Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The production of the world city : extractive industries in a global urban economy Surborg, Björn 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2012_spring_surborg_bjoern.pdf [ 759.13kB ]
JSON: 24-1.0072585.json
JSON-LD: 24-1.0072585-ld.json
RDF/XML (Pretty): 24-1.0072585-rdf.xml
RDF/JSON: 24-1.0072585-rdf.json
Turtle: 24-1.0072585-turtle.txt
N-Triples: 24-1.0072585-rdf-ntriples.txt
Original Record: 24-1.0072585-source.json
Full Text

Full Text

THE PRODUCTION OF THE WORLD CITY: EXTRACTIVE INDUSTRIES IN A GLOBAL URBAN ECONOMY by BJÖRN SURBORG A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Geography) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) February 2012 © Björn Surborg, 2012 Abstract This dissertation argues for a re-grounding of world city research in world- systems and dependency theory. It proposes to conduct 'vertical world city research', which explicitly investigates the spatial interconnectedness between world cities and peripheral locations of production, rather than focus on the relationships between different world cities. The idea of vertical world city research is partly a response to recent post-colonial critiques of world city research. Advanced Producer Services (APS) have long been considered command and control functions over global capital. Constructions of world city networks have largely relied on data based on large APS firms. While this research has made important contributions to our understanding of the interconnectedness of world cities, the argument of this dissertation is that it may not adequately capture the role of other essential sectors in the global economy. It is proposed that world city research needs to investigate the global control networks over monopolistically (or more correctly oligopolistically) organised processes. The focus of this dissertation is therefore the urban control network of the monopoly over natural resources. The dissertation investigates the locations of ownership control over the world's ten largest non-fuel mineral producers. These ten firms account for more than one third of the world's non-fuel mineral production by value. It also investigates specifically the platinum industry, which is controlled by a very small number of firms. While cities that are often identified as world cities, including New York and London, feature in the lists of cities that host the owners of the global mining industry and the platinum industry in particular, a number of cities in middle income economies are at the top of the list. The last part of the dissertation focuses on the role of nation states in the formation of global cities and how corporate decisions are administered through a network of cities. The effect of these decisions on mining communities is explicitly studied. This part focuses on Johannesburg and South Africa. The research suggests that the spatial organisation of the world economy and the concentration of power is more dispersed than previously suggested. ii Preface Parts of Chapter 2 and 6 of this dissertation have been published in the journal Urban Forum. Surborg, B. (2011). World cities are just "basing points for capital": Interacting with the world city from the global south. Urban Forum, 22(4), 315-330. Research for this dissertation was approved by the UBC Behavioural Research Ethics Board (H08-03112) and the Human Science Research Council of South Africa (HSRC Ethics Committee Protocol REC 1/04/03/09).  iii Table of Contents Abstract............................................................................................................................ii Preface............................................................................................................................iii Table of Contents............................................................................................................iv List of List of Abbreviations.....................................................................................................vii Acknowledgements......................................................................................................viii Dedication.......................................................................................................................ix Introduction......................................................................................................................1 Part I ...................................................................................................................................... Chapter 1: New Socio-Spatial Constellations, Old Division of Labour .......................12 Development and dependency...................................................................................12 The world-systems perspective..................................................................................14 Polarisation at a world scale......................................................................................18 New socio spatial constellations and the global division of labour...........................21 Towards a new spatiality of world city research.......................................................27 Chapter 2: World City Research: A Review .................................................................29 World city research: remembering the agenda..........................................................29 The network view of world cities..............................................................................34 The producer services and other network articulators...............................................40 Cities and globalisation: multiple perspectives on cities in the world.......................43 Complex relations: where to start world city research?.............................................49 Toward vertical world city research: a renewed world-systems approach................53 Methodology..............................................................................................................58 Part II..................................................................................................................................... Chapter 3: Cities as Switchboards of Processes............................................................65 Trading the control over natural resources................................................................67 Control through ownership........................................................................................72 Power in the world city network................................................................................75 Types of ownership....................................................................................................77 Commanding or servicing capital?............................................................................81 World cities in the world of mining...........................................................................83 Satellite functions and national territory....................................................................86 Historical dimensions and shifts in power.................................................................87 Diversified mining, concentrated ownership and the old north-south divide ...........90 Chapter 4: Control over Platinum: Cities, Firms and Governments..............................92 Bottom-up world city research...................................................................................93 The locus of control and command..........................................................................101 Trade and consumption of platinum........................................................................109 Interdependencies: operational and strategic control...............................................110 Identifying interdependencies through vertical world city research........................114 Part III.................................................................................................................................... Chapter 5: The Wealth of World Cities: Nation States Make Cities...........................118 Transition in South Africa.......................................................................................119 iv Sowing the seeds of the government-business nexus..............................................121 First grow big, then get out: the case of Anglo American.......................................124 The rise and exit of Afrikaner capital from South Africa........................................127 Fattening the world city from afar...........................................................................129 The state-economy nexus and the rise of the world city network............................133 Chapter 6: Vertical Control in the World City Network..............................................135 The world city as a neo-liberal city..........................................................................138 Black Economic Empowerment and the new mining charter..................................142 Black but not broad-based.......................................................................................147 Labour relations and Johannesburg's "long arm".....................................................150 Mining communities and world cities......................................................................152 The state, cities and vertical control in a network of cities and communities.........156 Chapter 7: Conclusion..................................................................................................159 Bibliography................................................................................................................170 Appendix 1 ..................................................................................................................190 Estimation of shares traded at specific exchanges...................................................190 Data generation by company...................................................................................191 Appendix 2...................................................................................................................195 Data generation by company...................................................................................195 v List of Tables Table 3.1: The world's largest mining firms in 2006.........................................................68 Table 3.2: Stock exchanges of the ten largest non-fuel mineral producers.......................71 Table 3.3: World cities of mining......................................................................................74 Table 3.4: Ownership types of top-ten mining firms ........................................................78 Table 4.1: Global Platinum Production in 2007 by country .............................................94 Table 4.2: Global platinum production in 2007 by individual source...............................95 Table 4.3: Global platinum production by company.........................................................97 Table 4.4: Urban control over platinum production for 2007..........................................100 Table 6.1: Platinum production attributed to BEE companies, 2007..............................148 vi List of Abbreviations ADR = American Depository Receipt ADS = American Depository Shares AIG = American International Group ANC = African National Congress APS = Advanced Producer Services ASX = Australian Stock Exchange BEE = Black Economic Empowerment BMV = Bolsa Mexicana de Valores (Mexican Stock Exchange) BHP = Broken Hill Properties BRICS = Brazil, Russia, India, China and South Africa COSATU = Congress of South African Trade Unions ECLA = United Nations Economic Commission for Latin America GaWC = Globalization and World City Research Group GDP = Gross Domestic Product GEAR = Growth, Employment and Redistribution Programme HDSA = Historically Disadvantaged South African(s) IMF = International Monetary Fund JSE = Johannesburg Stock Exchange LMN = London Mining Network LPPM = London Platinum and Palladium Market LSE = London Stock Exchange MK = Umkhonto we Sizwe NGO = Non-governmental organisation NP = National Party NYSE = New York Stock Exchange NUM = National Union of Mineworkers OTC = Over the Counter PGM = Platinum Group Metals R = South African Rand RDP = Reconstruction and Development Programme SACP = South African Communist Party SAHRC = South African Human Rights Commission SARB = South African Reserve Bank SEC = Securities and Exchange Commission SPV = Special Purpose Vehicle STRATE = Share Transactions Totally Electronic TPSN = Territory, Place, Scale and Network UK = United Kingdom USGS = United States Geological Survey USA = United States of America US$ = United States Dollars WTO = World Trade Organisation vii Acknowledgements I wish to gratefully acknowledge the exceptional advice received from my research advisers Jim Glassman and Elvin Wyly, who provided direction, support and valuable constructive comments throughout the research and writing process. I would also like to thank Tom Hutton and Philippe LeBillon for their support and advice throughout the process. There are many other people who contributed to the progress of this dissertation and who helped me to sharpen my argument. My dissertation has certainly benefited from the inputs and critical thinking of many. In South Africa these people include Chris Rogerson, Patrick Bond, Sophie Oldfield, Susan Parnell, Mfaniseni (Fana) Sihlongonyane, Bridgit Kenny, Keith Beavon, Margot Rubin, Aly Karam and Dan Hammett. In other parts of the world there are Anver Saloojee, James Sidaway and Simon Parker. I also wish to thank my post-doctoral supervisor Padraig Carmody at Trinity College Dublin, who provided helpful comments, as well as the university examiners Leila Harris and Jamie Peck, who engaged constructively with the argument presented here. Thanks as well to the external examiner, Peter Taylor, who provided equally constructive comments. I would also like to thank my fellow graduate students Jamie Doucette, Pablo Mendez, Kathy Sherrell, Markus Moos, Nick Lynch and many others. The staff and faculty in the Department of Geography at UBC need to be gratefully acknowledged as well and I am thankful to the School of Geography, Archaeology and Environmental Studies at the University of the Witwatersrand in Johannesburg for hosting me as a visiting researcher. Moreover, I would like to thank everyone who has participated in this research, most importantly the interviewees who generously shared their time with me. And, most importantly, there are Astrid, Kaja and Kioni. Thanks for bearing with me. viii Dedication To my parents ix Introduction "It is the argument of this book that all cities are best understood as 'ordinary'." Jenny Robinson, 2006, p. 1 of Ordinary Cities "I don't think, there is such a thing as ordinary cities, all cities are extraordinary." Peter  Taylor,  at  the  2008  Annual  Meeting  of  the  Association  of  American Geographers Contemporary world cities form a network that facilitates the most influential command and control functions of today's world economy. And while this network is highly influential and depends on the strength of each node in the network to maintain this influence, world cities do not function autonomously from their surrounding hinterland. New York depends on London and London depends on New York, or more specifically certain actors in these cities depend on each other. It would be difficult to argue otherwise. The two cities are intrinsically linked to each other through a multitude of processes and mutual dependence on each other's expertise. Similarly, London depends on Hong Kong and Tokyo and in turn Tokyo and Hong Kong depend on London. The network of world cities spans the entire globe. However, the relationships between world cities and within this seemingly exclusive network are by no means the only relationships of importance for world cities or any other place on earth. World cities remain closely linked to their surrounding territories, sometimes articulated and negotiated through the nation state, sometimes through supra-national organisations, sometimes through direct corporate influence and sometimes through the military. It is this relationship of the contemporary world city to the peripheral lands of the global economy, that this dissertation is about. Since its inception world city research has been a response to changes in socio- spatial configurations across the globe, partly shaped by a declining influence of the nation state, partly by a growing presence of supra-national organisations, but often also by an interplay between states and cities, which for all intents and purposes influence each other. The focus of this research has been on the role of large, central and well connected cities in organising the world economy and to a much lesser extent global political, social and cultural processes. The research has evolved from an almost entirely 1 place based analysis of cities within their respective nation states (Hall, 1977 [1966]), to an approach of locating cities within a network of cities (Friedmann and Wolff, 1982; Friedmann, 1986; Sassen, 2001 [1991]) and a conception of world cities as almost entirely determined by their position within a world city network (e.g. Taylor, 2004a). The central idea has always been that world cities are the most important locations of decision making processes. This research is of course fascinating: trying to identify the most important locations of centralised decision making within a world of decentralised production has attracted attention from a wide cross-section of society far beyond the academy. The term world city became shorthand for a place of greatness, a desirable place and a place that others should emulate. Academics had not conceived of the world city as such a place and John Friedmann, author of The World City Hypothesis (Friedmann, 1986), experienced firsthand how his idea was interpreted by city governments. In Singapore he had the following experience: A few years ago, I was invited by the government of Singapore to speak on world cities. In private conversations with senior government officials it became clear to me what the government really wanted. Singapore was embarking on 'the next lap' (Government of Singapore, 1991), and officials hoped to hear from me how their city state might rise to the rank of a 'world city'. The golden phrase had become  a  badge  of  status,  just  as  'growth  poles'  had  been  in  an  earlier incarnation.  There  was little  I  could  say that  the government  did  not  already know. (Friedmann, 1995, p 36) Later the term world city was supplemented by global city, a term that its academic author Saskia Sassen (2001) chose because it characterises the articulation of global flows as supposed to being merely a dominant city in the world. Both terms remain widely in use and I apply them interchangeably in this dissertation. Sassen (2001, p. 90) identified the Advanced Producer Services (APS), including insurance, banking, financial services, real estate, legal services, accounting and professional associations, as control and command functions in the global economy, because as production (of both manufacturing and services) has dispersed considerably to multiple sites across the globe, firms need specialised input from producer services to "manage and control global networks of factories, service outlets, and branch offices" (Sassen, 2001, p. 11). These APS are concentrated in the world cities. 2 But are the APS really the strongest indicator for world city network formation? What about other economic factors or social, political and cultural ones? Why are private business jets rushing back and forth between Turin (Italy) and Detroit (USA)? The answer for this connection is an easy one: There are close links in the automobile industry and business executives fly directly between these cities that are both places of production and head office locations, i.e. places of decision making and strategic control. So, there are clearly important direct links between cities that do not regularly feature in contemporary world city research that builds its analyses almost entirely on the producer services. One of the contributions that this dissertation seeks to make is taking a closer look at how different processes contribute to forging, establishing and maintaining connections between places and forming networks, while consciously considering their internal characteristics, their surrounding territory and perhaps most importantly their relative position to other places. Accepting that world cities are the command and control centres of the global economy or more generally the most important places of decision making, I have decided to investigate the world city network through the lens of one of the most elementary inputs into the global economy, the extractive industry. The world's construction or automobile industries could not be maintained without the extraction of iron ore. Mobile phones would not be as ubiquitous across the globe without the input of coltan. More and more high-tech and medical applications require specialised minerals, such as rare earths. Moreover, there is the domain of pure luxury, which to a considerable proportion fuels the extraction of diamonds, gold, silver and platinum, and there is of course speculation, which has been a driver for the extraction of gold for a long time. The APS are often considered the capital commanding functions of the global economy, but the direct links between the APS and their command over capital are not always made explicit in world city research. By researching the extractive industry I am investigating this relationship between capital and capital commanding functions and trace these connections from place to place to establish what I call vertical world city research. These are not the horizontal connections between world cities, which are all hosting APS firms, fulfil similar functions and are likely of very similar importance in the network of cities. These are connections from the places of strategic decision making to the places of implementation and production. But where are these strategic decisions 3 made? There are for sure the board rooms of the world's largest mining companies and there are the general assemblies of these firms, but there are also the offices of pension and investment fund managers, the living rooms of individual shareholders or government offices. All of these places contribute to the decision making process, which is for the most part driven by the motive of profit. At the other end of this vertical world city research are the places of implementation, in this case the places of resource extraction. There are the mines themselves, where people work to extract minerals and then there are the communities around them, where people live and depend on the jobs in the mines and where people may now be drinking contaminated water that was once clean. There is no question that these places are linked and there is a network of cities that facilitate these links. I am therefore integrating aspects of the global commodity chain literature (e.g. Derudder and Witlox, 2010) and the concept of transnational urbanism (Smith, 2001) into world city research. I began this research about two years before the 2008 financial crisis, when Wall Street was still high on what seemed like a never ending supply of cheap money. The APS, especially the banking and insurance sectors, were high growth industries with healthy profit margins and other sectors of the economy trailed along this growth. Then something triggered a crisis. Suddenly, repackaged securities, sub-prime mortgages and credit default swaps were no longer considered such safe investments. Lehman Brothers collapsed. Banks stopped lending to each other. The American International Group (AIG) had to be rescued by the United States government. Those industries that produced material goods were clearly affected by the crisis, but to a much lesser extent than the financial industry. The idea that the network of world cities should be investigated through the lens of the extractive industry seemed even more feasible than before, because the sector is clearly still playing an important role in the world economy. The fact that the largest investment banks on Wall Street and the AIG had to be rescued by the US government underlines another dimension in world city formation often only considered as an addendum, the role of nation states in the production of world cities. Had the US government not stepped in to rescue the large investment banks on Wall Street and other financial institutions, New York's position as a financial centre would have changed more dramatically. 4 A considerable part of this dissertation is tracing the origins of world city research and its evolution over roughly the last quarter of a century. For John Friedmann (1986) world cities had become the 'basing points' for global capital. This is an important conceptualisation of world cities, because it means that world cities are not necessarily the location of capital itself, but these basing points are the places from where capital is being controlled. World city research is also closely related to world-systems theory (Friedmann and Wolff, 1982) and by extension to dependency theory, but this theoretical connection has not been emphasised in more recent writings on world cities. Importantly, the world economy today needs to be considered a single system and cities play an important role in organising this system. I therefore want to return to world city research's origin in the world-systems perspective. Furthermore, this dissertation is in many respects a response to the relatively recent critique on world city research, which charges that world city research is overly deterministic, based on outdated concepts of development such as first and Third World and pays too much attention to ranking cities according to their economic might, which is often based on inaccurate data (Robinson, 2006, pp. 95-97). This critique also posits that world city research establishes some cities as exemplars, which others should follow, but world city research was in its inception not intended to lead to such a conclusion. Tracing world city research to its origins in dependency and world-systems theory, will contribute to providing a more nuanced view of this research. The idea that some cities are exemplars and others are followers, implies that there is a single pre-defined path to development, which all societies will follow in an evolutionary process. The neo-Marxist traditions of dependency and world-systems theory have rejected this view and were in fact a response to these evolutionary views of development from both mainstream liberal as well as orthodox Marxist economic traditions. Dependency and world-systems theory instead consider the world economy a single system, where the distribution of wealth is based on unequal exchange, which contributes to specific development paths of cities around the world. A network of cities facilitate this uneven exchange. Viewing the articulation of the world economy through a network of world cities by using the extractive industry as a starting point provides an opportunity to view this network of world cities from the periphery. The aim is to provide a perspective from the bricks and mortar capital investment, in this case the mines of large mining companies in 5 South Africa, and analyse how decisions that are made far away and implemented through a network of cities affect people in these communities. While this research was conducted within the theoretical framework of world city research and world-systems theory, the focus was on the vertical connections from the world city to the periphery or the places of production. Conducting this research within the extractive industry was an opportune choice. On the one hand it is a basic, elementary ingredient into the world economy and its nature of being extracted from the earth makes an easy to trace commodity. Moreover, it is an industry that is largely organised in a monopolistic fashion with very few large companies dominating global production. Following a world-systems perspective, I suggest that world city research needs to establish an urban geography of the control over all monopolistic processes in the world economy, as these yield the highest profits and are the mechanism to maintaining a dominant position in the world economy (see Amin, 1997; Arrighi, 1994; Wallerstein, 2004). The dissertation is divided into three parts. Part I, consisting of Chapters 1 and 2, outlines the theoretical foundations of the research and the methodological approach. Part II (Chapters 3 and 4) investigates the ownership of the global mining industry with a particular view to the spatiality of control over the industry. Part III (Chapter 5 and 6) focuses on how the control and command functions exercised through the network of world cities are partially shaped by the South African State and  specific actors in the city of Johannesburg. It also examines how life in the mining communities of South Africa is affected by corporate and government decisions. As mentioned above this work is theoretically grounded in dependency and world-systems theory. This theoretical foundation is important in several ways. It emphasises the interconnectedness of places within the world economy and considers these relationships in terms of uneven development across the globe. This emphasis also stresses that processes in one part of the world can profoundly affect life in other parts of the world, no matter how distant. Andre Gunder Frank (1972) termed this process The development of underdevelopment, emphasising that the perceived underdevelopment of of places in the third world is a result of unfair trade relationships rather than of the traditional way life in these places. He further pointed to a metropolis-satellite structure through which these relationships were facilitated. According to this structure large cities in Latin America (and other parts of the third world), act as satellites to the global 6 metropolis (consisting of North America and Western Europe) and smaller cities in turn act as satellites to the larger satellites. Through this process surplus is channelled from the periphery to the core. Moreover, dependency and world-systems theory emphasise the role of monopolies in the capitalist world economy as a separate layer above the market economy and conceptualise this layer as one of the mechanisms to maintain uneven relationships between core and periphery. Chapter 1 outlines these theoretical underpinnings in more detail and provides a foundation as to how this theory should still inform world city research. Chapter 2 outlines the evolution of world city research and the methodological approach applied in this dissertation to reground this research in the theoretical traditions of dependency and world-systems theory. World city research as we know it today dates back to roughly the early 1980s, when John Friedmann and Goetz Wolff (1982) published their paper World city formation: an agenda for research and action. While world city research from that time onwards emphasised the role of a network of cities, there was very little research undertaken that would empirically investigate the connections between cities. Instead the research has largely relied on attribute data for individual cities, as for example the prevalence of corporate headquarters. Later world city research was focused almost entirely on connections based on common office locations of APS firms emphasising a city's position within the network of cities, rather than the strength of cities as nodes in the network. However, this research focused strongly on the relationship between world cities with virtually no empirical investigation of the relationships between world cities and peripheral areas of the world economy. The chapter outlines how the concept of vertical world city research can contribute to a better understanding of world cities' relationships to places around the world. Methodologically the research applies the extended case method (Burawoy, 1991) to contribute to the theoretical development of world city research. In order to establish a bottom-up approach to world city research and trace the connections from the sites of capital investment to the places of control, I am tracing the ownership of the world's ten largest non-fuel mineral producers. These ten firms account for more than one third of the world's non-fuel mineral production. Chapter 3 investigates the geography of control over these ten firms and the complex relationship between the locations of the owners, the corporate headquarters and the stock exchanges, where these 7 firms can be traded. The result is by no means a list of the usual suspects in world city research, but a mix between 'traditional' world cities, large cities in middle-income countries and small towns. Cities are switchboards of processes for mining activities around the world. Chapter 4 takes a similar approach to Chapter 3, but with a particular focus on the platinum industry. The focus on this specific resource makes it possible to trace almost the world's entire platinum production to the mine level and establish, which company accounts for which proportion of platinum production. This information in turn can be used to trace the ownership of these firms and establish a geography of control over the world's platinum industry. Much of the ownership over the industry remains in places close to the production, i.e. the mines and mineral deposits, but 'traditional' world cities and smaller places around the world include a significant share of the ownership as well. The approach to world city research taken in Part II is also a contribution to the discussion of what constitutes power in a world city network. In recent world city research with its focus on a horizontal network of world cities, power is generally conceptualised as 'power to', rather than 'power over'. 