the development aid workstream of the ENFOLDing project in CASA, UCL. She was appointed as the main researcher on a series of transport planning and policy research projects in Northern Ireland, before working as a main researcher at two nodes of UK National e-Social Science Centre: Genesis (Generative Simulation for the Social and Spatial Sciences) and MoSeS (Modelling and Simulation of e-Social Science). In 2012, she became the named researcher of the ESRC project SYLLS (Synthetic Data Estimation for the UK Longitudinal Studies) to produce the synthetic microdata to broaden the usage of the valuable UK Longitudinal Studies data for ONS (Office of National Statistics). She is also a Fellow of Royal Geographical Society (FRGS) and a Member of Institute of Logistics and Transportation (MILT).

Acknowledgements

I am grateful to the following publishers for permission to use material.

INDECS, Interdisciplinary Description of Complex Systems, Scientific Journal: A Review of the Maritime Container Shipping Industry as a Complex Adaptive System, INDECS, 10(1), 1–15, used in Chapter 2.

CASA, UCL: Shipping as a complex adaptive system: A new approach in understanding international trade, CASA Working Paper 172, used in Chapter 2; Global migration modelling: A review of key policy needs and research centres, CASA Working Paper 184, used in Chapter 5.

Pion Ltd: A multi-level spatial interaction modelling framework for estimating inter-regional migration in Europe, Environment and Planning A 45: 1491–1507, used in Chapter 6.

Springer: Space-time modelling of insurgency and counterinsurgency in Iraq, Journal of Quantitative Criminology, 28(1), 31–48, used in Chapter 12.

I am very grateful to Helen Griffiths and Clare Latham for the enormous amount of work they have put into this project. Helen began the process of assembling material which Clare took over. She has been not only an effective administrator but an excellent proof reader and sub-editor!

I also acknowledge funding from the EPSRC grant: EP/H02185X/1.

Part One

Global Dynamics and the Tools of Complexity Science

Chapter 1

Global Dynamics and the Tools of Complexity Science

Alan Wilson

The populations and economies of the 220 countries of the world make up a complex global system. The elements of this system are continually interacting through, for example, trade, migration, the deployment of military forces (mostly in the name of defence and security) and development aid. It is a major challenge of social science to seek to understand this global system and to show how this understanding can be used in policy development. In this book, we deploy the tools of complexity science – and in particular, mathematical and computer modelling – to explore various aspects of change and the associated policy and planning uses: in short, global dynamics.

What is needed and what is the available toolkit? Population and economic models are usually based on accounts. Methods of demographic modelling are relatively well known and can be assumed to exist for most countries. In this case, we will largely take existing figures and record them in an information system. An exception is the task of migration modelling. National economic models are, or should be, input–output based. We face a challenge here, in part, to ensure full international coverage and also to link import and export flows with trade flows. In the case of security, there are some rich sources of data to report; in the case for development aid, the data are less good. In each case, we require models of the flows – technically, models of spatial interaction.

Flow models represent equilibria or steady states. Our ultimate focus is dynamics. There will be imbalances in the demographic and economic accounts, and these become the drivers of change in dynamic models. Typical combinations of systems and models that we explore are as follows:

multi-layered spatial interaction models of trade flows – in the context of rapidly changing ship, port and route ‘technologies’;

dynamic models of trade and economic impact using a variant of spatial Lotka–Volterra;

input–output models linked by spatial interaction models of imports and exports;

spatial interaction models of migration combined with biproportional fitting;

models of riots (i) using epidemiological and spatial interaction modelling, (ii) using discrete choice models, (iii) using spatial statistics and (iv) using diffusion models;

models of piracy (i) using agent-based models and (ii) using spatial interaction models with ‘threat’ as the interaction;

models of ethnic contagion using spatial statistics;

modelling the impact of development aid through input–output models;

spatial Richardson (arms race) models;

Colonel Blotto game-theoretic security models.

We introduce each of these in a little more detail, noting the actual or potential planning and policy applications of each.

In the case of shipping (Chapters 2 and 3), we can use the models we develop to explore the consequences of changing patterns of trade and changing transport technologies. There are rich, albeit disparate, sources of data. The global trade system is complex – through the variety of goods, commodities and services that are carried and through the set of transport modes deployed – sea, air, rail, road and telecommunications. This means that we have to choose levels of resolution at which to work and particular systems of interest on which to focus. In making these decisions, we are, to an obvious extent, constrained by the availability of data. We also wish to connect – and make consistent – any predictions from a model of trade with the import and export data which form part of the input–output tables to be outlined in the next chapter. We focus on a coarse level of aggregation based on seven economic sectors, and we present these sectors and volumes of trade in money terms. We focus mainly on ‘container shipping’, though container routes usually include road and/or rail elements as well as sea. This covers 80% or more by volume of trade flows. We proceed in two stages beginning with a review of the evolution of the container shipping system (Chapter 2) and then by building a multi-layered model of international trade (Chapter 3).