'Power to' refers to the power to service global capital, i.e. the ability to provide the inputs required by capital to function smoothly as would be done by the APS (Taylor, Walker, Catalano & Hoyler, 2002, p. 232). On the other hand direct ownership can be conceptualised as 'power over', i.e. the ownership power over capital, which may, however, still depend on the specialised input of the APS. In Part III the attention of the dissertation turns more directly to South Africa and more specifically to Johannesburg. Chapter 5 covers the historical development of the mining industry in South Africa. It also investigates the role of specific institutions and actors from the apartheid years in shaping the political economy of contemporary South Africa and the role of Johannesburg in the global economy. While the apartheid state provided long-term protection and a stable supply of cheap labour for the mining industry, it maintained stringent exchange controls and South African capital had to re- invest within the country. However, with the end of apartheid, exchange controls were lifted and much South African capital could move overseas, especially to London, contributing to the growth and status of this distant world city. 8 The relationship between Johannesburg and the South African state as well as the affects of mining on communities in South Africa are the focus of Chapter 6. When South Africa re-engaged in the world economy as a democratic nation, it took a distinctly neo- liberal path. Under the new government exchange controls were abolished, policies of nationalisation that had been the corner stone of African National Congress (ANC) policy for years were shelved and market mechanisms were seen as the solution for socio- economic development that would eventually benefit all South Africans. The success of these policies remains to be debated, but it is clear that they shaped the role Johannesburg plays as a business centre in both the national and global economy in very specific ways. Johannesburg is clearly the leading business location in southern Africa, but it also occupies a very specific position in the control and command chain of global capital. On the one hand it is the location for strategic control over a large sector of the mining industry in South Africa, southern Africa and further afield. On the other hand, it is the location for operational control, often receiving inputs from other places, especially London, on what strategic goals should be achieved by the corporation. The decisions made in Johannesburg will affect mining communities across the sub-continent and the city must therefore be considered a command and control centre. In the final concluding chapter I summarise how the idea of vertical world city research can provide new perspectives in world city research. While the construction and analysis of world city networks based on APS firms has significantly enhanced our understanding of the horizontal connections between world cities, the connection between the capital commanding functions supposedly hosted in world cities and capital itself, i.e. bricks and mortar investments, was largely assumed rather than empirically investigated. Moreover, the conceptualisation of power as the ability to service global capital, has greatly tilted world city research towards those places hosting large service sectors, especially APS, while ownership power or direct ownership control over productive assets has been relatively neglected. This dissertation integrates factors such as production, labour or the environment as well as political forces into world city research more strongly than before. The result is a mapping of the command and control centres in the global mining industry, with cities as diverse as Santiago, Johannesburg and New York emerging as the most important centres of control over the mining 9 industry. This spatial organisation of the mining industry can be applied to identify a changing spatial organisation of command and control functions in the world economy. 10 Part I 11 Chapter 1: New Socio-Spatial Constellations, Old Division of Labour There are two basic premises that will underlie the main argument of this dissertation. The first one is, following the tradition of the dependency school and world- systems theory, that the global economic system is based on a relationship of unequal exchange between those places constituting the core of the global economy and those constituting the periphery (Wallerstein, 2004, p. 17; Emmanuel, 1972). The second basic premise is, following Arrighi (1994) who bases his observation on Fernand Braudel and partly on Adam Smith and others (e.g. Wallerstein, 2004, p. 23 - 28), that contrary to mainstream social science analysis capitalism must be understood as distinctly different from a market economy and should be considered an additional layer above the market. I will provide a more detailed discussion of these two premises later in this chapter, which will trace the development of dependency theory and world-systems theory over the last half century or so and establish a theoretical basis for arguing that contemporary world city research needs to be undertaken in the tradition of these theories. In the following chapter I will review the specifics of the world city literature since roughly the 1980s and sharpen the main theoretical argument of this dissertation. Development and dependency There has long been a prevalent view in wealthy countries that poorer countries are "underdeveloped", "less advanced" and simply at an earlier stage of a natural path of development that the rich countries have passed some time ago. The publication of Walt Rostow’s (1960)  The stages of economic growth: a non-communist manifesto is probably the primary example from the mainstream liberal view that considers economic development a natural course. However, a similar view prevails in orthodox Marxist theory (e.g. Warren, 1980), according to which capitalist industrialisation will have to precede socialism, which would be achieved through a revolutionary process. A detailed discussion of these views is not necessary here. The point of departure for this thesis is the introduction of dependency theory in the 1950s and 1960s. 12 Dependency theory emphasises the interrelationship of the poorer parts of the world with the wealthier ones, rather than viewing them as separate entities that develop in isolation and along a pre-determined path. Moreover, dependency theory points to an empirical vacuum in the formulation of mainstream liberal development theory, which at that time was mainly based on examples of the so called "developed metropolitan" countries rather than the so called "underdeveloped" world, but established theoretical claims about poorer and "underdeveloped" countries (e.g. Frank, 1969, p. 3). Dependency theory incorporated some of the major traits from the work of Latin American structuralists in the United Nations Economic Commission for Latin America (ECLA), but took a more radical view on the potential for development of third world nations under capitalist conditions. The ECLA’s director Raul Prebisch, explained the relative lack of development in Latin American countries with the unequal effects of international trade on the so-called developed and developing countries. The "developed" countries had a comparative advantage in manufactured goods and increases in productivity should have reduced the price of these goods. However, organised labour was able to achieve real wage increases for workers in the developed countries and prices would stay high. As these goods were exported to Latin America and other third world regions consumers in these regions paid the prices for rising wages elsewhere. While Latin American countries have a comparative advantage in primary goods, prices for these goods have been declining and because these goods were primarily exported, the decline in prices aided primarily producers in the developed world. The rich or developed countries were able to maintain this structure through their political and economic clout in the international market (Sheppard, Porter, Faust and Nagar, 2009, pp. 84-85). Central to the argument of dependency theory is that poverty is not primarily caused by internal conditions of poor countries, but the expansion of capitalism by means of colonialism and imperialism, which leads to the channelling of  resources to richer countries while leaving the poor countries dependent on the rich ones. Andre Gunder Frank (1969, p. 6), who popularised dependency theory in the anglophone world, albeit fairly late, compared to its diffusion in Latin America, described this mechanism as a metropolis-satellite structure. For Frank, the countries of Western Europe and North America constituted the world metropolis. Through a system of satellites resources would be channelled to the world metropolis. Often this system would be multi-tiered with large 13 primate cities in colonies or former colonies being the direct satellite to the world metropolis, but at the same time multiple smaller cities or towns in the colonies would serve as satellites to the larger satellites. These smaller towns in turn could have additional satellites (see also Armstrong and McGee, 1985). An often ignored aspect of Frank's work - Armstrong and McGee (1985) being a notable exception - is the inherently urban structure of this metropolis-satellite relationship. While he, perhaps inappropriately, aggregated all of Western Europe and North America into the world metropolis, there is a strong emphasis on the role of cities as satellites. He quotes two examples (Frank, 1969, pp. 5-6) from Latin America, where the city serves as the point of exploitation for the local indigenous population. He continues with the examples of Buenos Aires, Sao Paulo and Montevideo, which established themselves as regional metropolises, but remained satellites to London and the United Kingdom as a whole and later the USA. These hints at an early theory of a hierarchically organised world city network were not developed by Frank or any other scholar of dependency, and have rarely, if ever, been taken up in the contemporary world city literature. I will engage in a more detailed discussion of this argument in dependency theory in the next chapter but will first focus here on another major mechanism of dependency: the role of monopolies or quasi-monopolies in capitalist production. The role of monopolies is emphasised, but relatively poorly developed in Frank's early work. He mainly refers to the monopolistic nature of the metropolis-satellite structure, suggesting that the metropolis produces goods and services under monopolistic conditions, while the satellites compete fiercely with each other. This work was part of the development of the “monopoly capitalism” perspective first articulated by Paul Baran and Paul Sweezy (1966). Later world-systems theory deals in much more detail with the role of monopolies (see e.g. Braudel, 1992 [1984]; Taylor, 1996; Wallerstein, 2004). The world-systems perspective While dependency theory provided a persuasive model of globally uneven development and provided much needed alternatives to both mainstream perspectives of modernisation theory as well as orthodox Marxists' views on development in pre- 14 determined stages, it remained a somewhat deterministic model as it viewed the world economy as a fairly static and inflexible structure. The theory was not able to account for some newly industrialised countries that seemed to have broken out of their dependent position. World-systems theory, building on dependency theory, addressed this deficiency by developing a more flexible model and emphasising the role of the global social division of labour. According to most world-systems theorists (e.g. Wallerstein, 2004, p. 23) the modern world-system has its origin in sixteenth century Europe.1 While the term world-system (or Fernand Braudel’s economic-monde) refers to a system that is a world and does not necessarily need to include the entire globe, it is largely agreed that at the beginning of the 21st century there is a de-facto global economy that constitutes a single world-system. Crudely put, this means that all parts of the globe are linked to each other through the global division of labour. This world-system, which is now de-facto incorporating the whole globe, is a capitalist world-system. This emphasis on it being a capitalist world-system is essential for understanding the dynamics created by the system. Dependency and world-systems theorists also distinguish themselves in this respect from more orthodox Marxists as they include all regions that are in some way tied through capitalist processes into the capitalist world- system, even if there is no wage labour involved in a specific region. Classical Marxist interpretations of capitalism differ in that they consider capitalist processes as taking place primarily within societies, a view increasingly expressed in the 1970s, for example by Ernesto Laclau (1971), as a response to the emphasis on external factors in dependency and world-systems theory (Armstrong and McGee, 1985, pp. 27-29). Capitalism, in the definition of the world-systems theorists, is distinctly different from a truly open market economy. The capitalist system gives priority to endless accumulation, defined as accumulating capital with the main goal of accumulating more capital in a continuous and repeating cycle, but this endless accumulation depends on markets that are only partially free. Wallerstein (2004, pp. 25-26) argues that a truly open and free market would allow buyers to bargain down the price of any product to an "absolute minuscule level of profit" and would therefore make endless accumulation 1 There is some historical debate about the origin of the modern world-system and the role of Asia prior to the sixteenth century (e.g. Frank and Gills, 1993; Frank, 1998). However, for this dissertation the contemporary organisation of the global economy is of primary importance and while the historical debate is useful in its own right, it does not need to be discussed in detail here. 15 impossible. Thus monopolies or quasi-monopolies are a necessary pre-condition for the existence of a capitalist system, because monopolies allow producers to set the prices higher and achieve higher profits, which in turn can be invested again into accumulating more capital. In practice there are few, if any, true monopolies of one firm controlling a global market or even a segment of it, but there are plenty of quasi-monopolies, where a small number of firms share the market for a specific product. Usually over time these monopolies weaken, but last long enough for significant profits to be accumulated, which can be re-invested for further, expanded accumulation, rather than simply small profits for immediate consumption or small scale re-investment. Capitalism can thus be viewed as a separate layer above the market, where such competition-driven, small-scale accumulation predominates (e.g. Arrighi, 1994, p. 10; Wallerstein, 2004, pp. 24-27). Contemporary examples of monopolies include software production, genetically modified foods and seeds, pharmaceuticals and commercial aircraft. These monopolies can only exist with the active support of states and their regulatory environments that protect these monopolies and defend them by force, if necessary. States for example protect software producers with copyright laws, the pharmaceutical and genetically modified food and seed industry with patents, and aircraft producers with tariffs and subsidies. Of course, copyrights can be undermined, patents run out eventually and even tariffs and subsidies do not provide endless protection from cheaper competitors. Nonetheless, these mechanisms are effective enough and last long enough, perhaps roughly 25 to 30 years, for their producers to make significant profits. Highly competitive industries by comparison have very low profit margins allowing for significantly less capital accumulation. Most products become more competitive over time and thus less profitable. For example, while most manufacturing in the 1950s was still conducted under quasi-monopolistic conditions, parts of the manufacturing industry, e.g. textile and garment production or low-end electronics, have become highly competitive and a large proportion of global manufacturing is now scattered around the globe with the market seeking the locations with the lowest cost (see especially Wallerstein, 2004, pp. 23-41). What has to go hand in hand with the existence of monopolies in global production and endless accumulation is the global division of labour, where generally speaking those employed in monopolistic production, e.g. a software programmer, earn high wages, whereas those employed in highly competitive industries, e.g. a worker in 16 textile manufacturing, earn low wages. Based on this division of labour, world-systems theorists, have divided production processes into core-type and periphery-type production. Core-type processes are those producing under monopolistic or quasi- monopolistic conditions, while those producing under competitive conditions are periphery-type ones. Hence, there are core states as well as periphery states, depending on the type of production that is dominant in the respective states. And there are semi- peripheral states, which have an about even mix of the two types of processes. However, there is no semi-peripheral production type (Wallerstein, 2004, p. 29). The idea of core and periphery as types of production processes signifies a considerable theoretical development compared to dependency theory as it helps account for states that are moving up or down the world-system hierarchy from one category to the other. However, as the capitalist system depends on the division of labour, it is not possible for all states to become core states at the same time. The core-periphery model also allows us to picture the world more as a mosaic of states rather than a star shaped pattern with the world-metropolis at the centre and a multiplicity of satellites around it. In fact, it already hints at a network structure, a theme that was not fully developed in early world-systems arguments but became a central paradigm in later work by Hopkins and Wallerstein on global commodity chains (Hopkins and Wallerstein 1986), as well is in related work by authors like Manuel Castells, who conceived of the global economy as a networked "space of flows" (Castells, 1996). Network approaches also became central to world city research, as this evolved out of world-systems approaches from the 1980s onward. World-systems theory (especially Wallerstein, 2004; see also Amin, 1997) stresses the role of the state as the protector of the system of capitalist production. State power is used to maintain quasi-monopolies, grant corporations access to natural resources or suppress labour unrest. Hence, there is a strong nexus between political power and economic power. At the same time capitalists need a multiplicity of states to allow them to assert pressure on individual states by threatening to move capital to other locations, in case of non-cooperation by a given state. The development of the abstract construct of the state as a sovereign power (rather than a monarch) is relatively new and generally dated back to the signing of the Treaty of Westphalia in 1648 (Wallerstein, 2004, p. 42; Taylor, 2004a, p. 15). Others argue that dynastic and absolutist political 17 communities dominated the cultural landscape for much longer and existed until World War I (Teschke, 2003). There is little attention paid to the role of cities in world-systems theory, but Braudel (1992, pp. 25-35; see also Taylor, 2004a, pp. 13-15) already emphasised the role of a leading word city in a world-system, successively centred around cities in southern and north-western Europe. Braudel’s view is that there was always one city that was dominant—or, in his terms a "super-city"—to which all other places were subordinate, a clearly hierarchical structure. However, the rise of the modern inter-state system after 1648 undermined the power of cities. Braudel’s analysis (1992) focuses on examples from Europe and to a much lesser extent from other world regions. While this Eurocentric nature of the analysis is problematic and may be somewhat limiting for generalisations at a global scale, it is important to remember that his historical conception of a world-economy stresses that this economy is a system stretching across boundaries, but not necessarily including the entire world. Polarisation at a world scale Samir Amin (1994, p. 90) is partly critical of the view of the world-systems school (e.g. Wallerstein, 2004; Arrighi, 1994; Braudel, 1992) that there was typically a single hegemonic power, either a city or a state, in the capitalist world economy. Amin’s work emphasises polarisation through an uneven division of labour as a pre-condition for a capitalist world economy. Thus, his views are hardly in fundamental disagreement with the world-systems school, but there are disagreements on details (see also Amin, 1974). Amin’s theory of Accumulation on a world scale (1974, see also 1994) emphasises the differential remuneration of labour at equal productivity in the global core or the centre and the periphery leading to polarisation in worldwide accumulation. Although Amin (1994, 1997, p. 1) recognises that there has been unequal development in different regions since antiquity, he maintains that capitalist polarisation has only occurred since the mercantile era beginning around 1500. With reference to Lenin, Amin (1994, pp. 86- 87) pays particular attention to the emergence of monopoly capitalism around 1880, when colonialism was not a new phenomenon, but saw a significant expansion. From this point on wages in the centre kept pace with productivity. Previously real wages were 18 stagnant or declining in the centre allowing for capital accumulation through the extraction of surplus value, but with "a new social compromise between capital and labour" (Amin, 1994, p. 87), wages increased with productivity forcing a reduction in wages in the periphery to allow continuous capital accumulation in the centre and leading to polarisation. After World War II Fordism manifested the system of continuously rising wages in the centre, putting pressure on the peripheral countries to lower wages and increase productivity to keep the system of endless accumulation at pace. Fordism was also a new form of regulation and functioned to even out the regularly occurring deep fluctuations in the global economy that occurred in the first half of the 20th century - or "the extended nineteenth century" - through sustained consumption power in the core countries (Amin, 1994, p. 54). Relations prior to 1880 were not equal but more directly structured so that production in the periphery would serve accumulation in the centre, rather than through the difference in wages for equal productivity. This means that prior to 1880 wages for the European and North American working class were declining as they were in peripheral countries, serving accumulation. After around 1880 wages kept pace with productivity in the industrialised centre. A development that had to come at the expense of the working class in the agricultural and mining periphery to facilitate accumulation (Amin, 1994, p. 87). The result according to Amin was polarisation at a global scale. This system was assured to function because commodities and capital can move freely, while labour remains confined to its own national containers (Amin, 1994, p. 75). While it can be argued that labour was certainly not completely immobile at the end of the 19th century and the beginning of the 20th century, goods and capital were likely more mobile than people. Amin (1994, p. 88) argues that Marx "relegated primitive accumulation to pre capitalist history, underestimated polarization, and proposed a law of pauperization in general without clarifying that in fact this law operates at a world level but not at a level of centres artificially separated from the global system", because Marx wrote at a time when the industrial revolution was taking place and before monopoly capitalism manifested itself. Hence, Marx falsely interpreted development as taking place in stages, a paradigm that remains highly influential with orthodox Marxists and mainstream liberal economists and was the basis for much critique from the dependency and world-systems schools on mainstream development theory. Unlike the later born Lenin, Marx could not see the 19 level of geographic polarisation on a world scale, which just manifested itself around the time of his death. Hence, he relegated primitive accumulation to pre-capitalist history. Others have given Marx more credit in actually viewing primitive accumulation as partly a historical phase, but also a continuous condition under capitalist production (e.g. de Angeles, 2001; Glassman, 2006). Simultaneously to the beginning of wages keeping pace with productivity from around 1880 onwards, came the manifestation of monopoly capitalism as Lenin emphasised (Amin, 1994, p. 87). Innes (1984, p. 17) summarises Lenin’s conception of monopoly capitalism as consisting of five "basic features": 1. The development of the tendencies of concentration of production and capital to the stage where monopolies emerge which play a decisive role in economic life; 2.  the merging of  bank capital  with  industrial  capital  to create finance capital giving rise to the emergence of financial oligarchy; 3.  the  increasing  importance  of  the  export  of  capital  (as  opposed  to  that  of commodities); 4. The rise of international monopoly combines which share the world among themselves; and 5. the complete territorial division of the world among the major capitalist powers. There are many limitations to using this simplified concept of monopoly capitalism as an analytical basis. Moreover, the world has changed since Lenin wrote on the topic, but the basic points remain valid and can be useful for the analysis of the world economy at the beginning of the 21st century. For Lenin, monopoly capitalism was closely linked to colonialism and imperialism (Amin, 1994, p. 86). Around 1880 colonialism saw a substantive expansion and states used their political and military power to protect the development of monopoly capitalism in the imperial countries. Amin (1997, pp. 3- 5) argues that in the contemporary world the core countries use their "five monopolies" to continue to extract value from the periphery through unequal exchange. These five monopolies are: 1. technological monopoly; 2. financial control of worldwide financial markets; 3. monopolistic access to the planet's natural resources; 4. media and communication monopolies; 5. monopolies over weapons of mass destruction. By this definition the role of monopolies goes beyond the monopolies in production processes, enjoyed by the global core and emphasised by Wallerstein’s (2004) world-systems theory. These five monopolies include the nexus between political 20 and economic power, between state and private sector, and they all work in concert to protect one another and by extension the power of the global core. The dominant unit of analysis in this world-systems approach remains the nation state. While the nation state maintains its role as the sovereign power that it assumed after 1648 through jurisdiction over military power, foreign trade and foreign relations and territorial border control, it has lost some of its political and economic influence to both supra-national as well as sub-national jurisdictions. Although Amin (1994), with his thesis of polarisation, has perhaps relied most on the nation state as the primary unit of analysis, he recognises (1997, p. 3) that the "auto centred nation-state" is eroding and that there are "new dimensions of polarisation" which do not neatly follow the "great divide" of "industrialized centre/non-industrialized peripheral regions". Moreover, Amin (1994, p. 88) speculates that we are "returning to a system similar in some ways to that of Marx's time". In other words stagnating or declining labour costs in the centre due to the decline in Fordist production methods allows for some higher wages in newly industrialised countries, while still ensuring accumulation. While this means that we are still seeing polarisation at a world scale, the nation state as a unit of analysis seems to be more inadequate than ever before. New socio spatial constellations and the global division of labour Geographers have often argued that the nation state is overly privileged in social science research and that greater attention needs to be paid to other scales above and below the nation state (e.g. Brenner, 1998, 2004; Swyngedouw, 2004). In addition to scale, other spatial paradigms have influenced the work of geographers. Eric Sheppard (2002) examines three prominent paradigms in geography through which to analyse globalisation: place, scale and networks. He then draws attention to positionality, a fourth spatial paradigm that has been underprivileged in geographic research. Geographers have stressed that place is not an obsolete concept in the current phase of intensified globalisation as suggested by some popular and scholarly authors (e.g. Friedman, 2007; Cairncross, 1997). The characteristics of places still matter regarding locational decision 21 making and Marxist geographers have long argued that capital requires spatial fixes in economic agglomerations to foster accumulation (e.g. Harvey, 1982; Smith, 1984). Similarly, geographers have been battling with the question of scale and identification of an appropriate unit of analysis for some time. One of the most detailed discussions on the question of scale in political economy research is offered by Neil Brenner (2004) in his book New State Spaces: Urban Governance and the rescaling of Statehood. While a simplified analysis of scalar changes in state administration suggests that the nation state is increasingly sharing or handing over a wide variety of policy responsibilities to both the urban or sub-national level (downscaling) as well as to supra- national organisations, like the World Trade Organisation or the European Union etc. (upscaling), a more nuanced view proposes that state rescaling consists not only of up- and downscaling, but a fundamental "qualitative transformation" (Peck, 2002, 332) of the state. Processes of globalisation do not just shift from one scale to another and take place at one specific level, they are also intrinsically linked across scales, connecting these scales and actors at various levels. Research on global processes should reflect this (Dicken, Kelly, Olds and Yeung, 2001, p. 90). These processes of re-scaling are directly connected to processes of globalisation or more correctly the current round of globalisation, as globalisation is a process that has existed for at least 500 years since the ongoing expansion of capitalism and what we are seeing now is a period of transition and intensification of global processes (see Wallerstein, 2000). John Urry (2003, p. 4), for example, suggests that there “has been an increase in structural globalisation”, i.e. an increased density of international and global interactions compared to the increase of such interactions at the national or sub-national scale. He further considers (the current round of) globalisation to be “an emergent feature of the capitalist economy as a whole” (Urry, 2003, p. 4). Eric Swyngedouw (2004) proposes the concept of 'glocalisation'. He argues that institutional and regulatory processes are shifting both upwards, e.g. from the nation state to the global scale, as well as downwards to the individual body and to local, urban and regional configurations. This rescaling will "alter social power geometries in important ways". Swyngedouw (2004, p. 42) concludes: 22 The socio-spatial transformations that have characterised the past two decades or  so  are  testimony  to  such  scale  restructurings  through  which  older  power relations are transformed. The disturbing effects of these recent ‘glocalisation’ processes  suggest  that  the  spaces  of  the  circulation  of  capital  have  been upscaled,  while  regulating  the  production/consumption  nexus  has  been downscaled, shifting the balance of power in important polarising or often plainly exclusive ways. The rescaling of the state and the production of new articulations between scales of governance, in turn, redefines and reworks the relationship between  state/governance  and  civil  society  or  between  state  power  and  the citizen. Sassen (2000) argues that globalisation is not a zero-sum game, in which any gains in the global economy are a loss in national economies and vice versa. The global and the national are interconnected and dependent upon each other and "the global economy to a large extent materializes in national territories" (Sassen, 2000, p. 374). Similar shifts of jurisdiction are occurring in terms of governance, which has seen a "reconfiguration of political power" and a shift of political power to various scales without eroding the nation state (Held, 2000). While globalisation or an intensification of global processes is occurring, it is necessary to examine the causes for globalisation, i.e. what decisions enable this intensification of global processes, rather than viewing globalisation as a self-propelled, 'natural' process or the explanatory variable, i.e. it is not globalisation that causes inequality, rather inequality enables globalisation to occur in its current form. Thus, globalisation is enabled by the "switchability" of geographic scales, i.e. the ability to organise social and material processes at different - smaller or larger - scales than before (Yeung, 2002). The third important lens through which to study globalising processes is by focusing on networks. Manuel Castells's  The rise of the network society (1996) and other works (e.g. 1989, 1997, 2001) was probably the single most influential contribution to the study of global networks, introducing the concept of the space of flows. This work suggests that with the rise of more efficient communication technologies, place based decision making is becoming less important. Thus the space of flows is created in horizontal networks that touch on individual places but is not place bound and represents a permanent flow of ideas and decisions. Thus places, as for example world cities, become highly interdependent upon each other (see Taylor, 2004a; Derudder & Taylor, 2005). It is then the strength of the network that a person, a city or a place participates in, which is essential for the success of this person, city or place (see Sheppard, 2002, pp. 23 315-318). Dicken, Kelly, Olds and Yeung (2001) advocate a network methodology to studying globalisation, in which the network is the "foundational unit of analysis" (p. 91). Moreover, this network methodology must incorporate multiple scales, i.e. not overemphasise horizontal connections, and it must not privilege a "single institutional or organizational locus of analysis" (p. 89). While place, scale and networks have received considerable scholarly attention, the concept of positionality has been underprivileged in research on economic globalisation (Sheppard, 2002). Positionality provides a viewpoint that requires us to view all prospects for local development by considering a place's relationship to other places. Sheppard (2002, pp. 325-326) outlines four examples that illustrates how considering positionality affects the study of globalisation. - Space remains important. Technology has enabled humans to overcome greater distances faster. However, differences in access to telecommunications and transportation systems increase disparities amongst people and places. Therefore, positionality in this system is still important for an individual's or a place's access to resources. - While neo-liberal policies like the Washington Consensus emphasise local conditions as the key to development in specific places, unequal development is not just a result of local conditions but of a place's positionality in the global economic system. Changing local conditions will have little to no effect, if positionality matters, and increased competition amongst places will not result in increased prosperity for all, but force some locations to compete on the basis of exploiting their human and environmental resources and to participate in the so called 'race to the bottom'. - Attention to positionality refutes the view that there is only one single path to development that all places have to follow in order to achieve local development. The conditions in one place influence those in others and thus prosperity in one place is often enabled by poverty and exploitation in others. Hence, capitalism with its dependence on a spatially differentiated world cannot bring prosperity for all and the concept of positionality thus provides a counter narrative to the 'globalisation-as-modernisation' narrative. 24 - Positionality is important for successful resistance to globalisation in its current form. Consciousness of one's position in the global system enables more effective resistance through better cooperation amongst groups of people and more strategic targeting of one's opponents. Sheppard is amongst a minority of geographers who directly engage with the work of the dependency school, recognising that Andre Gunder Frank and James Blaut (Sheppard cites Frank, 1967, 1978; Blaut, 1976) came to the conclusion that the positionality of one place makes it dependent on the conditions in other places and that there cannot be a single pre-determined path to development, a conclusion that was also reached much later, through post-colonial theory, by Doreen Massey (Sheppard, 2002, p. 326; Shepard cites Massey, 1999). The emphasis on positionality also speaks to much feminist and post-colonial critique of concepts such as the space of flows or time-space compression in political-economy research. Positionality as a lens for globalisation research opens an avenue of inquiry that allows us to study global processes without privileging a specific scale, deriving theory on the basis of only a few places, or overemphasising the network or territorial nation states as the dominant spaces. The way globalisation works in specific places has been neglected in globalisation research by generalising experiences across large areas, if not generalising it across the global scale. Cindi Katz (2001) critiques David Harvey's (1989a) concept of time-space-compression as being informed by the perspective of capital. Her study of a village in Sudan provides a notable exception in that it engages in “scale-jumping” (Katz, 2001, p. 1216) from the specific village to global processes playing out in specific ways in that village. She argues that "from the vantage point of capital, the world may be shrinking, but, on the marooned grounds of places such as Howa [Sudan], it appeared to be getting bigger every day" (Katz, 2001, p. 1224). Katz therefore brings forward the idea of time-space-expansion. Emphasis on positionality and the theoretical underpinnings of the dependency school speak directly to this kind of critique and allow us to incorporate the views from different actors and places into research on globalisation, rather than relying on a single meta-narrative of processes at the global scale. 25 Gillian Hart's (2002) study of a global network in the garment industry, which she has located in two towns in KwaZulu-Natal in South Africa, is another notable exception. It provides an analysis of globalising processes from the vantage point of multiple actors in these two towns. Hart (2002, p. 293) charges that writing on globalisation is usually informed by the western experience of (late) modernity and that there is a “core centric 'global centrism'”. Her non-explicit emphasis on positionality provides a counter narrative to Western experiences and draws connections between various places across the globe without privileging a specific place, territory or scale. Similarly, positionality can incorporate the study of networks, but does not consider networked spaces or the space of flows as dominant. Hence, positionality as emphasised by Sheppard (2002), or earlier the dependency school, does not exclude place, scale and networks as lenses for analysis of global processes, but suggests that these are analysed in concert with each other. While some authors (e.g. Dicken, Kelly, Olds and Yeung, 2001; Sheppard, 2002; and for a more cultural perspective see Axford, 2005; Hannerz, 1996; Urry, 2003) already recognised the limits to over-emphasising one specific conceptual dimension in the research on globalisation, Jessop, Brenner and Jones (2008) go a step further and propose an approach that does not privilege a single dimension and recognises "the inherently polymorphic, multidimensional character of sociospatial relations". Their discussion of territories (T), places (P), scales (S) and networks (N) suggests that these four ordering principles of socio-spatial research "must be viewed as mutually constitutive and relationally intertwined dimensions of sociospatial relations" (p. 389). Jessop, Brenner and Jones (2008) argue that a TPSN research agenda, in which none of these ordering principles will be privileged, will enable researchers to shed new light on the historical periodisation of capitalist development. After a specific constellation of territories, places, scales and networks enabled the post-war Fordist-Keynesian spatiotemporal fix, "new TPSN combinations seem to be emerging that are more suited to a postnational, unevenly developing global economy" (p. 397). Samir Amin's (1994, p. 88) speculative reference to a return to a situation bearing greater similarities with that of Marx's times than the one characterised by the historical compromise between capital and labour in the global north reaching its highpoint in the post world war II Fordist era, is nothing else but a new or reshaped TPSN constellation. 26 This new TPSN constellation enables the current form of globalisation2, which is on the one hand a reaction to a crisis of accumulation in the 1970s with a new regulatory environment based on neo-liberal principles (see e.g. Harvey, 2005; Duménil and Lévy, 2004a, 2004b), but at the same time based on the global division of labour and uneven development long emphasised by dependency and world-systems theorists (e.g. Frank, 1969; Arrighi, 1994; Wallerstein, 2004; see also e.g. Harvey, 2006; Smith, 1984 for a more geographic perspective). However, the spatiality assumed by world-systems theorists, one that primarily recognises the nation state as the ordering principle, can no longer explain the current dynamics of the international division of labour and polarisation at a world scale. The early works of the dependency school perhaps viewed the world economic structure as too static and fixed and was not able to account for the up and down movements of some states or territories later recognised by world-systems theory, but it was more attuned to the socio-spatial details of places, territories, scales and networks. For example, Frank (1969) recognised the relationship of satellite cities, such as Buenos Aires, Sao Paulo and Montevideo, to the territorial hinterland within states or colonies and their scalar and networked relationships, with a particular emphasis on positionality. It is useful to pick up some of the traits outline by Frank and others in contemporary world city research. Towards a new spatiality of world city research The development of world city research from the 1960s (Hall, 1979) onwards, with significant new developments in the 1980s (see Friedmann and Wolf, 1982; Friedmann, 1986), is in many ways an attempt to grasp a new TPSN constellation within a world-systems framework, where the locus of political and economic power is shifting from the territorial nation state to a place, the world city. A process that is also a shift in scale as well as a shift towards a networked organisation of processes. Recent world city research, however, has usually emphasised one of two organising principles: place or network. On the one hand many world city studies are case studies of individual cities and local reactions to global change, their attempts at promoting themselves in a global 2 Following Yeung (2002) I take the view that it is the TPSN constellation that enabled the current form of globalisation. It is not globalisation (as some kind of natural force) that led to the the current constellation. 27 economy or discursive attempts to become a world city (e.g. Newman and Thornley, 2005). On the other hand, in particular the work of the Globalisation and World City (GaWC) Research Group emphasises the role of the network of world cities as an organising principle (e.g. Taylor, 2004a). What I am proposing in this thesis is a more nuanced world city research that pays attention to a wider variety of shifts and changes in organising principles of socio-spatial relations. While the emphasis on world cities as loci of command and control functions of the global economy is plausible, too much emphasis has perhaps been put on their networked character or their internal conditions. Two important aspects of global change are currently only poorly reflected in world city research. First, the relation of world cities to smaller cities, villages, settlements, agricultural land, forests and any other point on the surface of the earth. This means, there is little emphasis on scale and interactions between scales with a privileging of the horizontal space of the network. Second, world city research has almost categorically excluded the nation state (or in some cases regions, provinces or federal states) as a unit of analysis and constructed an unproductive dualism between the two, instead of recognising the deep connections and interrelationships between states and cities. Territory as a unit of analysis has therefore largely disappeared from world city research, although Taylor (2004a, p. 58) specifically considers states one of four agencies that are primarily responsible for shaping the world city network and the relationship between world cities and their respective nation states was initially recognised as one elementary dimension of world city research by Sassen (2001, p. 8). Parnreiter (2002) provides a notable exception to this trend as he explicitly analyses the link between Mexican state policy and Mexico City's rise as a world city. Moreover, in spite of the origins of world city research in world-systems theory (see Friedmann and Wolf, 1982), positionality is a declining concept in world city research with a greater emphasis on the horizontal organisation of the network (see Beaverstock, Doel, Hubard and Taylor, 2002; Taylor, 2005b). The following chapter will outline the evolution of world city research as well as recent critiques of the concept, before proposing a new research agenda for a vertical world city research. 28 Chapter 2: World City Research: A Review Our paper  concerns  the  spatial  articulation  of  the  emerging  world  system of production and markets through a global network of cities. Specifically, it is about the principal urban regions in this network, dominant in the hierarchy, in which most  of  the  world's  active  capital  is  concentrated.  [...]  These  vast,  highly urbanized  -  and  urbanizing  -  regions  play  a  vital  part  in  the  great  capitalist undertaking to organize the world for the efficient extraction of surplus. Our basic argument is that the character of the urbanizing processes - economic, social and spatial - which define life in these 'cities' reflect, to a considerable extent, the mode of their integration into the world economy. John Friedmann and Goetz Wolff,  1982, p.309,  first  paragraph of  their  paper "World city formation: an agenda for research and action" World city research: remembering the agenda Although the term world city has been in use for at least two centuries, being used by Goethe at the turn from the 18th to the 19th century, the term has only achieved its contemporary significance within human geography a little more than two decades ago. Peter Hall (1979 [1966]), building on Patrick Geddes (1915), gave the term significance with the publication of his book The World Cities, which was an examination of seven of the world's largest agglomerations at the time. Hall characterised the world cities as the places where "a quite disproportionate part of the world's business is conducted" (p. 1), including the exercising of political power. For Hall world cities were also the centres for banking and finance, transport and communication and the world media. While he recognised all the global or international influences, his analysis was situated primarily within these seven cities in the context of the sovereign nation state and he did not bring forward a world city hypothesis of interconnected cities forming an urban network of world cities, as John Friedmann and Goetz Wolff (1982) did one and a half decades later (see also Friedmann, 1986, 1995). The work by Friedmann and Wolff (1982) was arguably the starting point for world city research as we know it today and when they wrote their now seminal paper "World city formation: an agenda for research and action" they drew attention to a number of socio-spatial processes and relationships that were articulated through "a global network of cities". I want to draw attention to some key traits in the quote above. 29 Friedmann and Wolff considered this network the backbone of "the emerging world- system of production and markets", which played a central role in "the great capitalist undertaking to organize the world for the efficient extraction of surplus". To Friedmann and Wolff the network of world cities is the central component of the global economy and there is no longer an imperial centre such as Rome or Venice, or even London (cf. Braudel, 1992). Without the network of world cities, the "world spanning system of economic relations would be unthinkable" (p. 312). The authors provide a world-systems perspective for the analysis of urban regions and an urban perspective for the analysis of the world-system. I wish to follow this view in this dissertation. The 1980s and early 1990s were in many ways the formative years of world city research. In addition to the works by Friedmann (1986) and Friedmann and Wolff (1982) the publication of Saskia Sassen's (2001) book The Global City at the beginning of the 1990s was a pivotal step in the evolution of world city research. Although Sassen tries to deliberately shift the terminology to global city rather than world city, arguing that a global city is characterised by articulating global flows (partly enabled through improvements in global telecommunication) as opposed to being merely a dominant city in the world as the historical use of the term world city would suggest (cf. Hall, 1979; Braudel, 1992), both terms remain widely in use. Perhaps the single most important contribution of the book to world city research is the identification of the APS, including banking, accounting, insurance and law, as being dominant in articulating global flows, due to their capital commanding functions. In short, these producer services are understood as elementary services for the successful investment of finance capital. Other scholars include a much larger group of services in the producer services category (e.g. Daniels, 1993, p. 26), but the relatively narrow selection presented by Sassen became prevalent in later world city research. It is this narrow focus on these producer services that is central to my critique of contemporary world city research. In order to broaden the scope of world city research, I will revisit the works by Friedmann and Sassen, which are perhaps amongst the most widely cited and the least read. Friedmann and Wolff (1982) set out to draft "an agenda for research and action" and Friedmann (1986) stresses that "the world city hypothesis" is intended as a "framework for research" and that "it is neither a theory nor a universal generalization about cities, but a starting-point for political enquiry" (p. 69). Friedmann (1986) presents 30 seven loosely connected theses. While some of these theses feature prominently in contemporary world city research others have received less attention within the world city framework and are primarily discussed under different research agendas within urban studies. However, these theses are interrelated and it will be useful to reconnect at least some of these themes in world city research. Friedmann (1986, pp. 69, 71) describes the world cities as "basing points" for global capital. Considering world cities "basing points" for global capital has two important implications: a) The world city itself is not (or at least not necessarily) the location of capital investment, although obviously some capital investment is always required, of first importance being office buildings and other real estate, and b) the world city is the location of control for this capital. The role of the world city as a location of control has received considerable attention in world city research, however, its relationship to the sites of capital investments, i.e. the sites that world cities (or more precisely actors in world cities) presumably control, has not. Friedmann (1986, p. 70) considers world cities within the capitalist world-system, i.e. world cities will have certain functions within the international division of labour assigned to them. On the one hand this will greatly influence the internal development of the city, as Friedmann suggested, on the other this division of labour links the world city to numerous sites of production around the world, where the respective other side of the division of labour is performed.3 Considering that the international division of labour is so central to characterising the world city this link appears to have been studied only insufficiently. Friedmann (1986, p. 73) lists a number of economic sectors as the "global control functions" concentrated in world cities. These include corporate headquarters, international finance, global transport and communications as well as business services, such as advertising, accounting, insurance and law. He also mentions ideological penetration and control as an "important ancillary function", which resonates with Peter Hall's (1979) description of world cities as hosting amongst other important functions the world's largest media networks. Friedmann used a relatively broadly defined group of functions to determine world city status, although he clearly and deliberately privileged economic functions in his analysis as he viewed these as the dominant or most influential 3 I am not suggesting that only one side of the division of labour is performed in world cities, but certain aspects (the control functions) are more prevalent in world cities than elsewhere. 31 ones with the greatest effects on urban life. Sassen (2001) narrowed the focus even more to only a certain set of the producer services (referred to as business services by Friedmann). These include insurance, banking, financial services, real estate, legal services, accounting and professional associations (Sassen, 2001, p.90).  