A key component of an integrated global model will be a submodel that gives us the state of economic development of each country. The ideal model for each country is an input–output model and these, of course, would be linked through trade flows. It seems appropriate, therefore, to report our response to this challenge in this section along with trade (Chapter 4). The basis of this development has to be the existence of national input–output tables. WIOD 2012 provides an excellent source for 40 countries. However, there are enormous gaps of course – 40 out of 220 – and these gaps embrace the whole of Africa. We have sought to handle this situation by developing new tools, based on high-dimensional principal components' analysis, which enable us to estimate the missing data. The detail of this method is presented in Chapter 5 of our companion book Geo-Mathematical Modelling (Wilson, ed., 2016).

The policy challenges facing governments associated with migration are essentially of three kinds: the effective integration of in-migrants; limiting the inflows of some types of migrant; encouraging inflows of others. There are forces driving migration which, from governmental perspectives, are more or less controllable in different circumstances. It is important, as ever, to seek to provide a good analytical base to underpin the development of policy. There has been extensive research on migration, and we first provide a background to our own work by surveying this research in the context of the policy questions that arise (Chapter 5).

A typical problem facing the global systems' modeller is the situation in which the data available are not sufficiently detailed. In this case, bearing in mind the nature of the policy challenges, in Chapter 6 we take on the task of estimating flows at a regional (sub-national) level. We do not have the data to achieve this on a global scale, but we have good European data and so we develop the methodology on this basis. This is a classical biproportional fitting problem. Migration data have to be assembled from a variety of sources and different ones are more or less reliable. In order to build as complete a picture as possible of global migration, we explore a variety of sources and seek to integrate them (Chapter 7).

Security challenges vary in scale from the urban – even street level – to the international, for example, through the global deployment of a country's military forces. These different scales, in general, demand different modelling methods and we seek to illustrate a range of these. Security has rich but disparate data. We have developed a two-pronged approach: first to develop some new theoretical models by taking some traditional ones and adding spatial structures, and second, we have assembled a wide range of data that has allowed us to carry out some preliminary tests. We recognise that in this case, there will be government agencies around the world who are modelling these systems with far richer resources than we can bring to bear. What we hope to have achieved is to demonstrate some new approaches to security modelling that may be taken up by these agencies. In this case as well, we have been able to develop models at finer scales in relation to riots, rebellions and piracy. A key concept in this work is the representation of ‘threat’ and in particular, threat across space. We introduce this in broad terms in Chapter 8.

We then present five distinct applications which between them offer a wide range of methods. In some cases, we can apply different methods to the same problem and so discover the strengths and weaknesses in a comparative framework. Chapter 9 offers a variety of approaches to the London riots of August 2011. We built a three-stage model – propensity to riot (from epidemiology), where to riot (a version of the retail model) and the probability of arrest. We use Monte Carlo simulations to determine whether the counts of observed patterns are more or less frequent than might be expected under conditions in which the extent of spatio-temporal dependency of offences is varied.

In Chapter 10, we shift scale and location again and examine the Naxalite rebellions in India. The data on Naxalite terrorism include the date on which events took place and the district (of which there are 25) within which they occurred. Events include Naxalite attacks and police responses. A key idea in the insurgency literature concerns the contagion of events. This can occur for a number of reasons. For example, conflict may literally spillover from one locality to a nearby other, leading to an increase in the area over which the conflict occurs or moving from one location to those adjacent. In this case, we explore a number of hypotheses. We can test whether there are non-spatial effects of police action on insurgent activity. Moreover, we may test the hypothesis that police action is triggered by insurgent activity. If only the latter is observed, this would suggest that police action is reactive but has little effect on insurgent actions (at least on a short-time scale). For such models, the count of attacks per unit time is described by two components: (i) the first is a baseline risk – which may be time invariant or not, but where it changes it will tend to do so over a relatively long-time scale; and (ii) a self-excitation process, whereby recent events have the potential to increase the likelihood of attacks today considerably.

In Chapter 11, we explore a very different system of interest: piracy in the Gulf. An important question is the security of shipping in relation to pirate attacks. There are two possible approaches to this problem: first, to develop an agent-based model with a given (and realistic) pattern of shipping, and pirates as agents; and secondly, as adopted in this Chapter, to develop a spatial interaction model of ‘threat’ and to use this to explore naval strategies.