Sassen's rationale for the selection of producer services as key indicators of world city status is fairly straight forward. As production (of both manufacturing and services) has dispersed considerably to multiple sites across the globe, firms need specialised input from producer services to "manage and control global networks of factories, service outlets, and branch offices" (Sassen, 2001, p. 11). These functions with "global control capability" are highly centralised and concentrated in the global cities, as increasingly dispersed, post-Fordist production has a need for centralised control. The producer services became later the primary analytical component of world city research, especially for the analysis of a network of world cities. While the obvious emphasis for both Friedmann and Sassen is the globalisation of urban space and the shift towards globally oriented business functions in world cities, neither considers the nation state obsolete. Sassen (2001, ch. 6) devotes a whole chapter to the changing roles of New York, London and Tokyo within their national economies, especially the changing concentration of producer services in these cities compared to others in their respective countries. She then considers such changes in the global context (ch. 7). Although it has been argued in an early critique of Friedmann's work that "city development occurs in competition with the state" (Korff, 1987, p. 485), Sassen (2000) and others have argued that globalisation is not a zero-sum game, where gains in the global economy necessarily come at the expense of the nation state. Brenner (2004, p. 3) argues in this context that the formation of globally oriented urban economies that try to position themselves optimally within global circuits of capital, are the result of national, regional and local state strategies, i.e. the nation state has a profound interest in positioning its cities as world cities. Although the relationship to all things global is a logical focus in world city research, the relationship to world cities' respective nation states should not be neglected. For example, the strongest influence on London business remains the national economy and London is by no means detached from the rest of the United Kingdom (Gordon, 2006, p. 193). 32 Friedmann (1986) brings forward three additional theses that I will treat as one interrelated theme. First, he suggests that the global control functions in world cities directly result in a dichotomised labour force: highly educated, well paid professionals on one side and a "vast army of low-skilled workers" on the other (p. 73). Second, world cities are the destinations for a large number of domestic and international migrants (p. 75). Third, the evolution of jobs in world cities results in spatial and class polarisation, which also occurs at the regional and global scales. Central to all three of these theses is their relationship to the division of labour and resulting inequalities at various scales. Perhaps most importantly for the analysis of world cities is the urban scale. Friedmann does not provide many details regarding the characteristics of migrants, but distinguishes between international and interregional migrants. Sassen (2001, p. 322) specifically identifies the emergence of a "serving class" in global cities that is often made up of immigrant men and women. Immigrants are 'absorbed' by the global cities to service the strategic sectors, i.e. the producer services and other key elements of the economy. Although immigrants have contributed to world cities by revitalising parts of the city that would otherwise be largely vacant and revitalised specific sectors of the economy, as for example specialised retail, the division of labour into strategic sectors and those servicing the strategic sectors has resulted in highly polarised cities with polarised job markets. Friedmann's observation that polarisation occurs at three different scales is probably not ground breaking, but in the context of world cities it emphasises that similar processes occur at various spatial scales and that the world cities are not disconnected from their peripheries. The ongoing connections of world cities to their peripheries are facilitated by what Sassen (2001, p. 173) calls gateway cities. While there will be a number of leading financial centres around the world, an emerging division of labour between cities means that the leading financial centres of many countries fulfil gateway functions that link national investment opportunities to global circuits of capital. While gateway functions are of fundamental importance for global (and national) capital flows, their primary function will be the linking of global flows to local or national investments rather than the innovation of new financial products. These linking functions tend to be executed by leading producer services through their network of branches and affiliates (Sassen, 2001, p. 173). 33 Long before the world city view became prominent in the social sciences Andre Gunder Frank (1969, pp. 3-17) already identified the same mechanisms of an urban structure in a globally uneven development. The satellite-metropolis structure describes exactly the phenomenon that satellite cities function as gateways to the resources (or investment opportunities) of the periphery and if necessary the satellites to the world metropolis utilise the specialised services of smaller satellites further down the road, thus connecting more peripheral economies to the world cities. A major difference between Frank's and Sassen's analysis, however, is the description of the urban structure at the top of the hierarchy. Frank pays little attention to the globally leading urban centres - which would now be called the world cities - although he mentions the relationship of London to its satellites. Frank does not provide much of an explanation as to why he aggregated all of western Europe and North America into the world metropolis, but one explanation may be the much flatter distribution of wealth, enabled through Fordist production, in these regions as opposed to the capitalist periphery. Friedmann (1986, p. 76) reminds us that regional income gradients in core countries rarely exceed a ratio of 1:3, while the ratio for semi-peripheral countries would more likely be 1:10. Hence, the role of cities as gateways or what we may call linking cities seems more pronounced in the periphery or semi-periphery as supposed to the core.4 The networked character of this system is continuously emphasised by both Sassen and Friedmann and both suggest that the network is strongly hierarchical, but neither analysis provides any data specifically on inter-city relationships. The network view of world cities In addition to Friedmann and Sassen's work, the concept of a 'network society' developed by Manuel Castells (1996) gained significant influence in world city research. Core to Castells's argument is that cities should be viewed not as places, but as processes, i.e. as a network of interlinked places, in which the flow of ideas, goods and people are the central element in shaping a city. Castells (1996) famously used the term space of 4 Compare Amin (1994, 1997) in this context. Amin suggest that the historical compromise between labour and capital since the 1880s and especially pronounced in the Fordist mode of production in North America and Western Europe reduced the income gap within core countries while increases it at a global scale. 34 flows to describe this view of cities, compared to the until then dominant view of cities as bounded spatial entities. Castells identifies three layers of material support for dominant social practices that constitute the space of flows. The first layer is made up of infrastructure, such as communication technologies, especially the internet, and transport networks, such as global airline traffic. The third layer is constituted by the spaces occupied by global elites (technocratic, financial, managerial), such as luxury residential areas and exclusive leisure spaces. Sandwiched between the first and the third layer is the most important second layer, "constituted by agents who use the infrastructure networks to link together specific places to carry out 'well-defined' economic, cultural and political functions" (Taylor, 2004a, p. 26 referencing Castells, 1996, pp. 413-415). Although Castells did not consider his work as a specific contribution to world city research, he does engage with Sassen's Global City. For him, however, research on global cities cannot be carried out as comparative studies, where attribute data for individual cities can determine a ranking of cities, and he postulates that world city research cannot be limited to a few cities at the top of the urban hierarchy (see Taylor, 2004a, pp. 25-27). The view of a network society became increasingly dominant amongst world city researchers and world cities themselves were increasingly analysed from a network perspective. Thus the lack of data on inter-city relationships became a serious concern in the late 1990s (Short, Kim, Kuus and Wells, 1996; Taylor, 1997; Beaverstock, Smith, Taylor, Walker and Lorimer, 2000). Sassen's work The Global City was a valuable comparative study and Friedmann's work remained an agenda for research, but neither analysed the network. A major challenge in the collection of data on inter-city relationships is that socio-economic data are usually collected by states for states and while there is a wealth of data at the city level from census and other statistics, these are attribute data and are not collected in a consistent manner across inter-state boundaries. Neither do these data represent flows. Data on flows, as for example trade, are aggregated to the level of the nation state allowing for an analysis of the connectedness of states but not cities (Taylor, 1997, p. 324). Moreover, with respect to this paucity of data, world city research has been critiqued as privileging cities in core countries of the world-system (Godfrey and Zhou, 1999). One of the main challenges associated with the absence of inter-city data is the reliable establishment of a hierarchy amongst world cities, because attribute data alone 35 cannot show the structure of a hierarchy. Ranking by attribute data only produces an ordered list, but does not provide information on the strength of the relationships between cities and thus does not provide any clues to the structure of the hierarchy (Taylor, 1997, p. 325). Beaverstock, Smith and Taylor (2000, p. 124) point out: “By concentrating on the attributes of world cities and neglecting their relations, we learn a lot about the nodes in the network, but relatively little about the network itself.”  Taylor (1997) therefore refers to hierarchical tendencies amongst world cities rather than a set hierarchy. Given the absence of any agencies collecting data on inter-city relationships, urban researchers had to collect their own data. One of the first methodologies suggested was a newspaper content analysis of the business pages of the leading newspapers in world cities (Taylor, 1997). In this analysis locational references to other cities were counted and the number of counts for specific city pairs (i.e. the 'home' city in which the newspaper originates and the other city mentioned in the paper) was used as an indicator of the strength of the relationship between the cities. The initial study by Taylor (1997) only included a handful of cities within the United States. It was partial, but it was a starting point for collecting inter-city data and considered a methodological test from its inception (see also Beaverstock, Smith, Taylor, Walker and Lorimer, 2000). Although Taylor suggested that the methodology of newspaper content analysis for cross referencing of cities has not been replicated to any substantial degree, there were multiple attempts to establish evidence of a world city hierarchy. Two other methodologies for measuring a world city hierarchy based on relational data, rather than attribute data, were proposed and carried out by Beaverstock, Smith, Taylor, Walker and Lorimer (2000), reflecting their affirmation for a network perspective of world cities based on Castells's space of flows. The first methodology concerns the analysis of migration of highly skilled labour between cities. The main challenge with this method is again that migration data are collected at the inter-state level only. The authors suggest a combination of quantitative and qualitative data to analyse migration flows. Specifically, the authors reviewed the use of postal questionnaires as well as semi- structured interviews (see also Beaverstock, 1990, 1996; Beaverstock and Smith, 1996). The other methodology and one that became prevalent in world city research concerns the office locations of APS firms. The distribution of head offices and branch offices of such global producer services firms are used to suggest strong links between cities, if any 36 of these firms have offices in at least two cities. This method was first applied to construct a hierarchy amongst cities of certain regions (e.g. Lynch and Meyer, 1992 for North America; and Daniels, 1993 for Western Europe). At a global scale this method was first applied by Beaverstock, Taylor and Smith (1999) and gained considerable influence in the research on world cities (e.g. Beaverstock, Smith and Taylor, 2000; Beaverstock, Doel, Hubbard and Taylor, 2002; Taylor, Walker, Catalano and Hoyler, 2002; Derudder, Taylor, Witlox and Catalano, 2003). Given Castells's influential view of cities as processes, the network structure of the world cities has been considered the new metageography of economic globalisation. Although the world cities do not eliminate the role of the state or make the state an obsolete entity, there has been a rescaling of economic power to the nodal structure of cities from the mosaic structure of states (Beaverstock, Smith and Taylor, 2000). The network perspective became dominant in world city research in the early 2000s, especially through the work of GaWC at Loughborough University5, which engaged in a significant effort of data collection on inter-city relationships, most of which are based on the office location of producer services firms. World city research reached a preliminary climax in its theoretical development with the conceptualisation of the world city network as an interlocking network (Taylor, 2001). The concept of an interlocking network is based on the role of APS firms as providing the command and control functions in the global economy, as discussed above. This conceptualisation divides the world city network into three levels: 1) the network level, this is the world economy in which the network is placed; 2) the nodal level, which is made up of the cities as multiple centres in the network; and 3) the service firms, which are the agents forming the network. This triple configuration makes the world city network an unusual network. Most networks operate on two levels only (Taylor, 2001, p. 182). The conceptualisation of the world city network as an interlocking network with three levels is based on Burt's (1983) study of overlapping directorships in the United States, where the US economy constitutes one level (the network level), firms with their boards the second level (nodal level) and directors that sit on two or more boards constitute the third level (the agents). However, in Burt's conception the middle level is 5 For more details see: 37 the most influential one as firms, through the decision making of their boards, are the main actors. In the world city network, the sub-nodal level, i.e. the service firm, is the prime actor (Taylor, 2001, p. 184). Firms can dismiss board members, but cities cannot dismiss firms (other than perhaps with the aid of their respective nation states and only at potentially high costs to a city's locational value). Hence, the world city network is articulated by global services firms that as a locational strategy chose multiple cities as their nodes for their operation and thus create the links within the network. A city's position is strengthened by the strength of its links, rather than the pure size of decision making capacity, as measured through attribute data such as stock market size or the number of headquarters. Over the three decades since the publication of Peter Hall's (1979) book, which focused on a variety of functions in world cities including the manufacturing sector, media and political power, the focus became the service industry as it had become the major growth sector in the 'advanced' economies (Beaverstock, Smith and Taylor, 2000, p. 123). Thus producer services (accountancy, advertising, banking/finance and law) became the focus to assess a city's role and power position within the network of world cities, as these services give firms “global competence” and are a sign of the “global capacity” of cities. “Global capacity is then defined empirically in terms of aggregate scores and interpreted theoretically as concentrations of expertise and knowledge” (Beaverstock, Taylor and Smith, 1999, p. 446). However, assessing the global capacity of cities in this way, which resulted in the ranking of cities into alpha, beta, gamma and delta cities, assumes "global competence" of producer services firms, which is the sole indication of strength in the network. It still says very little about the network itself. In order to measure inter-city relations, the GaWC group constructed a data matrix that included 316 cities cross tabulated with 100 firms. Again these firms were producer services firms, but tabulating their office locations in a matrix with cities allowed for the conceptualisation of the world city network as an 'interlocking network'. The service firms are the network articulators and a firm's presence in a city indicates a city's integration into the network. In addition, size and extra-territorial functions of an office were ranked between 0 and 5, whereas 0 indicates no presence of a firm in a city and 5 the presence of the headquarters. Scores between 1 and 4 were allocated depending on an office's importance and extra-territorial functions (Taylor, Catalano and Walker, 38 2002; Taylor, 2004a). This matrix allows the construction of two specific rankings of cities. The first one measures a city's rank by the size of cities as service nodes, i.e. how important is the city in terms of what is located in it. The score for this ranking is derived from adding the values for each firm in a city. The second ranking is based on the "situational status of a city within the network" (Taylor, Catalano and Walker, 2002, p. 2371), where the level of connectivity is measured as a proportion of all interlocking connections between cities expressed by the presence of offices of the 100 firms analysed. The results of these two rankings are similar with London, New York and Hong Kong taking up the top three positions in both. However, there are some examples where cities take different positions in the two rankings (Taylor, Catalano and Walker, 2002, pp. p. 2371-2372; for a detailed explanation of the measurement methodology see also Taylor, 2001). More importantly than the actual ranks of cities may be that in this matrix of 316 cities by 100 firms even the top two ranked cities (London and New York with almost the same score) each account for only slightly more than 1.5 percent, meaning that connectivity is by no means highly concentrated, but rather diffused across a large number of cities. Would the number of firms and by extension the number of cities be increased, the proportions of the top ranked cities would likely decrease even more. It has often been argued that the world city network is not strictly hierarchically organised, rather that there are hierarchical tendencies in the network (Taylor, 1997; Taylor, Hoyler, Walker & Szegner, 2001; Taylor, Walker, Catalano & Hoyler, 2002; Derudder, Taylor, Witlox & Catalano, 2003). Based on analysing seven different angles of measuring power differentials amongst world cities, Taylor, Walker, Catalano and Hoyler (2002) conclude that there are clearly hierarchical tendencies within the world city network, but power differentials are "not easily discerned" (p. 239). The seven angles of analysing the network are global network connectivity, banking/finance connectivity, dominant centres, global command centres, regional command centres, high connectivity gateways, and gateways to emerging markets. While identifying 'command and control' centres is relatively easy and transparent, this does not mean that this power should be analytically privileged over the "more opaque network power" (p. 239). This means of course that there are many different ways of participating in the world city network with specific roles and niches for many different places (for a review of the literature examined in this section see also Derudder, 2008; Lai, 2009). 39 The producer services and other network articulators The construction of evidence regarding an interlocking world city network has been based almost entirely on data of APS firms. Although this choice is based on a plausible explanation, the empirical evidence for this explanation appears rather thin. Sassen (2001) makes the point that the producer services are essential to organising and controlling the increasingly globally dispersed system of production and therefore these services are the global control and command functions of the global economy. There has certainly been a shift in importance towards services, measured as a proportion of world trade (Daniels, 1993) and although services were traditionally considered location bound, because they require face-to-face contact, they have increasingly become a tradeable and international commodity (Taylor, 2004a, p. 130). Sassen (2001) views the global cities as command and control centres, because they host a large proportion of the producer services. However, her emphasis is not on power, but production. The production of advanced producer services is in Sassen's view the core ingredient to exercising global control. Hence, being the location for this production gives a place a powerful position. By concentrating on the production of producer services, Sassen seeks to divert the attention away from the power of large corporations towards service providers and by extension specific cities. Taylor (2000, pp. 7-10) supports this point of view, but he goes one step further and views producer services as a new production monopoly. He reconsiders the role of producer services in light of what he calls “Braudel's provocation”, arguing that the largest profits are not made under conditions of competitive markets, but under conditions of monopoly, and - following the kind of world-systems analyses outlined in chapter 1 - that capitalism is distinctly different from and in direct opposition to a market economy. Accordingly producer services are not just an input into the economy to make capitalist processes work more efficiently, but are new knowledge intensive products that operate under monopolistic conditions and thus reap large profits. Moreover, Taylor, building on Braudel, suggests that the most successful capitalists are those that are flexible and not bound by location, since these are conditions conducive to seeking new monopolistic 40 products. This locational flexibility is easily attained with financial capital, which depends on efficient producer services. While the argument for using producer services as the primary articulator of the world city network remains plausible, it leaves out many other sectors that produce under monopolistic conditions and are therefore potential network articulators in their own rights. These may not have been vitally important in the re-organisation of the global economy from Fordist production methods to more dispersed forms of production, but sectors like software production, pharmaceuticals and biotech research or even agriculture and natural resources, which are at least partly being produced under monopoly conditions, remain elementary ingredients to the functioning of the global economy and capitalist expansion. While there is no sustained research on the urban control over other monopolies, other than through the proxy of the producer services, there is some research on an interlocking network of world cities articulated through political, cultural and civil society arenas (Taylor, 2004b, 2005a, 2005b). Although cities and states are sometimes viewed as dualistic concepts in contemporary globalisation, this research has taken a more nuanced view of this relationship and has not ignored state power, the power of civil society or cultural influences. In an attempt to map out global civil society in the world city network, Taylor (2004a) applied an interlocking network model for 74 global non-governmental organisations (NGOs) across 178 cities. In this model Nairobi is the most connected city followed by Brussels, Bangkok, London and New Delhi. Taylor (2004a, p. 273) views this potentially as an emancipatory global agenda by the NGOs, which runs counter to corporate globalisation as well as other geographies of NGOs6. In an attempt to map out the leading world cities articulated through multiple networks, Taylor (2005b) considers four types of globalisation - economic, cultural, political and social - to arrive at a complex taxonomy. This taxonomy identifies New York and London as the leading duo of global cities, while others are categorised as global cities with "smaller contributions and with cultural bias", "incipient global cities", "global niche cities - specialised global contributions" with a focus on economic or political and social. Beyond these global 6 Taylor refers to the Global Civil Society Yearbook, which concluded that NGOs are highly concentrated in north-western Europe. 41 cities, a category of world cities is divided into "subnet articulator cities" within specific globalisation arenas as well as "world-wide leading cities" with primarily economic or non-economic global contributions (see Table 12 in Taylor, 2005b, p. 1606 for details). Taylor differentiates between global and world cities in this taxonomy, with global cities representing the top echelons of the world city network and being comprehensive in making global contributions across multiple arenas, while world cities still have world- wide influence, but are more specialised in a specific arena of globalisation (Taylor, 2005b, p. 1605). The multiple globalisation arena approach is important, because it recognises various arenas of globalisation and goes far beyond the focus on producer services or even economic aspects more generally. The result is a much more complex, but perhaps also messy, ordering that opens up many new questions about the influence of specific spheres or arenas of globalisation. In yet another study that analyses the world city network outside the sphere of influence of the producer services, Taylor (2005a) examines global civil society and global governance as network articulators. Using an interlocking network approach, he analyses three world city networks: 1. inter-state city network with state departments (diplomatic representations) as network makers; 2. supra-state city network with UN agencies as network makers; and 3. trans-state city network with NGOs as network makers. The three networks examined show quite different spatial organisations, from intra-regional and horizontal (inter-state network) to inter-regional and hierarchical (supra-state) and inter-zonal (across the North-South divide, trans-state). While networks of political power, civil society and cultural and social influences provide an intriguing perspective of the world city network with multiple arenas of power and influence, there is little concrete evidence of what degree of power and influence over global processes these specific arenas or network articulators (i.e. actors like NGOs, UN agencies, state departments, cultural organisations etc.) exercise. Friedmann (1986) deliberately privileged the economic sphere in his initial conception of world cities and subsequent work on the world city network also emphasised this sphere and control over production. The focus on producer services, i.e. those services that are needed to produce other goods and services efficiently, seems logical, but perhaps too narrowly focused, because corporations need states to protect production (see Wallerstein, 2004, p. 48). It does not reflect a political-economy approach to research on globalisation nor the 42 "cultural turn" in economic geography (Crang, 1997; Barnes, 2001). World city research has recently been critiqued for focusing too narrowly on the economic sphere and specifically the producer services. Cities and globalisation: multiple perspectives on cities in the world In the late 1990s a number of studies appeared that referred to 'globalising cities' rather than 'global cities' in response to the increasingly dominant global city paradigm (see eg. Öncü & Weyland, 1997; Afshar, 1998; Yeoh, 1999; Marcuse & van Kempen, 2000). Marcuse and van Kempen (2000, p. xvii) state in the preface to their book Globalizing cities: a new spatial order? that they chose the term globalising deliberately over global, as they view globalisation as “a process, not a state, and a process that affects all cities in the world, if to varying degrees and in varying ways, not only those at the top of the 'global hierarchy'.” The concept of the world city or global city is thus critiqued for its relatively static view of the world urban network and lack of breadth. Early world city research set out to understand the city in the current round of increasing globalisation, i.e. in a time of increasing density of international and global interactions compared to the increase of such interactions at the national or sub-national scale (Urry, 2003, p. 4). Friedmann and Wolff (1982) pointed to a number of forces that they viewed as connecting cities with each other, amongst them footloose professionals. Similarly, Hannerz (1996) identifies four human types of inter-city flows: 1. transnational business people; 2. low-wage immigrants; 3. cultural professionals; and 4. tourists. Smith and Timberlake (1995a) identified three processes for studying world cities, namely human, material and information processes, which occur in four different spheres, economic, political, cultural and social. Thus they suggested 12 areas for empirical research, which in practice, however, proved difficult to study, because of a persistent lack of data (Derudder, 2008, p.560). Considering these types of processes in more detail would increase the breadth of indicators on which current world city research relies. Moreover, it would provide opportunities to go much further to explore yet other networks. Perhaps more importantly though, these processes include flows of human actors rather than firms, organisations 43 and aggregate numbers of air traffic volumes or even cities themselves. Marcuse and van Kempen point to this problem, when they write that they view globalisation as a process, not a state. Current world city research is relatively static or deterministic, often a snapshot of a point in time. This is not suitable for analysing highly dynamic global flows. Marcuse (2005) argues that cities cannot be treated like actors themselves in the process of globalisation, neither are they some kind of component of it (see also Norlund, 2004). Taylor (2001, 2004c) is very clear that he does not treat cities as agents, but that in his analysis the global services firms are the agents. However, perhaps even these firms should not be considered agents. Cities are inhabited by people and people are employed by firms. The people are actors and have agency in the process of globalisation. The analysis of cities or firms becomes relevant, because certain actors have more agency or influence than others and analysing firms and cities as places to group these actors is an important tool to identify loci of power. However, cities or firms themselves do not have agency as such, only through the collective action of certain groups of people within these entities. Michael Peter Smith (2001, p.10) argues that the work of some influential writers such as John Friedmann, Saskia Sassen, David Harvey or Manuel Castells constructs capitalism as an "omnipresent" force to which cities are inescapably subjected. He argues that such an analysis relegates any other societal forces to the local and conceptualises these as helpless spaces that are all subject to the grand force of capitalism. He argues for a reconstruction of urban theory that does not privilege class or capital or capital as analytical dimensions. He argues that global city theories focus too much on the effect of global processes "from above" on urban processes while neglecting myriad other transnational processes (Smith, 2001, pp. 46-47). Moreover, he argues (Smith, 1998, p. 485) that the "global city is best thought of as a social construct, not as a place or an object consisting of essential properties that can be readily measured outside the process of making meaning". His critique is primarily aimed at the methodology that possession of certain attributes (e.g. headquarters of transnational firms or offices of advanced producer services firms) determines a city's global status. Taylor's work on the interlocking network (e.g. Taylor, 2001), which appeared after Smith's critique, stresses that it is the connections between cities that are being measured, i.e. the strength of the 44 network. Nonetheless, subsequent world city research has been equally critiqued for being overly deterministic (e.g. Robinson, 2006). The debate between 'global' and 'globalising' cities then becomes one of a binary view of structure and agency. While global cities identify structures, globalising cities are analysed by examining agency. However, Glassman (2003a, p. 266) argues that this view of a binary between structure and agency is irrelevant, if structure is merely considered the agency of large groups of people. This view is explicitly Marxist and refutes conceptualisations of structure by authors such as Anthony Giddens. The main issue then is "whether, where and how some subset of a larger collectivity can gain enough support in its actions to substantially alter relatively long-standing features of the social relations that constitute 'the structure'" (Glassman, 2003a, p. 266). One strength of the more structural approach to studying cities in the current round of globalisation is the identification of uneven development at a larger scale. The grouping of people as agents acting within a firm or a specific city based on certain common characteristics can then be helpful in identifying structures. A strength of the approach focusing on the agency of individuals or small groups and networks is its ability to engage in multiple scales at the same time, but is unlikely to provide an overview of larger inequalities. It is thus imperative for the kind of world city research suggested here to consider both the structural power and unevenness within the world city system as well as the agency of individuals acting within these structures. Given the focus on only a small number of select cities, primarily in western Europe and North America, but increasingly also in East and South-East Asia, another major thrust the critique of world city research takes is a charge of ethnocentrism in the approach to ranking world cities (e.g., Larner and Le Heron, 2002, p. 415). The increased interest in East and South-East Asia is, however, not matched by an interest in South America or Africa and other world regions, including South Asia, the Arab world, Central America and the Caribbean (see also Davis, 2005, p. 100). Furthermore, geographic concentration is not the only critique that points to limits in the world city paradigm. The limited focus on economic processes has been considered exclusionary by many observers. Jenny Robinson (2006, chapter 4) formulates two specific critiques of world city research. First, the methods applied in narratives and network analyses of global cities are 45 limited to “a relatively small range of economic processes with a certain 'global' reach” (p. 96), i.e., the producer services identified as 'command and control' functions by Sassen (2001) and applied widely by others. Second, the research organises cities in a hierarchical order and thus it “implicitly establish[es] some cities as exemplar and others as imitators” (p. 94). Robinson (2006, chapter 3) also examines the work of the Manchester School of urban anthropologists that conducted their field work in the Zambian Copperbelt in the 1950s, where urbanisation was shaped to a large degree through the immigration of workers from rural areas of Zambia. The Manchester School found a specific form of modernity shaping urban life in the Copperbelt. This was partly influenced by rural traditions and partly by new realities of social interactions amongst various ethnic groups (Robinson, 2006, p. 48). The Manchester School also pointed to the shortcomings of the Chicago School, which was highly influential in urban studies, and noted that Wirth's (a Chicago School researcher) universal theory of urbanism was far from universal (p. 46). Robinson's main point of critique remains that urban studies, including world city research, remains primarily shaped by the western experience and that the experiences of cities in the Global South surface at best as case studies or add-ons to established theory. Robinson concludes the chapter on 'comparative urbanism' in her book Ordinary Cities with a call to open up the theoretical lenses of urban studies (2006, p. 63): A post-colonial urbanism will be open to theory travelling in any direction: from Chicago to the Copperbelt,  but  also back again.  Of course like all  travel  in a globalising world, there are obstacles to overcome, historical ties and entrenched power relations that shape the directions of influence, and limited channels to follow (Tsing 2000). And there is no a-priori reason why ideas and writing in one context have to make sense in another. But the opportunity is there, to establish conversations  because  of,  and  across,  strong  empirical  and  theoretical differences (see for example Oliviera 1996). As a tactic for post-colonial critique, a form of comparative studies that rests not on a-priori categorical and structural similarities within groups of cities but on a cosmopolitan and curious theoretical endeavour,  inclusive  of  all  kinds  of  cities,  might  stimulate  and  transform the divided form of urban studies. While entirely sympathetic to Robinson's approach to developing urban theory, it is important to not only “travel” across the globe to develop urban theory, but to trace the connections between places,  without - as Robinson suggests - a-priori assumptions on what cities should be like. The ability to trace connections between places is - of course - 46 one of the quintessential strengths of contemporary world city research. And, as Jim Glassman (2003b, p. 691), revisiting Andre Gunder Frank, argues, “the relations between places are crucial to the ways in which we understand the significance of unevenness”. Tracing relations allows us to examine a place's positionality, which will influence people's life in a given place (see Sheppard, 2002). Comparative studies, which compare two separate entities, are not the most suitable form of research for tracing connections. World city research builds on the earlier work of world-systems theory, according to which countries are divided into core, periphery or semi-periphery type economies and can both move 'upward' or 'downward' within the hierarchy. Similarly, cities in the world city system would be able to move within their own hierarchy. The strength of these analyses is their ability to provide an overview of globally uneven development, in the case of world city research at the urban level. The analytical risk of this work, however, is the implicit assumption that in order to move up the hierarchy resource poorer places need to 'develop' in the same way as richer places do, resulting in a mimicking of places 'higher up' in the urban hierarchy. Although dependency theory and by extension world- systems theory and world city research have in their inception rejected such notions of progressive, linear development within both capitalist or orthodox Marxist paradigms, they are often interpreted as suggesting that the experience of the western industrial core or increasingly often that of world cities is the only possible path to development (see Cowen and Shenton, 1996 for an overview of development doctrines). This interpretation remains problematic as it undermines the main theoretical advances the core-periphery models have made in analysing uneven global relationships. Concepts of delinking and independent development paths in fact suggest that a different or autonomous development needs to take place in cities and countries of the global south. Although there have been a number studies that specifically examine or consider non-western cities and those in the global south (e.g. Grant and Nijman, 2002; Gugler, 2004; van der Merwe, 2004; Amen, Aecher, & Bosman, 2006) and some interlocking world city networks of social, political or cultural domain have identified cities in the global south to be highly connected and at the top of the urban hierarchy in their specific domain (see Taylor, 2004b, 2005a, 2005b), there is a further critique that world city research and urban studies more generally ignore the urbanisation experience in the global south (Robinson, 2002, 2004, 2006, McFarlane, 2008). One of the main charges of 47 this critique is that empirical evidence from cities in the global south has not had any influence on formulating urban theory and that, if cities in the global south are considered, these analyses are often nested in theoretical underpinnings derived from examples of western cities (Robinson, 2006; McFarlane, 2008). McFarlane (2008, p. 341) advocates for learning across the north-south divide to make it more global and more situated in its claims. He argues that there is a problem with measuring southern cities "in a particular relation to Western counterparts and theoretical constructs". World city research, for example, privileges a certain type of economic development, while ignoring other often more important sectors. While Taylor's theoretical framework of an interlocking world city network shows no signs of directly building on empirical evidence from cities in the global south, several studies still apply this framework and associated methodologies to cities in the global south (Gugler, 2004; van der Merwe, 2004; Amen, Aecher, & Bosman, 2006). On the other hand, the interlocking world city network is a theory of a global network, rather than of cities themselves. However, the identification of the producer services as performing 'command and control' functions and using these services as an elementary basis for constructing a world city network and identifying hierarchical tendencies, gives world city research a rather narrow focus. McFarlane (2010) advocates for comparative research as both a methodology in urban studies as well as a "mode of thought that informs how we construct knowledge and theory of the urban" (p. 726). There is a danger of overgeneralising on the basis of one or a few examples and of using just one particular type of space or understanding of time (see also Amin and Graham, 1997). Learning about cities, McFarlane (2010, p. 738) argues, should include learning through comparisons of different contexts. While a broad theoretical base for developing urban theory is certainly a welcome development that should enhance learning about cities, it is easy to overemphasise place as a spatial domain in comparative research and lose sight of other domains, especially positionality but also territory, scale and network (see chapter one). Following Jessop, Brenner and Jones (2008) I will propose a methodology for studying world cities that does not privilege territory, place, scale or network and pays profound attention to positionality. 48 Complex relations: where to start world city research? Network studies have enjoyed popularity in geography, sociology and anthropology for some time and several authors have suggested considering the city (Mitchell, 1987, p. 310 building on Craven and Wellman, 1973) or global relations (Hannerz, 1992; Castells 1996, 2000a, 2000b) as a network of networks. I want to employ this notion of a network of networks to argue for a 'scale jumping' (see Katz, 2001, p. 2016) and place/territory inclusive approach to world city research. Currently, constructions of the world city network produce a single meta-network that is spanning the globe, but by no means includes all parts of the globe. Rather it focuses on a small number of select places, which make for rather small points on a map (even if they are large city regions and don’t focus on inner cities), leaving vast territories in between these places not considered. Graham and Marvin (2001, p. 201) discuss a new system of infrastructure provision that does not provide more or less homogeneous and standardised services to the population in a territory. Instead, there is now a system of unbundled infrastructure, that links certain points to each other over long distances, while leaving territory in between unserviced. They refer to this phenomenon as a system of 'hubs and spokes' that creates 'tunnel effects'. With reference to Castells (1996, pp. 147, 380), who suggests that in extreme cases people in these in-between territories will be "reduced to devalued labour", Graham and Marvin (2001, p. 307) write that “the infrastructure networks that support distant linkages, while always local and always embedded in space and place, may actually provide 'tunnel effects' which bring valued spaces and places closer 'together' whilst simultaneously pushing physically adjacent areas further 'apart'”. Ferguson (2006, p. 42) makes a similar point with reference to places of resource extraction and conservation areas in Africa, which increasingly resemble fortresses that are cordoned off from their surroundings and connected (sometimes by helicopter) to larger urban centres rather than their adjacent region. Ferguson considers the term space of flows a “peculiarly poor metaphor” (p. 47), because a flow would indicate some kind of a stream that would flow through a territory and connect places along its way, while the space of flows (as coined by Castells) provide a point-to-point geography. 49 A somewhat different perspective is provided by Anthony King (1990, pp. 6-7), taking a historical route to analysing urban development. While the influence of the colonisers on colonial cities is well known, he argues that “[i]t was from among this colonial urban system [cities, especially externally oriented port cities, in the colonies] that a majority of today's 'world cities' were to develop”. Thus development in the cities of the west or the metropolitan core cannot be understood outside the context of the colonial periphery. Similarly, Ulf Hannerz (1992, pp. 50-51), writing from a cultural perspective, examines the human flows of managerial elites, poor migrant workers, cultural professionals as well as tourists and finds that these flows create chains between 'world cities' and more peripheral areas that work in both directions, to and from world cities. He mentions the latest “ethnic cuisine” originating not in the world city or “Third World music” becoming first “ethnopop” and then “world music”. He summarises (p. 51): “The world city is a world city not by being the origin of all things but as much by being a cultural switchboard through which the more or less peripheral cultures can reach one another”. There is clearly a discrepancy between the cultural or historic perspective provided by King and Hannerz and the more economic perspective provided by Ferguson and Graham and Marvin. Both perspectives have validity in their own rights, but should probably be fused to arrive at a more comprehensively informed urban geography or more particularly a more comprehensively informed world city research. The economic perspective demonstrates how capital (or the agents directing capital) seek out the places providing the greatest return without any or only marginal attention paid to the spaces in between or as Castells would put it to “devalued labour”. The cultural and to a lesser extent the historical perspective to the contrary demonstrate ways, in which the people in the places between the points valued by capital push into, integrate and most importantly influence these places, characterised as the world cities. Anthony King argues that “today all cities are 'world cities'”. Given the myriad connections, his point is a logical conclusion, but perhaps one can go even one step further and argue that all places are world cities or at least all places contribute to the world cities. Thus world city research could begin at any point, anywhere, any time and only needs to trace the network of networks from there. 50 While the analysis of a world city system as a hub and spoke system with tunnel effects or a global meta-network is analytically helpful in understanding the logic of capital flows, our analyses do not need to do what capital does, i.e. devalue the territories in between. Moreover, there are of course myriad economic connections from the headquarters in world and other large cities to the peripheral places in between, ranging from export processing zones to mineral extracting mines, the fields of modern agriculture and village production systems. In other words, the important contributions world city research has made in terms of charting the urban connections of the global economy should be much expanded beyond its current scale and sphere. Some of Saskia Sassen's work, including The Global City (2001), stresses the role of labour in the formation of global cities, but this aspect has somewhat been lost in the focus on producer services and a reconsideration of labour would certainly enrich current world city research (see also Samers, 2002). In order to arrive at a more comprehensive world city research, connections need to be traced across scales. There are always connections between places, even if these do not consist of capital investment and may be considered weak. The lorry driver transporting the goods from the mines to the ports stopping for lunch in a village, the community nurse paid by an international NGO or the local store owner importing manufactured goods from far away all provide connections and link places. They may not link them directly into the network of world cities, but through the many networks of networks most places are integrated into the processes of world cities, even though many connections may also by-pass the small number of world cities at the top of the hierarchy. Axford (2005), building on Hannerz (1996), argues for a network approach to globalisation, which can cut across intra-, inter- and transorganisational levels as well as “conventional levels of analysis to link different personal and institutional domains" (Axford, 1995, pp. 78-82 cited in Axford, 2005, p. 195). Urry (2003, p. 122) goes as far as arguing that there should not be such notions as “agency”, “micro” or “macro” levels, “system-world” or “life-world”, as all these presume that there are “entities with separate and distinct essences that are brought into external juxtaposition with each other”. Instead research on globalisation should focus on connections outside any of these predefined realms (cf. discussion of Urry in Axford, 2005, p. 189). This notion of studying connections “rests upon a profound 'relationality'” and is closely related to actor-network 51 theory and post-structuralism (see Urry, 2003, p. 122; see Dicken, Kelly, Olds and Yeung, 2001; Murdoch, 1997a, 1997b, 1998 for a discussion of actor-network theory). Richard Smith (2003a, 2003b, 2007) considers ideas from poststructuralism, actor-network theory, nonrepresentational theory and complexity theory to produce a spatiotemporal pattern or topology that can account for the sheer endlessly interconnected world that produces contemporary world cities. According to actor-network theory networks are constantly produced and reproduced by human and non-human actants. Non-human actants, e.g. telecommunication devices or goods, are shaping networks as agency is attached to these goods and technologies. Considering the example above, agency would be attached to the lorry transporting goods from the mine to the port, the medication and equipment of the nurse and the groceries from some far away place in the store. Hence, everything is made up of a complex set of relationships and so are the world cities (see Smith, 2003a, pp. 35-36). Moreover, because everything is intrinsically interconnected, actor-network theorists view the concepts of local and global as obsolete, because one is only constructed in relation to the other and certain assumptions as to what we think should be a local process and what should be a global process (Smith, 2003a, p. 35). The critique of this constructed dualism is extended by Smith (2003a, pp. 33-34) to Castells's concept of the space of flows, where capitalism is constructed as the overarching powerful realm and actors are trapped in the local. Using the example of a CEO, who would be powerless, if all his assisting devices and human assistants would be stripped away from him, Smith (2007) continues to argue that power is not a top-down relationship (in fact he argues there is no such thing as scale), but an asymmetrical set of relationships. He defends poststructuralists and actor network theorists, including Foucault, Deleuze, Guattari and Latour, against the critique that they were concentrating on details at the local scale to the detriment of the global and therefore the real sources for global inequality. He argues that it is "highly disempowering to shift to the big picture [...] rather than follow a network, precisely because the political point is to question the relations that produce asymmetries" (Smith, 2007, p. 266). Poststructuralism and actor-network theory transform the research on power and thus question the common assumption of world city research that agglomerations of producer services or headquarters means a city has power and is in command and control. Expanding this theoretical framework to the work of Badiou, 52 Smith and Doel (2011, p. 35) conclude that the world city is by no means in control (in spite of all the producer services): "As it turns out, what is transforming the world’s economy is not the 'global city' (it is certainly not in 'strategic control'), but rather the 'sub-prime' or NINJA (No Income, No Job or Asset) American 'home owner' who, whilst belonging to the world's financial networks, was not included in those networks until it was too late." Toward vertical world city research: a renewed world-systems approach There seems to be an inherent contradiction between world city research applying interlocking network approaches and poststructuralist views of global connections and complexity. It is not my aim to create a false dualism between the two, but to bridge the approaches. They certainly have more in common than might appear at first sight. First, it is important to remember that dependency theory and world-systems theory rejected any notions of predetermined development having to take place in the so called 'underdeveloped' countries. Yet, contemporary world city research is critiqued for establishing a set of characteristics for model cities that others need to follow if they want the status of world city (Robinson, 2006, p. 94). Poststructuralist approaches to studying globalisation equally reject notions of a pre-determined path to development. I will propose an approach here that can bridge the two by re-introducing a world-systems perspective into world city research. Second, I do not view structure and agency as contradictory concepts (see Glassman, 2003a). I am therefore applying a methodology that examines large scale financial structures as well as the agency of individuals acting within these structures. Collective agency (i.e. the actions of groups of people) constitutes structural power. I will therefore treat the poststructuralist critique that the works of Harvey (1989a), Castells (1996) and more generally world city research (e.g. Sassen, 2001; Taylor, 2004) overemphasise capital as an agent of social change with some caution. Capital is certainly not the only agent of social change, but the dependency on capital investments for countries, regions, cities and their inhabitants makes it a dominant one in contemporary capitalism. However, if capital is a primary social agent we need to 53 consider capital itself, not just its financialised forms and the functions that presumably control capital. I therefore want to start my analysis of world cities at the bottom of the investment chain, the places of bricks and mortar investments. There is much value in the myriad contributions that world city research has made to the field of urban studies and globalisation. It has engaged in an unprecedented data collection specifically for cities, which could shine considerable light on connections and flows between cities. World city research as a research agenda was proposed to analyse the effects that increasingly transnational flows of capital had on these interconnected world cities (Friedmann and Wolff, 1982; Friedmann, 1986). It was not intended to establish some cities as exemplars that others should imitate. World city research is also specifically concerned with the role of labour and the relationship between cities and states (Sassen, 2001). Moreover, the analysis of world cities as an interlocking network, held together primarily by management functions or the APS, is unprecedented in scope. However, the almost exclusive focus on the APS in contemporary world city research is highly problematic and limiting in scope, which has - contrary to its intentions - limited the theoretical development rather than enhanced it. I therefore suggest to considerably broaden the scope of world city research and to trace the connections not only between cities that are included in the world city rankings under current criteria, but also the multiplicity of other places that connect, feed into or by-pass the world cities. This also means that the original theoretical underpinnings of world-systems theory in the network of world cities (see Meyer, 2003 for an overview; Smith and Timberlake, 1995b) needs to be revisited. Current world city research is global, but it is selectively global and by no means global enough. It focuses, as Jenny Robinson points out, on the top few dozens of cities in the world city network. What we have never achieved in world city research is to provide a perspective that cuts across scales as world-systems theory intends, although earlier works started to make important contributions in this respect (e.g. Armstrong and McGee, 1985); for the most part world city research has stopped in the semi-periphery or focused on the core cities in the world-system. In order to undertake a more broadly defined world city research that traces the connections to the peripheries and hinterlands, we need to move away from assumptions of unilateral power relations of command and control functions toward a theory of 54 interdependent power relations in world city research (see e.g. Smith and Doel, 2011). This means the focus of world city research can no longer be solely on the producer services. The producer services and especially the financial sector can hardly be considered in control, as some of the most powerful institutions, including New York investment banks and insurance giants such as AIG, would have met their Waterloo just as Lehman Brothers did in September 2008 without the intervention of national governments. In fact, the crisis would have likely spread even further and affected more banks and other financial institutions, had governments mainly in the west not rolled-out an unprecedented rescue package of trillions of dollars to maintain the current structure of financialisation. Had the state not intervened, capitalism as a layer above the market economy would have been considerably weakened or ceased to exist. So far, world city research has neither captured the link between economy and politics well, nor the powerful combination that the two arenas occupy. Producer services can hardly be viewed as in control. A broader, more global world city research agenda needs to reconsider its origins in world-systems theory. Giovanni Arrighi (1994, pp. 4-5), discussing Braudel, revisits Marx's formula of capital: MCM'. In this formula M = money