In Chapter 12, we explore a different kind of security issue with a different method: the impact of IEDs (improvised explosive devices) in Iraq. The null hypothesis is that they are independent in time and space. We use Knox's method of contagion analysis to seek evidence of clustering – an important issue in the assessment of response to this kind of threat – and find that there is evidence for clustering in space, time and space-time.

Another kind of security issue is posed in nearby countries where there is a threat of cross-border contagion fuelled, for example, by social networks and this is the subject of Chapter 13. We consider whether ethnic conflict is contagious between groups in different countries and if so, how? And then, whether governments react pre-emptively to potential conflict contagion by increasing repression of specific groups? The argument to be tested is that ethnic groups that are discriminated against in a society identify with groups fighting against the same grievance in other countries and become inspired by their struggle to take up arms against their own domestic government as well. For this process, information about foreign struggles is important, not geographic proximity as such. The empirical test involves using a statistical model on country-years from 1951/1981 to 2004 and this gives some support for the argument. The test will be repeated using data on the analytic level of ethnic groups in different years and improving on the measures of information flows. In the case of government reaction, the argument to be tested is whether governments pre-emptively increase repression against ethnic groups they expect to become inspired by foreign conflicts in order to deter them from mobilising. The empirical test in this case is through a strategic model using data on the behaviour of governments towards domestic ethnic groups.

Development aid (Chapters 14 and 15) offers different challenges: first, defining categories; and then assembling data from very diverse sources. In this case, the ultimate challenge is to seek to measure the effectiveness of aid, and a starting point is to connect aid to economic development. This creates a demand to ‘measure’ development, and we have done this by constructing input–output models for each country which can then be integrated with our trade model. It then becomes possible to compare the magnitudes of different kinds of aid flows with other trade flows and with flows within national economies. Not surprisingly, aid is much more significant in developing countries than in those with advanced economies. The value of the global input–output model now becomes apparent: in a selected country, we can compute the multiplier effects of increased demand or investment in particular sectors and then begin to address the question of whether investment aid is most effectively targeted. We model aid allocation in Chapter 16.

We finally seek to move beyond our investigation of the impact of aid on development in particular countries and explore the extent to which it has any impact on trade, on migration flows or on helping to maintain security. It has been necessary to drive our work in developing particular submodels by assembling relevant data in each case. Our global input–output and trade system then provides the basis for integrating the main submodels so that we explore the interdependencies which make the global system so complex. It is foolish to think that we (or anyone else) can offer a detailed and convincing ‘model of the world’ in all its aspects. But what we can do is to offer a demonstration model that reveals some of the complex system consequences of interdependence and points the way to further research, possibly to be carried out by government and inter-governmental agencies that can bring far greater resources to bear. In Chapter 17, therefore, we draw together the different submodels into a comprehensive model which enables us to incorporate the key interactions. Some of the most obvious interdependencies to be picked up are as follows:

the impacts of net migration on economic development through the labour elements of the national input–output models;

security-led pushes in outward migration;

security-led changes in economic development – whether from damage from attack or because of more intensive development of the arms industry;

many aspects of changing trade patterns – for example, from investment in new ports as well as changes in economic development levels;

the re-targeting of development aid.

We proceed by establishing a base model and year – taking 2009 as the latest year for which input–output data are available at the time of writing. As noted in Chapter 4, the model is rooted in WIOD 2012 data but then enhanced through a principal components' technique to cover all countries. The import and export flows are integrated with those from a trade model by a biproportional fitting procedure. We assemble base year data and models (as appropriate) for the flows of migrants, military dispositions and development aid. These become drivers of change for subsequent time periods (which we take to be years). At each year end, a number of indicators are calculated and particular attention is paid to imbalance as these will provide the basis for driving the system dynamics. Each year end ‘model run’, for this reason, is likely to involve iterations driving the system to a new equilibrium.

Reference

WIOD (2012) The World Input–Output Data Base: Content, Sources and Methods, Technical report Number 10.

Wilson, A. (ed) (2016) Geo-Mathematical Modelling, Wiley, Chichester.