capital, i.e. liquidity, flexibility and freedom of choice; C = commodity capital, i.e. capital invested in a particular venture with a specific input-output combination and a view to make a profit; M' = expanded money capital, i.e. expanded liquidity, flexibility and freedom of choice. Braudel states that one of the essential features of capitalism is its unlimited flexibility and that those capitalists that do not confine themselves to a single choice of investment and remain adaptable and non-specialised are the most successful ones. The choice for emphasising financial capital in particular and producer services more broadly in world city research thus seems to be a logical one. However, when focusing on the producer services as a proxy for money capital, the analysis ignores an elementary step in between, i.e. commodity capital. Simply put, in spite of all the financial capital and re-packaged securities, somebody has to produce something at some point. And, most firms listed at stock exchanges around the world do that. This means world city research only represents M and M' in Marx's formula. Leaving out the C (commodity capital) between M (money capital) and M' (expanded money capital) in world city research means leaving out a substantial part of the world economy, and thus a substantial part of what shapes world 55 cities. Hence, if the financial capital and producer services have considerable leverage in commanding capital in the world economy, there needs to be more empirical investigation into what kinds of capital are commanded and how this is happening in relation to a world city network. The recent publication of a special issue in Global Networks is starting to explicitly address this empirical vacuum by suggesting greater convergence between world city research and works on global commodity chains (e.g. Derudder and Witlox, 2010; E. Brown et al, 2010). There can be little doubt that capital is being channelled through the world's financial centres as these host the institutions that facilitate these flows. However, it is doubtful that world cities are unilaterally in control and act as command and control centres. Rather they are switchboards (see Hannerz, 1992; Smith, 2007), i.e. a central node in a global network made up of networks of networks. If the goal of world city research is to investigate the spatial organisation of world order (or disorder) through an urban network, we need to first identify the ordering principles of this world (dis)order. Following the world-systems perspective that capitalism is inherently anti-market and dependent on monopolies controlled by core countries, I suggest to investigate urban networks through functions controlling monopolistic processes, without an apriori assumption of certain places being in control. As a starting point five networks that organise the five monopolies identified by Amin (1997, pp. 3-5) could be constructed: - The places directing the technological monopoly (likely places with high research activity protected by patents and copyright) - The places hosting the monopoly of worldwide financial markets (likely the places with high concentrations of banks and stock exchanges protected by the bail-outs and guarantees from nation states, including cities currently most often identified as world cities) - The places in charge of monopolistic access to the planet's natural resources (likely a combination of places hosting the offices and owners of large extractive industry conglomerates and national governments in charge of sub- surface rights) 56 - The places creating the content for the media and communication monopolies (likely the places hosting large publishers, entertainment industries and news agencies) - The places with decision making powers on monopolies over weapons of mass destruction (likely a combination of national capitals, army headquarters, supra-national organisations and weapons manufacturers) Organising world city research around these five monopolies is an attempt to apply a broader base to world city research compared to contemporary approaches primarily focused on producer services. Of course, these five monopolies only provide an ordering principle for research and cannot be viewed as an ultimate lens through which to view all things global. Moreover, it will be futile to consider any one of these in isolation as these monopolies interconnect and build on each other. The monopoly on natural resources will need to cooperate with the financial sector in order to attract investors and monopolies of weapons of mass destruction need the support from the world media to maintain public support. The connections, overlaps and interactions can be endless, but as a starting point to a broader world city research, these five monopolies will be sufficient. A world city research agenda organised around these five monopolies and perhaps other monopolies as organising principles remains open to new avenues of inquiry. While contemporary world city research has constructed networks (or earlier rankings) of cities at the top of the system, I suggest starting the investigation with the places where these monopolies play out, where the effects of them are felt. This means, I do not propose the construction of another horizontal network, but the tracing of vertical connections in world city research. A further challenge to this type of world city research is the availability of data. The challenges in getting adequate data at levels below the nation states has been sufficiently documented (e.g. Taylor, 1997) and have remained for a long time. However, while the availability of data will be a challenge, designing research on world cities around global monopolies allows for considerable freedom and creativity. I will outline a methodology for investigating a small aspect of the urban control networks over world monopolies in the next section. 57 Methodology The focus of this dissertation will be the monopoly of access to the planet's natural resources and the spatial organisation of the control over this monopoly. While the monopolistic control over natural resources is global in scale, the analysis of it will jump across various scales from the global to the local and individual people. The analysis will take the form of a mixed methods approach applying both quantitative and qualitative methods. In order to establish a broad overview of the monopolistic control over natural resources, Chapter 3 in Part II considers the ten largest mining firms in the world and their spatial organisation with a particular focus on the ownership of these firms and the location of their shareholders. The ten largest firms control about one third of the world's mining business according to their controlled share of the total value of global non-fuel mineral production (Reuters, 2008). While this is not the majority of world production, it is a substantial part of it and establishing a geography of ownership and control over these firms, will provide new insights into what places can be considered the world cities of mining or, in world-systems terms, the world cities of the monopoly over access to the planet's natural resources. Data that trace the ownership of public companies to a specific place are not readily available. These data were assembled from the companies' annual reports and tabulated. The primary and secondary stock exchange listings of public companies are listed in their annual reports and all stock exchanges, where shares of at least one of the ten firms can be traded, were tabulated and ranked depending on their importance for trading. This provides an indication of the locations of stock trades. Depending on the reporting requirements of a specific stock exchange, public companies have to report any holding of more than three or five percent of a company's shares in their annual reports. In the case of Australia the 20 largest shareholders have to be listed. For the data generated in Chapter 3, the addresses of all publicly known shareholders were traced and control was allocated according to their shareholding, which resulted in a ranked list of cities with the largest share of ownership control. This methodology is further discussed in Chapter 3 as well as Appendix 1. 58 The focus on an urban control network over natural resources is refined in Chapter 4 by narrowing the analysis to a single resource. The high concentration of Platinum Group Metals (PGM) in a few regions of the planet makes it possible to map out the geography of ownership and control over approximately 90 percent of the world's entire platinum production. The United States Geological Survey (USGS) publishes data of platinum production by country and for most countries the data is broken down to individual mines. World platinum production is disaggregated to the mine level and ownership of each mine is traced to the corporations that operate these mines and their shareholders. Ownership data for the firms producing platinum is assembled in the same way as for Chapter 3, using the companies' annual reports. Subsequently, the locations of the shareholders or controlling institutions are mapped out to arrive at a list of places hosting the control functions over the world's platinum production. The quantitative analysis in Chapters 3 and 4 allows for the construction of two specific city networks. While I consider the cities included in these networks the world cities of the global mining industry and the world cities of the global platinum industry respectively, these networks are hierarchically organised. The methodology is therefore distinctly different from Taylor's interlocking network of world cities (Taylor, 2004a, 2005b; Taylor, Catalano & Walker, 2002). Actors in cities included in this network exercise control over capital in the mining industry. While there are likely linkages between these cities, these horizontal connections are not the focus of the analysis. While this approach does not provide a comprehensive overview of a single world city network, it allows me to trace 'vertical' connections from the locations of finance to the locations of production. These are the direct links between command and control centres in a particular industry and capital on the ground. Unlike in previous approaches these links are not established through the proxy of the producer services and their assumed command and control functions, but direct ownership from the top to the bottom of the hierarchy or vice versa. While the analysis in Part II is designed to gain insights into the corporate connections within the extractive industry at a broad level, Part III provides in depth insights into these types of connections with a focus on the South African mining industry primarily headquartered in Johannesburg. Both parts are a study of global linkages, but Part III takes the analysis to a more local context, recognising that national, 59 regional and metropolitan scales shape globalisation and are in turn shaped by it. The agency of individuals working in the extractive industry sector is an additional scale of analysis. Chapter 5 explores the specific national-historical context of South Africa that has shaped the way the mining industry in Johannesburg operates today. It considers policy developments during the apartheid era an how these have shaped developments in a democratic South Africa. The chapter specifically considers the role of large corporations in shaping policy and their international ambitions once apartheid ended. Chapter 6 analyses the role of Johannesburg as a world city in terms of its command and control functions in the mining industry. The chapter also considers the role of the South African state in shaping the city and its position in the world economy. Moreover, the interactions of mining communities with the locations of control and command are analysed in Chapter 6. Key informant interviews, official documents and news articles form the main source of information for this part. The dissertation ends with a conclusion in Chapter 7. The key informant interviews were conducted with executives in the mining industry, mining consultants, unionists, activists, government officials and academics. All interviews took place in South Africa during two periods of field work from May 23rd to June 6th and November 22nd to December 6th, 2009. Interviews took place primarily in Johannesburg, where I had access to a comparatively large number of interviewees that could provide specific information about the global mining industry with a focus on the role of Johannesburg and its links to investors world wide and investments in southern Africa. A small number of interviews was also conducted in Durban. The interviews were semi-structured with a stronger emphasis on open ended questions rather than structured questions, often evolving into a conversation depending on the particular areas of expertise of the interviewee. While a relatively broad cross-section of interviewees was approached, there was a focus on people in management and executive positions of mining firms as these exercise considerable control over the overall direction of an organisation. The focus in terms of interview requests was on platinum mining firms, because of the overall focus of this dissertation. Initially, companies were contacted by e-mail through addresses published on their web-sites. Interviews with representatives of four mining companies were conducted. In addition to the company executives, key officers from labour unions 60 in the mining sector were contacted and interviews with representatives from two unions were conducted. Once the first interviews in these two groups were completed, the group of potential interviewees was broadened to include consultants in the mining industry, activists, government officials and academics. Details on the interviewees cannot be provided under the conditions imposed by the ethics review for this dissertation. A total of 72 people were contacted with an interview request and 25 interviews were conducted. Of these eight people can be classified as academics, seven as mining executives, three each as activists and analysts/consultants and two each as government representatives and labour union representatives. Many people contacted did not respond to interview requests or cited limited time as a reason for declining an interview. It became apparent that different organisations have different resources to allocate to requests like mine and especially labour union representatives seemed pressed for time. Methodologically this dissertation builds partly on the extended case method (see Burawoy, 1991), which uses anomalous findings of case studies to adapt and reconstruct existing theories (p. 280). In other words, while the dissertation is undertaken within the theoretical framework of world city research and world-systems theory, it uses the case of the mining industry with a focus on Johannesburg to inform these theoretical approaches, adapt them to account for possible anomalies and reconstruct the theory. Moreover, the extended case method is concerned with the effects of the macro level on the micro level, i.e. how are local situations shaped by wider structures, or in the terms of Jürgen Habermas the extended case study is concerned with the effects of the system world on the life world. This represents methodologically an opposite approach from grounded theory, which attempts to draw generalisations from micro level studies to inform the macro (see Burawoy, 1991, pp. 271-287). Burawoy warns of studies that overemphasise the power of the system. “Thus those who would stress science over hermeneutics – the objective over the subjective – risk stressing the supremacy of system over life world. [...] Too often the system is seen as all-determining, so that forms of resistance such as innovation, negotiation, and rebellion are not taken seriously” (Burawoy, 1991, p. 284). An overemphasis on the system world is in essence also one of the major critiques of contemporary world city research (see e.g. Robinson, 2006). By utilising the extended case method, this research will respond to these critiques without rejecting the theoretical foundations of world city 61 research in world-systems theory, but it will interrogate the narrow focus on a few select world cities with a view towards reworking some assumptions of world city research. The study of the extractive industry, especially in Johannesburg, and most importantly the global connections within this sector from the mines of southern Africa to the centres of global finance is intended to stress the role of the life world within the system and thus reduce the overemphasis on the system of an interlocking network of world cities, which has recently crowded out the original emphasis of uneven development within world city research. Following Jessop, Brenner and Jones (2008) my intention in this dissertation is not to overemphasise a single dimension of socio-spatial research. Territories (T), places (P), scales (S) and networks (N), in addition to positionality (see Sheppard, 2002), are directly interrelated and I do not privilege any of these dimensions. Moreover, the mixed methods approach in this dissertation allows for a thorough analysis of the interrelated nature of the global and the local dimensions in the mining industry, thus overcoming the socially constructed dualisms between structure and agency or life world and system world (see also Urry, 2003). This research may be considered a form of what Burawoy (2000) termed 'global ethnography'. This is not an ethnography in the strict anthropological sense, in which the researcher fully emerges him-/herself  into a local society and studies the social structures in isolation, but it nonetheless studies the human relationships and it takes these out of their spatial confinements and traces them across the globe to investigate how the global level shapes such relationships or, in other words, how the system world shapes the life world. Burawoy (2000, p. 9-10) traces this form of global ethnography to the sociological traditions of the early Chicago School, which documents the connections between rural life of Polish peasants in the mid-19th century to the immigrant communities of Chicago in the late 19th and early 20th century. This 'ethnography' did not rely solely on interviews and participant observation, but also on letters exchanged between family members, newspaper articles, court records and autobiography. I will make no claims of having conducted an ethnography for this dissertation research and consider key informant interviews as one of the main methods applied (especially Chapters 4 and 6), but I want to subscribe to the particular global outlook of this 'global ethnography' and trace the connections that become apparent in the interviews. The personal interview 62 method was selected as it is well suited to 'probe deeply into the processes and mechanisms of international business' (Yeung, 1995). Ideally global ethnography should include work in more than one locale, but the emphasis here is on the connections in the mining industry centred around Johannesburg and although the field work was limited to one place, connections between firms or individuals to other places were explored by probing the experiences of the interviewees. The key informant interviews provided insights into institutional arrangements as well as the social forces at play in articulating these types of network formation. 63 Part II 64 Chapter 3: Cities as Switchboards of Processes “The world city is a world city not by being the origin of all things but as much by being a cultural switchboard through which the more or less peripheral cultures can reach one another.” Ulf Hannerz (1992, p. 51) In the prologue of his book 'World City Network' Peter Taylor (2004a) stresses the "second nature of cities". This second nature refers to the external relations of cities and their tendency "to be connected to one another" (p.1). Taylor revisits a classic article by Harris and Ullman, which famously shows the concentric, sector and multiple nuclei models to describe the internal structure of the city. This article has been a standard reference in urban studies for more than half a century. Taylor lauds the article for its interdisciplinarity, but notes that use of the article in urban studies has been very selective and is remembered only for half its argument, namely the second part of the paper dealing with the internal structures of cities, while the first part that deals with external relations of the city has been largely forgotten. For Taylor the use of the Harris and Ullman paper is a primary example for much of the post-world war II literature on urban studies, i.e. urban studies has focused on the internal aspects of the city, analysing it as if it were a closed entity without outside connections. He stresses that his studies focus on the interconnections between cities. The study he presents in the book is empirically based on the study of a large number of contemporary financial and business firms. This means it is global in reach, but narrow in topic. Taylor has produced a taxonomy of cities based on a fairly narrow aspect of the world economy that has to some degree become dominant in world city research. His rationale for focusing on these services is borrowed from Saskia Sassen's (2001) The Global City and her argument for using the producer services as a proxy for command and control functions in the global economy. Although early rankings of world cities have constructed a fairly strict hierarchy into alpha, beta, gamma etc. categories of world cities (e.g. Beaverstock, Taylor and Smith, 1999), later research that focused more on inter-city relations rather than attributes within cities produced a relatively flat and horizontal network of cities with some hierarchical tendencies rather than a strict hierarchy. This means that the world city network is defined more by its connections 65 between its nodes, which are of relatively equal importance, than a strict hierarchical organisation where one city can be clearly defined as more powerful than the other, although some hierarchical tendencies emerged (Taylor, 1997; Taylor, Hoyler, Walker & Szegner, 2001; Taylor, Walker, Catalano & Hoyler, 2002). This network spans the globe and cities complement each other in their functions to efficiently organise the world economy. These world cities are therefore considered the command and control centres in the world. I want to follow the idea of tracing the external relations of cities. However, I also want to broaden the scope of world city research away from the narrow focus on producer services to actually producing industries. In order to this, I want to borrow from the cultural analysis of Ulf Hannerz. His quote above brings forward the idea of world cities as switchboards of processes, rather than the origin of all things. The argument is simple: while the world city network may well hold the world economy together by efficiently organising capital and commodity flows, control over the world economy lies to a much lesser degree in the world cities than analyses focusing on producer services suggest. I want to return to John Friedmann (1986), who considered the world city in the light of the (then) "new international division of labour". The world city brings together capital and labour in an abstract way, but I argue that control over capital is much more dispersed and often lies outside those cities typically identified as world cities. A major shortcoming in constructing world city networks based on the presence of producer services and their inter-city linkages is that connections between money and capital are assumed, not investigated, traced or followed. In this chapter I will empirically address the connections between money and capital by constructing a network of cities that features prominently in extracting more than one third of the world's non-fuel mineral resources. By considering this specific aspect of the world economy, I am investigating one of the major monopolistically organised global industries. Following a world-systems approach, there is presumably considerable command and control power concentrated in these cities, although specific monopolies do not function on their own, but in concert with each other. 66 Trading the control over natural resources The challenge of data availability at the urban level has not disappeared since the inception of the Globalization and World City research network. The aim of this chapter is to identify the places where control over the extraction of natural resources is concentrated. The focus is on non-fuel mineral resources, which makes it possible not only to relatively clearly delineate the scope of the research but to analyse a major part of the global economy. Non-fuel mineral resources remain a core ingredient for the world economy and it seems the world gets increasingly hungry for minerals to produce consumer electronics, fuel construction booms or decorate the world's rich with high-end jewellery, amongst many other uses. Traditionally iron ore and other metals were considered strategic for industrial development and arms production. These metals were used in relatively large quantities. More recently non-fuel mineral demand has shifted to include more specialised metals, such as gallium, indium and rare earths, which are used for high technology applications and are elementary for the functioning of high technology-driven society (Ericsson, 2010, p. 16). According to the Swedish consulting firm Raw Materials Group7 only ten firms control more than one third of the world's non- fuel mineral production by value (Reuters, 2008) and only about 150 firms control 85 percent of the total global industrial mine production (Ericsson, 2010, p.7). Table 3.1 lists these ten firms, their production share, head office location and countries of operation. 7 Raw Materials Group is a consulting company for the mining industry based in Stockholm, Sweden. It maintains the database Raw Materials Data and is frequently cited as a source of information. 67 Table 3.1: The world's largest mining firms in 2006 Company Share of Global Production* Head Office Location Countries of Operation** Mining Operations*** Future Operations, Exploration, Reserves Vale 5.4 Rio de Janeiro, Brazil Brazil, Canada, Indonesia, New Caledonia Colombia, Peru, Chile, Argentina, Brazil, Canada, Guinea, Gabon, Dem. Rep. of the Congo, Mozambique, Angola, South Africa, Oman, Kazakhstan, India, Mongolia, PR China, Philippines, Indonesia, Australia BHP Billiton 4.8 Melbourne, Australia Australia, Brazil, Chile, Peru, USA, Canada, South Africa, Colombia, Suriname, Mozambique Guinea, Australia, Chile, USA, Peru, Canada, Colombia, Brazil, Guinea, Liberia, South Africa, Zambia Anglo American 4.3 London, UK South Africa, Namibia, Botswana, Chile, Brazil, Venezuela, Australia, Canada Brazil, South Africa, Chile, Zimbabwe, Botswana, Peru, USA, Colombia, Russia, Canada, Finland, China****, Australia, Namibia Rio Tinto 4.3 London, UK Australia, Ghana, Canada, Brazil, Guinea, USA, Argentina, Chile, Indonesia, South Africa, Zimbabwe, Madagascar Australia, Brazil, Guinea, USA, Chile, Indonesia, South Africa, Mongolia, Canada, Zimbabwe Freeport McMoran 3.3 Phoenix, Arizona, USA USA, Peru, Chile, Dem. Rep. of the Congo, Indonesia USA, Peru, Chile, Dem. Rep. of the Congo, Indonesia Codelco 3.2 Santiago, Chile Chile Chile, Brazil, Ecuador Xstrata 3 Zug, Switzerland Canada, Dominican Republic, Chile, Peru, Colombia, Argentina, South Africa, Australia, New Caledonia, Philippines, Papua New Guinea Ireland, Brazil, Tanzania, New Caledonia, Canada, Tanzania, Peru, Argentina, Philippines, Papua New Guinea, Australia Norilsk Nickel 2.7 Moscow, Russia USA, Russia, Botswana, South Africa, Australia USA, Russia, Botswana, Australia Barrick Gold 1.8 Toronto, Canada USA, Canada, Dominican Republic, Chile, Peru, Argentina, Tanzania, Papua New Guinea, Australia, USA, Dominican Republic, Chile, Argentina, Pakistan, Tanzania, Canada, Australia, Peru Grupo Mexico 1.6 Mexico City, Mexico Mexico, Peru, USA Chile Total 34.4 * by value of production of non-fuel minerals ** Data from 2009 company reports; Reuters, 2008 *** May include mineral processing, if the company report does not clearly distinguish. **** discontinued in 2009 The Table provides an overview of the firms' administrative centres and operations. The head offices are in a relatively diverse list of cities with only London 68 hosting two firms' head offices. Although the operations of these ten firms are spread out over all continents (excluding Antarctica), they are fairly concentrated in a relatively small number of countries. Nonetheless, their reach is global and decisions made at the head offices will affect conditions at the mines and the surrounding communities. However, corporations are ultimately responsible to their owners or shareholders, who collectively exercise control over the company, and expect a return on their investment in form of dividends and share price increases. Hence, the control over capital is more dispersed than the head office location indicates. I suggest that there are at least two additional layers of strategic control over the capital investments of companies. First, in the case of publicly listed companies, the locations of the stock exchanges, where the company is traded, is important. These places provide an indication where the companies are expecting to attract investors and tap into new investment capital, for example through the floatation of new shares. They are also the places where information is exchanged and where the companies' performance and ultimately the level of extraction of surplus value are evaluated through price setting by trading shares. Second, the locations of the residences or offices of the owners are important, because decisions regarding voting rights and strategic voting in company assemblies will originate there. These locations can of course be very different from the stock exchange locations. Availability of comprehensive data of this type remains challenging as there is no institution that collects such data (see Taylor, 1997). The GaWC Research Group has engaged in significant data collection exercises based on large transnational APS firms (e.g. Taylor, Catalano and Walker, 2002; Taylor, 2004a) and I am engaging in a similar primary data collection for this dissertation, albeit for a different sector and with a focus on ownership. All data in this chapter have been assembled through company reports, company and other websites, news sources or government issued information. All publicly traded companies are required by their respective stock exchanges and regulatory authorities in the respective countries to publish information about their company, including their primary and secondary listings and their major shareholders, usually those holding more than three or five percent, depending on the reporting requirements of the respective exchanges. In the case of the Australian Stock Exchange, the names of the 20 largest shareholders have to be published. 69 In an effort to establish the most important cities administrating the world's monopoly on natural resources, the stock exchanges, where the ten companies in Table 3.1 are traded, are identified and ranked in Table 3.2. They are ranked by the proportion of stocks that are traded at the respective exchanges. This data collection method has some limitations, but the Table provides an indication of the geography of control over natural resources. While all exchanges are identified in the company reports, it is not usually published which proportion of shares has been traded through which exchange. This information is estimated based on the distribution of shareholders by country or region, company origins or news reports. Most companies are engaged in American Depository Share (ADS) or American Depository Receipt (ADR) programmes, which allows them to float a portion of shares on American exchanges and access American money capital more easily. The proportion of shares held through ADRs for all companies that engage in such a programme, is listed in the reports and these firms are obligated to comply with specific reporting requirements by the Securities and Exchange Commission (SEC) of the United States. In the case of Anglo American, American banks are administering an ADR programme without the explicit engagement by the company itself. Only one of the ten companies is not a publicly listed company, namely Codelco of Chile, which is fully owned by the Chilean government.8 8 For details on data collection and methodology refer to Appendix 1. 70 Table 3.2: Stock exchanges of the ten largest non-fuel mineral producers Company Total Vale (Brazil) 5.4 BHP Billiton Limited (Australia ) 3.1 BHP Billiton Plc (UK) 1.7 Anglo American (UK) 4.3 Rio Tinto Plc (UK) 3.2 Rio Tinto Limited (Australia ) 1.1 FreeportM cMoran (USA) 3.3 Codelco (Chile) 3.2 Xstrata (Switzerla nd) 3.0 Norilsk Nickel (Russia) 2.7 Barrick (Canada) 1.8 Grupo Mexico (Mexico) 1.6 % WO NoF % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC London Stock Exchange 7.40 5 0.00 0.00 1.35 79.7 1.51 35 2.69 84.0 6 0.00 0.00 0.00 1.65 55 0.20 7.5 0.00 0.00 NYSE (New York) 5.69 5 1.28 23.7 0 0.01 0.20 0.01 0.70 0.00 0.19 5.94 0.00 3.30 100 0.00 0.00 0.00 0.90 50.0 0 0.00 Australian Stock Exchange (Sydney) 4.19 2 0.00 3.09 99.8 0.00 0.00 0.00 1.10 100 0.00 0.00 0.00 0.00 0.00 0.00 BM&FBOVESPA (Sao Paulo) 3.58 1 3.58 66.3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Unlisted 3.20 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.20 100 0.00 0.00 0.00 0.00 Johannesburg Stock Exchange (JSE) 2.47 2 0.00 0.00 0.33 19.6 1 2.14 49.7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Moscow Interbank Currency Exchange & RTS (Moscow) 1.88 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.88 69.5 0.00 0.00 Bolsa Mexicana de Valores (BMV) (Mexican Stock Exchange) 1.60 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.60 100 SWX Swiss Exchange (Zurich) 1.35 2 0.00 0.00 0.00 0.00 0.1 0.00 0.00 0.00 0.00 1.35 45 0.00 0.00 0.00 OTCBB Pink Sheets (New York) 1.20 2 0.00 0.00 0.00 0.65 15.0 0 0.00 0.00 0.00 0.00 0.00 0.55 20.5 0 0.00 0.00 Toronto Stock Exchange (TSX) 0.90 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.90 50 0.00 Euronext (Paris) 0.59 2 0.27 5 0.00 0.00 0.00 0.32 10 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Latibex (Madrid) 0.27 1 0.27 5 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Berlin Stock Exchange 0.07 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.07 2.5 0.00 0.00 Botswana Stock Exchange (Gabarone) 0.00 1 0.00 0.00 0.00 0.00 0.1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Namibian Stock Exchange (Windhoek) 0.00 1 0.00 0.00 0.00 0.00 0.1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 % WO = percentage of world output NoF = Number of Firms % MC = percentage by market capitalisation 71 The ten firms control 34.4 percent of world output and London and New York emerge as the locations, where the largest proportion of shares are traded. London accounts for 7.4 percentage points of the 34.4 percent (or 21.5 percent of the ten firms) and New York (NYSE & OTCBB Pink Sheets) for 6.89 percentage points (or 20.0 percent of the ten firms). While a greater share of the stock value is traded in London, seven companies are traded in New York and only five in London. A number of other cities are the locations of a considerable share in the trade of these raw material companies, including several cities usually categorised as peripheral, namely Sao Paulo, Johannesburg, Moscow and Mexico City. Although representing an insignificant share of the trade, stocks of one of the ten companies can be traded in Windhoek and Garbarone, while Berlin makes the list and Germany's financial centre Frankfurt does not. Control through ownership The location of the stock exchanges identifies the place where companies are traded. This is in many ways an important location of strategic control. Nonetheless, as suggested above, there is a second - perhaps more important - layer above the places of trading. The location of the owners is arguably, where the direction of companies is ultimately decided. However, this depends on the ownership structure of a company. I will return to the question of where power over large corporations is exercised later and will begin with mapping of the geography of ownership. In order to identify the locations of the owners of the top ten non-fuel mineral producers, information was systematically extracted from the company reports and proportionately allocated to a city. For example, Vale controls 5.4 percent of the world's non-fuel mineral production and 23.11 percent of shares (23.11 percent of Vale's market capitalisation) are knowingly held by persons with addresses in Rio de Janeiro, this means that 1.25 percent (5.4 x 23.11 / 100) of ownership control of the total global non-fuel mineral production by the ten largest firms is allocated to Rio de Janeiro.9 Table 3.3 lists the cities that host the owners of the ten companies controlling 34.4 percent of the world's non-fuel mineral production. The first column 'Total' lists the share of global non-fuel mineral production knowingly controlled in a city 9 For details on data collection and methodology refer to Appendix 2. 72 by all ten firms, adding up to a total of 34.4 percent in the bottom row for all companies. Following this column is a breakdown by company, whereas the sub-column '% of world output' is controlled in a city through the specific firm listed in this column. The sub- column '% by market capitalisation' indicates which share of the specific company is controlled in the city of that specific row. The picture that emerges is a diverse list of cities that is in contrast to previous world city research. 73 Table 3.3: World cities of mining Total Vale (Brazil) 5.4 BHP Billiton Limited (Australia) 3.1 BHP Billiton Plc (UK) 1.7 Anglo American (UK) 4.3 Rio Tinto Plc (UK) 3.2 Rio Tinto Limited (Australia) 1.1 FreeportMc Moran (USA) 3.3 Codelco (Chile) 3.2 Xstrata (Switzerlan d) 3.0 Norilsk Nickel (Russia) 2.7 Barrick (Canada) 1.8 Grupo Mexico (Mexico) 1.6 % WO % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC % WO % MC Other / Unaccounted 16.51 3.32 61.55 1.28 41.41 0.71 41.78 1.86 43.27 1.98 61.89 0.75 68.26 2.94 89.09 0.00 1.55 51.68 0.55 20.36 1.56 86.77 0.00 Santiago 3.20 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.20 100 0.00 0.00 0.00 0.00 Johannesburg 2.08 0.00 0.00 0.35 20.46 1.73 40.33 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 New York 1.89 0.00 0.00 0.00 0.27 6.32 0.27 8.38 0.00 0.25 7.6 0.00 0.18 5.98 0.80 29.48 0.12 6.63 0.00 Sydney 1.64 0.00 1.39 44.92 0.00 0.00 0.00 0.25 22.66 0.00 0.00 0.00 0.00 0.00 0.00 Mexico City 1.60 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.60 100 Rio de Janeiro 1.25 1.25 23.11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Moscow 1.17 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.17 43.25 0.00 0.00 Baar 1.10 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.03 34.38 0.07 2.44 0.00 0.00 London 0.90 0.00 0.00 0.57 33.79 0.20 4.62 0.13 4.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Los Angeles 0.55 0.00 0.00 0.00 0.00 0.28 8.85 0.00 0.00 0.00 0.15 4.94 0.00 0.12 6.6 0.00 Melbourne 0.49 0.00 0.40 12.87 0.00 0.00 0.00 0.09 8.13 0.00 0.00 0.00 0.00 0.00 0.00 Sao Paulo 0.39 0.39 7.3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Beijing 0.38 0.00 0.00 0.00 0.00 0.38 12 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Tokyo 0.32 0.32 5.98 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Pretoria 0.30 0.00 0.00 0.07 3.97 0.23 5.46 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Paris 0.25 0.00 0.00 0.00 0.00 0.16 4.86 0.00 0.00 0.00 0.09 3.02 0.00 0.00 0.00 Shelekhov 0.12 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.12 4.47 0.00 0.00 Brasilia 0.11 0.11 2.06 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Boston 0.11 0.00 0.00 0.00 0.00 0.00 0.00 0.11 3.31 0.00 0.00 0.00 0.00 0.00 Brisbane 0.01 0.00 0.01 0.28 0.00 0.00 0.00 0.00 0.43 0.00 0.00 0.00 0.00 0.00 0.00 Canberra 0.01 0.00 0.01 0.31 0.00 0.00 0.00 0.00 0.24 0.00 0.00 0.00 0.00 0.00 0.00 Adelaide 0.01 0.00 0.01 0.21 0.00 0.00 0.00 0.00 0.28 0.00 0.00 0.00 0.00 0.00 0.00 Total 34.40 5.40 100 3.10 100 1.70 100 4.30 100 3.20 100. 1.10 100 3.30 100 3.20 100 3.00 100 2.70 100 1.80 100 1.60 % WO = percentage of world output % MC = percentage by market capitalisation 74 Interpreting the results from Tables 3.2. and 3.3 requires considering some methodological issues in the data collection. The data set does not provide an all encompassing overview of the monopolistic control over the world's mineral resources, but it does account for more than one third of the world's non-fuel mineral production. Only ten firms account for this share, which in itself is an indication of the monopolistic or more precisely oligopolistic nature of the industry. The allocation of the proportion of shares traded at a stock exchange relies to some degree on estimates, but the number of exchanges is definite. In other words, all exchanges, where shares of any of the ten firms can be traded, are included and the number of firms that can be traded at a specific exchange are accurate, although the ranking and the percentage of shares traded may be somewhat inaccurate in some cases. Nonetheless, the data still provide a solid overview of the places where these shares are traded. Finally, place allocation of ownership is also not complete. Almost half of the ownership of the ten firms cannot be allocated to a place (16.51 percentage points of 34.4 percent), but the majority can be allocated, which again should provide a reasonable indication of the ownership of these firms. I will discuss the results in more detail below. Power in the world city network Most of the cities that host a stock exchange for the natural resource companies listed in Table 3.2 are usually identified in world city rankings, although - with the exception of New York and London - many are not ranked in the top echelons of world cities, but as "incipient global cities" (Madrid, Moscow, Toronto), "world-wide leading cities" (Sydney, Zurich, Mexico City) (Taylor, 2005b, p. 1606) or "Beta world cities" (Sydney, Toronto, Zurich, Madrid, Mexico City, Sao Paulo, Moscow) and "Gamma world cities" (Johannesburg, Berlin) (Taylor and Walker, 2001). Windhoek and Gabarone are exceptions in that they do not usually make any list of world city rankings. The cities that host the owners of the top-ten mining firms on the other hand show much more diversity in size, world city ranking and to some extent regional representation. In his original conception of the world city concept Friedmann (1986) emphasised the hierarchical nature of world cities. World cities are used by "global capital as 'basing 75 points' in the spatial organization and articulation of production and markets" (Friedmann, 1986, p. 71). Hierarchy has remained an important paradigm in world city research since the 1980s (e.g. Beaverstock, Taylor and Smith, 1999; Derudder, Taylor, Witlox and Catalano, 2003; Poon, 2003; Alderson and Beckfield, 2004). Some, especially earlier analyses, ranked cities by the presence of command and control functions in the respective cities (Friedmann, 1986; Sassen, 1991), while later studies emphasised that power is constituted through connectivity (e.g. Taylor, Walker, Catalano & Hoyler, 2002; Taylor, 2004a). Conceptions of power and the question of how power is derived are central to world city research. Although Friedmann included transportation and communication and the headquarters of large transnational firms in his list of articulators of world cities, later world city research has been based almost exclusively on producer services. This reflects a conception of power in world cities as 'power to' rather than 'power over'. 'Power to' refers to the power to service global capital and exercising this power depends on a network of cities, in which each node plays a particular role in servicing global capital. The power lies in the network and is relatively horizontal, although some cities have more connections than others and can therefore be considered more powerful (Taylor, Walker, Catalano & Hoyler, 2002, p. 232). The list of cities in Table 3.2 reflects this conception of power with New York and London emerging at the top with the highest trading volumes and the most firms traded. Some other cities that are usually highly ranked in world city rankings do not appear in this table (e.g. Tokyo, Hong Kong, Singapore, Frankfurt), but this lies most likely in the nature and regional clustering of the industry. However, this conception is based on producer services, i.e. those who service global capital or who administer it. It does not reflect any 'power over' in terms of ownership of capital, i.e. the means of production. The cities listed in Table 3.3 reflect this type of power and cities in South America and Africa appear at the top. Although it may be difficult to decipher which conception of power reflects more accurately control over global economic processes, it is conceivable that those servicing global capital will lose their power and influence without anyone actually owning capital and those owning capital are - at least somewhat - dependent on those servicing global capital. However, in the final analysis ownership can hardly be ignored and the vertical connections between capital 'on the ground' and ownership shape the global economy and the lives of millions of ordinary people. 76 In order to investigate how 'power over' manifests itself through ownership of the means of production, it is useful to consider what types and structures of ownership are dominant in the world economy. The focus of this dissertation is on mineral resources, which provides insights into a particular industry that allows for some limited generalisation and will provide suggestions for the research of other industries. Types of ownership Five types of ownership can be distinguished amongst the ten largest producers of non-fuel minerals: 1) public corporations (including other mining firms, insurance companies); 2) large individual shareholders and private companies that are controlled by individuals, families or small groups of people); 3) pension funds, investment funds and nominees; 4) small individual shareholders; 5) governments. Table 3.4 cross-lists these types of ownership and their characteristics, control capacity and the firms owned under this type of ownership. In addition to these types of ownership, the distribution of shareholders should be analysed in an attempt to assess, where the control and command functions over the industry are located. A diversified ownership without a dominant shareholder most likely shifts the locus of control towards the board of directors, whereas the presence of dominant shareholders shifts the control to the location of these shareholders. There has long been a large literature on the question of where corporate control is located (e.g. Berle and Means, 1932; Herman, 1981), but there is no universally accepted taxonomy of control patterns. Zajac and Westphal (1996, pp. 507-508) identify three streams of research on the power of corporate boards. One stream has generally considered boards as passive tools of management interest. A second stream has focused on interlocking directorships across multiple corporations and how this network of directors on different boards affects corporate culture and practices. And, a third stream investigates the role of corporate boards in protecting shareholder interests. One of the main difficulties with any attempt to identify the locus of control in large corporations remains the variation in shareholder engagement and intra-corporate interactions across different parts of the world. Most studies are conducted within Anglo-American or 77 western European corporate culture, but the growing influence of other parts of the world in corporate wealth will change any patterns. Even within the western economies different cultures lead to different patterns of corporate governance. As Gordon Clark (1998) points out pension fund savings during the 1980s and early 1990s were increasing dramatically in the Anglo-American world, while they remained largely absent in continental Europe. The following discussion is partly informed by this literature, but draws also on key informant interviews with mining executives and analysts held in Johannesburg as well as company reports and news coverage. Table 3.4: Ownership types of top-ten mining firms Public corporations Pension funds, investment funds and nominees Large individual shareholders, private companies Small individual shareholders Governments Characteristi cs web of cross ownership, securities for financial industry, including insurance firms and banks represents a large number of individuals each accounting for minimal ownership firms are in individual hands or family ownership dispersed ownership of non- substantial shareholding the state maintains a major interest in the control over the means of production Control capacity board of directors or executives will decide on voting decisions on voting are usually transferred from the owner to a financial institution and decisions are made by fund managers or other representative s owners will make direct decisions individuals will make decisions on voting rights, but these are not of major influence control lies with government officials, ministers Companies Rio Tinto, Xstrata Vale, BHP Billiton, Anglo American, Rio Tinto, Norilsk Nickel (and likely Freeport McMoran and Barrick) Norilsk Nickel, Xstrata, Grupo Mexico (Anglo American used to be dominated by the Oppenheimer family) all publicly- traded companies Codelco (100 %), Vale (small proportion) 78 There is very little cross ownership amongst the top-ten mining firms. One exception is a holding by Xstrata's dominant shareholder in Norilsk Nickel through one of Norilsk major shareholders, Rusal (traded in Hong Kong and Paris). The Chinese mining firm Chinalco also owns a twelve percent share of Rio Tinto. There is also very little ownership by financial corporations, although the French insurance company AXA is holding shares in two of the companies. Six of the ten largest non-fuel mineral producers in the world, including the five largest ones, have a highly diversified ownership and no single shareholder exceeding a silent interest of more than 25 percent, with most owners holding a share smaller than ten percent. Four of these firms, the four largest producers, have major shareholdings through pension funds, investment funds and nominees (banks or other financial institutions representing a large group of individual shareholders with small holdings). Ownership could be traced to between 30 and 60 percent of all outstanding shares for these four companies. The financial industry in form of banks representing individual shareholders, investment funds as well as pension funds collectively play a major role in controlling these firms and their strategic decisions on investments or divestments. Remarkably, territory remains a significant spatial component of control over these firms. The largest owners of the largest of the companies, the Brazilian firm Vale, can be traced to Rio de Janeiro, Sao Paulo, Brasila as well as Tokyo. All of the 20 largest shareholders of the Australian arm of BHP Billiton are located in Australian cities, with most of them in Sydney (44.92 %) and Melbourne (12.87 %), giving an indication of the company's origins in the Australian firm Broken Hill Properties (BHP). All major shareholders of the UK listed arm of BHP Billiton are located in London (33.79 %), Johannesburg (20.46) and Pretoria (3.97), representing the company's London listing and its partial origin in the South African mining firm Gencor, which in the 1990s bought Billiton, which was then the non-fuel mineral arm of Shell. A similar pattern emerges for the firms Rio Tinto and Anglo American. All of the 20 largest shareholders of the Australian arm of Rio Tinto are located in Australian cities and Anglo American is owned with 40.33 percent in Johannesburg and 5.46 percent in Pretoria, while 6.32 percent ownership can be traced to New York and 4.62 percent to London. The UK listed arm of Rio Tinto 79 shows the most regionally diverse ownership with shares being held in Beijing (12 %), Los Angeles (8.85 %), New York (8.38 %), Paris (4.86 %) and London (4.02 %). Some of the ten companies are publicly listed, but controlled or largely owned by individuals, families or privately owned companies. These companies include Xstrata, Norilsk Nickel and Grupo Mexico. The locus of control in this type of ownership structure lies entirely or largely with the major shareholder(s) and their location can be quite different from the company's headquarters and can show quite different characteristics from typical world cities. Xstrata is owned to 34.38 percent by Glencore International AG, a privately held Group owned by its management and employees. Glencore's head office is in the small town of Baar, Switzerland, not far from Xstrata's head office in Zug. Glencore has about 50 offices in more than 40 countries and employs more than 50,000 people in its industrial operations in 13 countries. Its operations focus on commodity trade and it has majority or controlling interests in a number of publicly listed firms in the mining and energy industry (Glencore International AG, 2010). Norilsk Nickel is under the control of two dominant shareholders. Vladimir Potanin owns 25.0013 percent of the firm and Rusal Plc owns 25.1299 percent, each exceeding the 25 percent threshold of a silent interest. Rusal Plc is in turn owned to 47.41 percent by En+, a private company owned by Oleg Deripaska. Both men are amongst the richest people in the world10. The remaining parts of Rusal are owned by a Russian private investment fund, a mining firm from eastern Russia, Glencore (indirect) and roughly eleven percent public float (Norilsk Nickel, 2010a; En+ Group, 2008; Rusal, 2010). There is very little information about ownership published in Grupo Mexico's annual report, but news reports indicate that the company is owned by and under the control of German Larrea, member of the Larrea family. Larrea is Mexico's second richest person (Rosenberg, 2010). The ownership structure of these three firms suggest that there is substantial control concentrated in few hands and these individuals and families exercise the capital commanding functions without any major influence from the financial sector or the producer services in general. 10 Oleg Deripaska ranked 57th on the Forbes list of world billionaires. Vladimir Potanin ranked 61st on the same list. Source:Forbes. (2010). Deripaska_UCP9.html; Potanin_KZ32.html . [last accessed: 2010 December 16]. 80 While almost all companies have some individual owners holding small numbers of shares, two firms, namely Freeport McMoran and Barrick, seem to be held almost exclusively through this type of ownership. Only about eleven and 13 percent respectively can be traced to specific owners for these two companies, all of who are in New York, Los Angeles and Boston. Given the firms origins, main listings and assets, it can be assumed that most of the owners reside in the United States and Canada. However, these are likely dispersed across cities and towns in North America, without a dominant single city. Freeport McMoran's links to its places of origin in Louisiana and Arizona appear to remain strong, given the composition of its board of directors. Only one firm is owned by its respective government, namely Codelco of Chile. Control over this firm has been attributed 100 percent to the place of political control in Chile, i.e. the country's capital Santiago. Codelco is not publicly listed. The Brazilan government maintains a 5.4 percent interest in Vale, but control is allocated to an agency in Rio de Janeiro, not the capital Brasilia although ultimate control of the share may lie in Brasilia. The Brazilian government also maintains some veto rights, such as decisions on the head office location and corporate purpose relating to mining activities, through so called twelve golden shares (Vale, 2010a). This type of control lies with the governments and, should the mechanisms of democracy function well, with the people of the respective country.  Commanding or servicing capital? Those who own capital are usually in a position to command it. However, the relationship between capital and their owners may not be that straightforward, because in order to enable owners to efficiently organise their capital, producer services are required. At the same time, virtually none of those providing producer services, actually own a substantial part of the capital that controls more than one third of the world's non-fuel mineral production. The question, then, is who is in command of the monopoly over natural resources? The most common forms of ownership in the top-ten non-fuel mineral producers, are investment and pension funds, nominees and individual owners. Individual owners 81 can represent themselves at general meetings of the companies and are likely spatially dispersed. While nominees are usually located in large cities, the world cities, their power rests on their right to represent individual owners, who are likely not located in world cities, although the boundaries of the nation state still seem to matter. Even investment and pension funds represent the savings of individuals, who can divest from funds or, in the case of pension funds, collectively take action against the investment strategies of their fund.11 While considerable decision making power lies in the hands of producer services under this type of ownership, many individuals choose not to be represented through a nominee and the decision making power of bankers and fund managers could potentially be taken away with the stroke of a pen by those who actually own the shares. Although withdrawal of all shareholders from nominee arrangements and massive shareholder activism through investment and pension funds is unlikely, these institutions need to cater towards the interests of the actual owners to some degree and the system is potentially fragile. In addition to the potential vulnerability of the command functions exercised by producer services, there is substantial ownership concentrated in private hands that will require some producer services, but ultimately maintain control over the capital they own. These private owners, e.g. Glencore management and employees, Russian industrialists Vladimir Potanin and Oleg Deripaska as well as the Larrea family in Mexico, will retain the control and command functions in their own hands. Similarly, governments partially or fully holding stakes in companies may require some producer services, but will maintain control over their share. Corporations owning shares of other corporations will likely also maintain control over their assets and not transfer voting rights. While allocation of ultimate control over the monopoly of natural resources is difficult to establish, the ownership structure of the companies controlling more than a third of the world's non-fuel mineral production suggests that control and thus capital commanding functions are exercised in rather diverse ways and are by no means the exclusive domain of the producer services. 11 A number of trade unions requested from the Council on Ethics for the Norwegian Government Pension Fund that the fund divest from Grupo Mexico, because of human rights violations by the firm (International Metalworkers' Federation, 2010). Grupo Mexico is not currently on the list of "companies excluded from the investment universe" by the Council on Ethics for the Norwegian Government Pension Fund, but a number of firms, including Rio Tinto, are listed (Ministry of Finance, 2010). 