Part Two

Trade and Economic Development

Chapter 2

The Global Trade System and Its Evolution

Simone Caschili and Francesca Medda

2.1 The Evolution of the Shipping and Ports' System

Shipping volumes have grown dramatically in recent decades, and this growth has been coupled with changing technologies – particularly through larger ships and improved port logistics – and with changing geographies. At present, many authors estimate that maritime shipping ranges between 77% and 90% of the intercontinental transport demand for freight by volume compared to shipping in the 1980s when it was around 23% (Rodrigue et al., 2006, 2009; Glen and Marlow, 2009; Barthelemi, 2011). The total number of twenty-foot equivalent units (TEU) carried worldwide ranged from 1,856,927 in 1991 to 7,847,593 in 2006 (Notteboom, 2004), and the average vessel capacity has grown from 1900 TEU in 1996 to 2400 TEU in 2006 (Ducruet and Notteboom, 2012). Kaluza et al. 2010 attribute this substantial increase to the growth in trans-Pacific trade. The lower cost per TEU-mile in long distance transport for large quantities of goods has also driven this growth (Rodrigue et al., 2006) coupled with significant technical improvements in size, speed and ship design as well as automation in port operations (Notteboom, 2004; Rodrigue et al., 2009). For instance, in 1991, the use of vessels larger than 5000 TEU was unheard of, but by 1996 large vessels constituted about 1% of the world's fleet, increasing to 12.7% in 2001 and 30% in 2006.

A variety of independent agents (shipping companies, commodity producers, ports and port authorities, terminal operators and freight brokers) play roles in this process, and through their mutual interactions, they determine the patterns of development and growth. It is helpful to view shipping as a complex system of relatively independent parts that constantly search, learn and adapt to their environment while their mutual interactions shape hidden patterns with recognisable regularities that continuously evolve. In this context, the science of Complex Adaptive System (CAS) provides a useful framework to analyse a shipping system (Arthur, 1997; Dooley, 1996; Gell-Mann, 1994; Gell-Mann, 1995; Holland, 1992; Holland, 1995; Holland, 1998; Levin, 1998; Levin, 2003), a field of study in which strategic analysis is based on what is essentially a bottom-up perspective. CASs are generally assumed to be composed of a set of rational, self-learning, independent and interacting agents whose mutual interrelations generate non-linear dynamics and emergent phenomena.

Since the 1980s, maritime agents have continuously evolved in their organisation in response to both external stimulus and market competition. In the logistics and management perspective, a new form of inter-firm organisation has emerged in the shipping industry. Rodrigue et al. 2009 succinctly explain how this change has occurred:

[…] many of the largest shipping lines have come together by forming strategic alliances with erstwhile competitors. They offer joint services by pooling vessels on the main commercial routes. In this way they are each able to commit fewer ships to a particular service route, and deploy the extra ships on other routes that are maintained outside the alliance. […] The 20 largest carriers controlled 26% of the world slot capacity in 1980, 42% in 1992 and about 58% in 2003. Those carriers have the responsibility to establish and maintain profitable routes in a competitive environment Rodrigue et al. 2009.

The evolution of the shipping industry has gone hand in hand with the changes in port organisation. According to a study for the European Parliament 2009, from the growth of containerisation to what is known as the terminalisation era, where ports have become multifunctional operations through the development of highly specialised terminals, ports have undergone a major transformation in their management structures. The role of port authorities has also changed as the shipping system has evolved. Their main duties now involve the optimisation of process and infrastructures, logistics performance, promotion of intermodal transport systems and increased relations with their hinterlands.

In light of these observations, our objective in this chapter is to examine how maritime shipping can be modelled through the use of CAS theory. If we assume that emergent global behaviour such as international trade can be explained through bottom-up phenomena arising from the local interaction among individual agents, our main goal is to understand how patterns emerge in the global shipping system. The argument of our analysis is organised as follows. In Section 2.2, we summarise the main features regarding the worldwide movements of goods. Section 2.3 provides a detailed discussion of the CAS methodology for maritime trade, and in Section 2.4 we discuss the opportunity to apply CAS modelling to the maritime system and conclude with a research agenda for future studies.

2.2 Analyses of the Cargo Ship Network

Two recent articles (Ducruet and Notteboom, 2012; Kaluza et al., 2010) examine the main characteristics of the Global Cargo Ship Network (GCSN). Other studies focus on sub-networks of the GCSN. Ducruet et al. 2010 have analysed the Asian trade shipping network, McCalla et al. 2005 the Caribbean container basin, Cisic et al. 2007 the Mediterranean liner transport system and Helmick 1994 analysed the North Atlantic liner port network. But the two studies of Kaluza et al. 2010 and Ducruet and Notteboom 2012 are particularly interesting because they examine the complete global network and give us a view of its macroscopic properties. However, one drawback is their inability to forecast future trends or changes in the networks. Nevertheless, for our purpose, the aim of both studies is to characterise the global movements of cargo in order to define quantitative analyses on existing structural relations in the rapidly expanding global shipping trade network.

Table 2.1 highlights similarities and differences between the two studies on the GCSN. Kaluza et al. use the Lloyd's Register Fairplay for the year 2007, while Ducruet and Notteboom utilise the data set from Lloyd's Marine Intelligence Unit for the years 1996 (post-Panamax vessels period) and 2006 (introduction of 10,000+ TEU vessels).

Table 2.1 Overview of the main features of the GCSN as proposed in the studies of Kaluza et al. 2010 and Ducruet and Notteboom 2012

Both studies apply different approaches to the network analysis and sometimes reach different conclusions. Ducruet and Notteboom built two different network structures: the first (Graph of Direct Links – GDL) takes only into account the direct links generated by ships mooring at subsequent ports, and the second (Graph of All Linkages – GAL) includes the direct links between ports which are called by at least one ship. Kaluza et al. 2010 differentiate between the movements according to the type of ship. Four networks are subsequently constructed: all available links, sub-network of container ship, bulk dry carriers and oil tankers. Despite the clear differences between the approaches adopted in the two studies, in order to compare them we have considered the complete network of ships' movements by Kaluza et al. and the GAL network of Ducruet and Notteboom.

All the networks are quite dense (on average, the ratio between number of edges and nodes is 37.2). Some network measures indicate a tendency of the GCSN to belong to the class of small world networks¹ given the high values of the Clustering Coefficient.² Small world networks are a special class of networks characterised by high connectivity between nodes (or, in other words, low remoteness among the nodes). In an economic setting, this property has an underestimated value, for example, in the retail market, and connections among firms can create clusters of small specialised firms that gravitate around a big firm (hub). This large firm uses small sub-peripheral firms to sub-contract production. Thus, both firms (the hub and peripheral ones) reach their goals and increase the economic entropy of the system (Foster, 2005).

The degree distribution P(k)³ shows that "most ports have few connections, but there are some ports linked to hundreds of other ports" (Kaluza et al., 2010). However, when the authors examine the degree distribution in detail, they find that the GCSN does not belong to the class of scale-free networks.⁴ Both studies show low power law exponents or right-skewed degree distributions. In this sense, the degree distribution analysis would have had a higher significance if the authors had showed a ranking of the ports over time. This would have allowed them to understand if there was ever a turnover of dominant hubs that in turn had led to the detection of competitive markets in maritime shipping. Opposite results would have depicted a constrained market.

Kaluza et al. 2010 also studied the GCSN as a weighted network where the distribution of weights and Strength⁵ displays a power law regime with exponents higher than 1. These findings agree with the idea that there are a few routes with high-intense traffic and a few ports that handle large cargo traffic. The detection of power law regimes is often associated with inequality (i.e. distribution of income and wealth) or vulnerability in economic systems (Foster, 2005; Pareto, 1897). The correlation between Strength and Degree of each node also fits a power law; this implies that the amount of goods handled by each port grows faster than the number of connections with other ports. Hub ports do not have a high number of connections with other ports, but the connected routes are used proportionally by a higher number of vessels.

Unfortunately, Ducruet and Notteboom's work (Ducruet and Notteboom, 2012) does not provide results of the weighted network analysis over the years 1996 and 2006. Such as analysis would have allowed us to discuss relevant facts about the dynamics of flows in the main transatlantic routes and also address constructive criticism of the influence of the introduction of large loading vessels (post-Panamax era) on specific routes.

The centrality of ports in a network (i.e. the importance of a node) can be inspected with other topological measures rather than the crude number of connections per node (degree k). In the case of GCSN, both studies use measures of the betweenness centrality.⁶ Kaluza et al. 2010 highlight a high correlation between the degree k and betweenness centrality and then validate the theory that hub ports are also central points of the network. Ducruet and Notteboom detect interesting anomalies in the centrality of certain ports. Large North American and Japanese ports are not on the top rank position in terms of network centrality despite their traffic volume. The most central ports in the network are the Suez and Panama canals (as they are gateway passages), Shanghai (due to the large number of ships that visit the port) and ports such as Antwerp (due to its high number of connections.)

Although maritime shipping has been going through a tremendous period of expansion in the last decade, the underlying network has a robust topological structure that has not changed in recent years. Kaluza et al. 2010 observe the differences in the movement patterns of different ship types. The container ships show regular movements between ports, which may be explained by the nature of the service they provide. Dry carriers and oil tankers tend to move in a less regular manner because they change their routes according to the demand of goods they carry.

Finally, maritime shipping appears to have gained a stronger regional dimension over the years. In 1996, there was a strong relation between European and Asian basins while in 2006