82 Sassen's (2001) rationale for privileging the producer services in world city research is that these are the services that facilitate investment flows through a network of cities and are essential in organising a world-wide, dispersed post-Fordist production of goods and services. She thus considers them the control and command functions as they hold the system together. Taylor, Walker, Catalano and Hoyler (2002; see also Taylor, 2004a) go one step further and argue that power does not only lie in the size of a node, i.e. the concentration of producer services in a city, but its connectivity. They distinguish between "connectivity-through-dominance, connectivity-through-subordination, and neutral connectivity" (p. 236), where a city exercises connectivity-through-dominance, if it hosts more higher order offices (headquarters, regional headquarters etc.) than the other city that it is connected to. Similarly, if a city hosts fewer higher order offices, it is characterised by connectivity-through-subordination, and if both cities host equal numbers of higher order offices, they have a relationship of neutral connectivity. Both perspectives consider flow as the decisive characteristic in exercising command and control functions. While flow is certainly important, without capital investments, there will be no flow and the power of controlling the flow ultimately still depends on those owning capital. Some flow is generated through the financial industry itself and the creation of fictitious capital, i.e. the re-packaging of securities, but these are bubbles that regularly pop. The most recent examples are the Asian financial crisis of 1997, the bursting of the bubble in 2001 and most recently the global financial crisis following the sub-prime mortgage crisis of 2008. Given that any constructions of world city rankings and networks are (almost) exclusively based on producer services, the world cities are perhaps more the glue in the system, rather than the ultimate control points. Or, borrowing from the cultural analysis of Ulf Hannerz at the opening of this chapter, they are switchboards of processes, rather than the origin of all things or the ultimate locus of global control.  World cities in the world of mining After more than two decades of world city research, we know a lot about the flows that articulate the 'power to' service capital through the construction of world city 83 networks based on producer services. However, I consider ownership the main control mechanism and having 'power over' capital. Table 3.3 expresses this view with a list of cities in which ownership of the ten largest non-fuel mineral producing firms is concentrated. The result is surprising, with Santiago and Johannesburg topping the list, ahead of New York in third place. Of the ten cities knowingly hosting the largest share of the ownership of the ten firms, only four are located in Western Europe or North America, one of which is a small town, representing private wealth held by very few people. The other cities represent the dominant urban centres in the resource producing countries. While perhaps surprising at first sight, the prominence of cities in the global south hosting the owners of capital, is unlikely to mean that the world-system has been turned upside-down. Although the data clearly show a considerable concentration of ownership in cities outside western Europe and North America, ownership in these two regions is likely underrepresented, due to difficulties in data collection. Ownership of two firms, Codelco and Grupo Mexico, is attributed completely to one owner, a state and an individual. These two owners are located in cities in the global south. Moreover, owners holding shares in Anglo American and BHP Billiton through the JSE tend to be represented by a single nominee, because of the electronic trading system at the JSE. This means that small individual owners are all included for Johannesburg. On the other hand ownership of the two North American companies, Freeport McMoran and Barrick, could be established only to a very small degree, but owners can be assumed to be dispersed across North America with some minor holdings in other parts of the world. While ownership of Vale, BHP Billiton, Anglo American and Rio Tinto could be allocated to place to a larger degree than for the North American firms, the majority of owners of these firms are likely dispersed in Europe and Australia as well as Brazil (given their stock exchange locations, headquarters and places of origin). Given the ownership structures and some reporting practices, there is probably more ownership and thus decision making power concentrated in North America and Europe than Table 3.3 suggests. These results reflect to some degree the flatter urban hierarchy and more even regional distribution of wealth within the countries of Western Europe and North America, compared to semi-peripheral countries like Brazil, South Africa, Mexico and 84 Russia, where wealth might be more concentrated. Another distinguishing feature between cities is the number of companies that a city's influence rests on. The power attributed to Santiago, Mexico City, Rio de Janeiro and Moscow rests on the presence of single companies. For all of these companies head office location and ownership largely overlap. On the other hand, New York's power rests on ownership in six of the ten firms, while Los Angeles' and London's rest on ownership in three firms and the power of Johannesburg, Sydney, Baar, Melbourne, Pretoria, Paris and three more Australian cities rests on two firms. Because of the unique arrangements and reporting requirements for ADR programmes, New York may be artificially over represented compared to other cities, but generally those cities that are usually ranked highest in world city rankings show the most diverse portfolios in this particular construction of world cities. They are located in Europe and North America. Considering Braudel's (see Taylor, 2000, pp. 7-10) argument that the capitalists that show the highest degree of flexibility are the most powerful ones as they can re- organise quickest should a particular capitalist venture fail, New York's as well as to a lesser extent London's and Los Angeles' power rests on the most robust foundation, while those relying on a single firm have a lot less flexibility. Producer services represent this power through portfolio investment, which means flexibility, and this ranking of world cities in the mining industry confirms that view, as owners in these three cities are exclusively banks (nominees) and investment funds. At the same time, the concentration of wealth in individual or few hands cannot be ignored, as these persons have ultimate control over productive assets, while those focusing on portfolio investment do not actually hold controlling shares. Given the diverse types of ownership in these companies, power is likely constituted through a specific interplay between those who focus on portfolio investments because of the flexibility associated with this particular type of investment and those who choose direct investment because of its ultimate control over specific assets. 85 Satellite functions and national territory The distribution of wealth across more cities and a larger territory in Europe, North America and Australia, which in spite of its small population features five cities and a diverse ownership of relatively small shareholders, represents to some degree the remnants of the historical compromise between capital and labour in these countries (see, e.g., Amin, 1994). The wages of labour kept pace or exceeded overall growth until at least the late 1970s. This brought more people into a position to invest in capitalist ventures with a view to expand their share of the surplus and thus more flexibility at a later stage. Amin (1994) argues that such a compromise did not occur in the third world and that the relative equality in the rich countries could only be achieved through the ever increasing extraction of surplus value from labour in the third world. Those cities in Table 3.3, whose influence rests on a single company, mainly represent countries, in which this compromise did not take place and ownership is highly concentrated. Baar in Switzerland is clearly an exception to the rule, but perhaps an indication of the shifting distribution of wealth and the end of the historical compromise in Western Europe. Glencore International AG, the single private firm that Baar's influence rests on, was only founded in 1974. Regardless of the ownership structure of individual firms, it is only ten companies that account for more than one third of the world's non-fuel mineral production by value. This organisation of ownership over productive assets indicates a monopolistic or oligopolistic structure in the mining industry. Santiago, Mexico City and Moscow rest their influence on a single firm and concentrate national ownership and control in one city, whereas there is a significant difference between state ownership in Chile and individual ownership in Mexico and Moscow. The South African city pair Johannesburg- Pretoria, the Brazilian city pair Rio de Janeiro-Sao Paulo (plus Brasilia) and the Australian cities represent a variety of owners, including a large number of individuals and pension funds in their own countries. This group of cities, nonetheless, rests its influence on only one or two firms per country. The established world cities show greater diversity through portfolio investments. Although Mexico City and Moscow both rank highly as measured by concentration of ownership, they may still be fulfilling typical satellite functions (see 86 Frank, 1972), where the city facilitates the extraction of minerals and surplus value through large scale individual owners with the consent of the government. While actors in these cities have the potential to exercise considerable capital commanding power, they are likely to exercise it in a specific way, because their interest is in continuous business and their loyalties are to the customers of the raw materials rather than national labour, an arrangement that ensures a steady flow of income for the owners. Because of the more diverse ownership in the South African, Brazilian and Australian cities, actors in these cities may be forced to show greater loyalty to their local regions or countries, although the dynamics will be very specific in each case. In Brazil, Rio de Janeiro's and Sao Paulo's influence is based on one company that is headquartered and traded in Brazil. In South Africa, the influence is based on partial ownership in two companies, which are traded primarily in London with a secondary listing in Johannesburg, while the firm's headquarters are in Melbourne and London. If the example of these individual firms can be generalised, the relationship between city and the territory of the nation state will be quite different in both cases. Another feature that requires the attention of a world-systems analysis is that those cities that are not established world cities and represent the global south are all located in so called middle-income economies, not the poorest countries of the world12. They are mostly large functional cities that fulfil typical gateway functions (e.g. Sassen, 2001) or can be considered linking cities. There is certainly some 'power over' capital concentrated in these cities of the global south or the former second world. However, this 'power over' is also shaped by a complicated relationship between the state-territorial power, which facilitates capitalist ventures through labour laws and mineral concessions, and networks of cross ownership and capital flows, which is articulated through the efficient administration of world cities and its satellites. Historical dimensions and shifts in power There are historical dimensions that shape this specific ranking of cities by ownership over mineral production. There are also some new geo-political formations 12 Brazil, Chile, Mexico and South Africa are all included in the Upper-middle-income economies ($3,946 to $12,195) category of the World Bank (World Bank Group, 2010). 87 that slowly become apparent in this list of cities. With the onset of neo-liberalism across the world, large state owned corporations have been privatised over the past twenty years. Two of the ten largest mineral producers, Brazilan Vale and Russian Norilsk Nickel, used to be 100 percent state owned. The Brazilian government facilitated the privatisation of Vale in the 1990s and it was officially privatised in 1997, thus transferring productive assets from the control of the federal government into private hands. The government only holds a small non-controlling share in the company and reserves a veto right on specific decisions (Vale, 2010a, p. 108). Norisk Nickel was founded in 1989 as a state owned corporation in the former Soviet Union and privatisation was decided by decree of the Russian president in 1993. Shares were auctioned in 1994 (Norilsk Nickel, 2010b). Both examples, show the transfer of productive assets from state into private hands, whereas Vale has a more diverse ownership today than Norilsk Nickel, which is practically controlled by two men. The Chilean firm Codelco remains an exception as a firm that is 100 percent owned by its respective government. Nationalisation of mining assets in Chile was decided in 1971 under the democratic-socialist government of Salvador Allende. Privatisation was discussed after the military coup in 1973. The new government under General Pinochet started a neo-liberal experiment of privatisation of state assets and state ownership was inconsistent with the policies of the government, but the military favoured state ownership of this strategic sector and in 1976 Codelco was founded as a state owned company (Lagos, 1997, p.57; Codelco, 2010a). It seems like an irony of world history that the country that has first experimented with neo-liberal policies of privatisation, created a state owned mining firm that to this day remains the only state owned company of the top-ten non-fuel mineral producers in the world. Although Codelco is a financially successful company, its production is most concentrated in one country and based on the enormous copper reserves in Chile. Other firms, on the other hand, have a much more diversified production and access new capital through a network of world cities that host the world's largest stock exchanges. The absence of cities in Asia in Table 3.2 and the low ranking of only two cities in Asia in Table 3.3, suggests that a number of world cities are of very little importance in non-fuel mineral production. Because I have focused on a small, albeit elementary aspect of the global economy, this pattern reflects to some degree that the power of world cities is derived through the "conjoint actions of the attendants who maintain the world 88 city network (and thus the global space economy)", rather than the competitive advantages that each individual city has (Beaverstock, Doel, Hubbard and Taylor, 2002, p. 128). In other words Asian cities are not specialised in the trading of shares or ownership of natural resource companies. However, most of the publicly traded companies aim to access capital through secondary listings in various parts of the world and one of Norilsk Nickel's major shareholder, Rusal, is listed at the Hong Kong Stock Exchange (Norilsk Nickel, 2010a) and Vale announced a listing on the Hong Kong Stock Exchange through Hong Kong Depositary Receipts starting December 2010 (Vale, 2010b). This is perhaps an indication that profits from other Chinese enterprises are seeking an outlet for new investment in new sectors and geographical regions. Moreover, a Tokyo based firm is already holding shares in Vale, representing Tokyo in Table 3.3 and the Chinese state owned enterprise Chinalco owns a 12 % share in Rio Tinto, representing Beijing. Asia, especially China and India, is also of global importance through its role as a major market for the sale of non-fuel minerals, as the reports of all ten companies indicate. In addition to that, there is considerable power concentration in the production of non-fuel minerals in China, which is the largest mining country, if all metals are included (Ericsson, 2010, p. 11). Individual Chinese companies are too small, however, to be included in the top-ten list and although Chinese companies generally focus on production in China, they increasingly invest in projects overseas (Shankleman, no date, p. 22). China also seeks to gain greater influence in Africa through intensifying diplomatic relations and investments on the continent (Mawdsley, 2007; Carmody and Owusu, 2007). Currently the Chinese mining industry is in practice state controlled (Ericsson, 2010, p. 11), but given the importance of the sector to the Chinese economy some new city network formations may emerge in this part of the world. The global mining industry for metals, diamonds and uranium has grown significantly from approximately 80 billion US Dollars in 1995 to 463 billion in 2008, after a period of stagnating demand from the 1970s until the late 1990s (Ericsson, 2010, pp. 7-8). The post cold war period has seen a high level of concentration in the mining industry. While in 1995 the top-ten non-fuel mineral mining firms controlled 23 percent of the global market share, the ten largest firms controlled 35 percent in 2008 (Ericsson, 2010, p. 11). After a series of mergers and acquisitions, large diversified mining firms 89 started operating in all parts of the globe, while some previously state owned firms have broken out of their national confines. In South Africa, the mining industry went through a period of significant change, with the liberalisation of the economy at the end of apartheid. Under apartheid profits from the mining operations could not be invested overseas, because of exchange controls, and Anglo American was forced to invest its profits into other sectors of the South African Economy. By the 1980s Anglo American accounted for more than 30 percent of the republic's GDP (Jamieson, 1990, p. 2). At the head of this empire was still the Oppenheimer family. Although it nominally held only 8.3 percent of the company, their family's influence and the sense of patriarchy was immense (Jamieson, 1990, p. 5). With the end of apartheid Anglo American had only 11 percent foreign ownership, but - together with a handful of other companies - it was allowed to leave the country, relocate its headquarters to London and build a global empire efficiently administered through a network of world cities (I will discuss the corporate history of Anglo American in more detail in Chapter 5). Diversified mining, concentrated ownership and the old north-south divide Only ten companies account for more than one third of the global non-fuel mineral production. The data presented in this chapter are based on these ten companies and are intended to shed light on the urban control of this monopoly. Combined, these ten companies span the world and facilitate the efficient extraction of mineral resources around the world from their headquarters in nine different cities. Two distinct lists of cities emerged from the data collection exercise. The first one represents the cities with the 'power to' service global mining capital, the other represents the cities with the 'power over' this capital. The first list (Table 3.2) includes primarily cities that are commonly included in contemporary lists of world cities, with New York and London emerging at the top. These cities are world cities because they are hosting the functions that enable a dynamic exchange of control and ownership of productive assets. They are switchboards for money whose owners are seeking new opportunities to invest in capital with a view to 90 increase the surplus in its money form. The second list (Table 3.3) shows a considerable number of cities that are not usually at the top of world city rankings, but are important national or regional centres, many of these cities are in the global south. Considerable control over capital is located in these cities and firms originating in the global south or the so called former second world have expanded far beyond their national territorial boundaries. The spatial structure of ownership over productive mining assets suggest that polarisation at a global scale (Amin, 1974) is starting to take on new forms and shapes. As Amin himself observed, we may be moving towards a world that resembles that of Marx's times more than that of the post world war II period (Amin, 1994, p. 88). Non- fuel mining production is controlled in a diverse group of cities around the world, rather than in the so called global metropolis consisting of North America and western Europe, although the global metropolis is still on the receiving end of the global division of labour. Generally higher wages are paid in these regions than in areas of the global south. However, wages for workers in Europe and North America have declined and there is an increasing accumulation of wealth in the global south. The list of world mining cities in Table 3.3 may reflect this new organisation of the global economy, where a network of cities is facilitating the efficient extraction of natural resources and surplus value. At the same time, the network of world cities that has the 'power to' service global capital will retain considerable strategic control and facilitate an interplay between those capitalist that hold 'power over' productive assets and those who seek to invest their money.  91 Chapter 4: Control over Platinum: Cities, Firms and Governments Each time when they announce [financial] results, they have to go overseas [from Johannesburg] [Interview 23]. Natural resources are largely produced under monopolistic conditions and this trend is particularly pronounced in the production of Platinum Group Metals (PGM). Only ten firms controlled 93 percent of the global production in 2009, while the largest firm controlled more than 30 percent and the largest three had a share of approximately 65 percent (Ericsson, 2010, pp.11, 13). Chapter 3 established an urban geography of the places of command and control over natural resources, specifically non-fuel minerals. This chapter will narrow the focus to a single resource: platinum. PGM are mined in relatively small quantities, but their value is extremely high. Prices for platinum itself, for example, went up as high as US$ 2276 per troy ounce13 in March 2008. The highest price for gold during 2008 by comparison was US$ 1002 per troy ounce. Prices for other PGM are usually lower, as for Iridium, Palladium and Ruthenium, but Rhodium reached a price of US$ 10,050 per troy ounce in June 2008. This high price was partly caused by real demand and partly by speculation. In late November 2008 Rhodium traded for US$ 1000 per troy ounce. Although speculation clearly played a role in the extreme price fluctuations of Rhodium and other metals in this group, PGM are widely used in industrial applications and prices are largely determined by demand and supply. This is unlike gold, which has few practical uses and is mainly used as an investment instrument. About 50 percent of PGM are used in the production of catalytic converters in the automotive industry. It is also used for jewellery and other industrial applications including in the medical sector (T.J. Brown et al., 2010, pp. 38, 80-81; United States Geological Survey, 2010a). In 2008 PGM accounted for about two percent of the total US$ 463 billion global mine production of metals, diamonds and uranium (Ericsson, 2010, p. 7). The monopolistic control over PGM production can partly be explained by the geography of the mineral deposits in the earth. About 88 percent of known PGM reserves 13 1 troy ounce (tr oz) = 1.09714 avoirdupois ounces or 31.103 grams (United States Geological Survey, 2009, p. 190). 92 are in South Africa. Russia accounts for about eight percent, while the remaining parts are located in the United States, Canada, Zimbabwe and Colombia (United States Geological Survey, 2009, p. 123). In 2008, South Africa accounted for 59 percent of all PGM production and 77 percent of platinum. The only metal within the PGM for which South Africa was not the largest producer was palladium. Russia produced 43 percent of the world's palladium, while South Africa produced 37 percent. South Africa accounted for 75 percent of the production of all other PGM. Russia was the second largest producer of all PGM combined, accounting for 26 percent, followed by Canada with six percent and the United States and Zimbabwe with three percent each (United States Geological Survey, 2010a). The high level of corporate concentration in a few firms in combination with the small volumes being mined makes it possible to trace the world's entire platinum production to its places of origin, i.e. the mines. This also means that almost every ounce of platinum produced can be allocated to a firm and the firms that produce platinum have shareholders, which in turn can be allocated to a place. This tracing of connections from the places of production to the places of control makes it possible to establish a list of world cities in the platinum industry, a list of cities that is in control of a specific monopoly. These are just the world cities of the platinum industry and this network of cities is just a snap shot of a small aspect of the global economy, but unlike in other world city research, the connections between capital commanding functions and capital itself are traced and empirically researched, rather than assumed. The numerical tracing of ownership in the platinum industry is followed by a qualitative analysis of key informant interviews conducted with a diverse group of actors in the platinum industry. This part evaluates the power relationships between locations identified in the first part of this chapter. Bottom-up world city research Most studies of world city networks focus on establishing a comprehensive network of cities that are linked through the offices of large producer services firms (e.g. Taylor, Catalano and Walker, 2002). Some studies utilise other data such as airline traffic 93 (e.g. Derudder, Witlox and Taylor, 2007) or the office locations of political and civil society organisations (Taylor, 2004b, 2005a), but these almost always produce horizontally organised networks with some hierarchical tendencies (see Taylor, 1997; Taylor, Hoyler, Walker & Szegner, 2001; Taylor, Walker, Catalano & Hoyler, 2002; Derudder, Taylor, Witlox & Catalano, 2003). In this chapter connections from the world cities to the peripheral locations of the global economy are studied and these connections are vertical and hierarchically organised. Methodologically, such a construction of data involves several steps tracing the connections from the places of production (i.e. the mines) to the places of control (i.e. ownership). Global platinum production in 2007 was only 212,000 kilograms (see Table 4.1). The five largest producer countries accounted for more than 98 percent of total production, with South Africa being the largest producer accounting for approximately 78 percent and Russia the second largest with about 13 percent. The table indicates some small amounts of production in countries that do not have platinum reserves. Production in these countries is usually a by-product of processing other ores (usually nickel ore), which are in some cases imported in unprocessed condition, as well as recycling. Table 4.1: Global Platinum Production in 2007 by country Country Production in kgs Australia 130 Botswana 300 Canada  6,200 Colombia 1,400 Ethiopia 3 Finland 800 Japan 770 Poland 20 Russia 27,000 Serbia 1 South Africa  165,833 United States 3,860 Zimbabwe  5,300 Total 212,000 Source: United States Geological Survey, 2008. Table 4.2 breaks down the world's platinum production for 2007 by individual mines and some other sources. The main type of source is primary production from mining operations, accounting for about 78.3 percent. Some production can be attributed to recycling and the processing of imported ores in Japan. These types of production have 94 a share of 13.4 percent. Only about 8.3 percent of production cannot be attributed to a particular source. The Table indicates the name of the mine or other source, the country where the mine is located as well as the ownership and production value for 2007. Ownership could not be established for production derived from recycling and the processing of imported ores. Table 4.2: Global platinum production in 2007 by individual source Mine/Source Country Ownership Production in kg Mimosa Mine Zimbabwe Impala 2430 Zimplats Mine Zimbabwe Impala 3000 Unki (Gweru)* Zimbabwe Anglo Platinum 0 Rustenberg Section South Africa Anglo Platinum 20700 Amandelbult Section South Africa Anglo Platinum 17900 Mogalakwena South Africa Anglo Platinum 5090 Lebowa Platinum Mines South Africa Anglo Platinum 2900 Twickenham Phase I South Africa Anglo Platinum 289 Union Section South Africa Anglo Platinum (85%); Bakgatla-Ba- Kgafela Traditional Community (15%) 9600 Bafokeng-Rasimone Platinum Mine South Africa Royal Bafokeng Resources (Pty.) Ltd. (50%); Anglo Platinum (50%) 6000 Modikwa Platinum Mine** South Africa Anglo Platinum (50%); African Rainbow Minerals Ltd. (41.5%); Modikwa Communities (8.5%) 3660 Kroondal Platinum Mine*** South Africa Anglo Platinum 4050 Marikana Platinum Mine*** South Africa Anglo Platinum 722 Mototolo Platinum Mine South Africa XK Platinum Partnership (Xstrata 37% & Kagiso Platinum Venture 13 %); Anglo Platinum (50%) 2960 Western Limb Tailings Retreatment Plant South Africa Anglo Platinum 1410 Marula South Africa Impala (majority) 2022 Leeuwkop Project* South Africa Impala (74% via African Platinum plc (Afplats)) 0 Two Rivers Platinum South Africa African Rainbow Minerals Ltd. (55%); Impala (45%) 2370 Northam Platinum (all operations) South Africa Northam Platinum 6610 Lonmin (all operations excl. Pandora Joint Venture) South Africa Lonmin plc 19217 Pandora Joint Venture South Africa Lonmin (42.5%); Anglo Platinum (42.5%); Mvelaphanda Resources (7.5%); Bapo Ba Mogale tribe (7.5%) 2383 Everest Mine South Africa Aquarius 2940 Kroondal Platinum Mine*** South Africa Aquarius 8210 Marikana Platinum Mine*** South Africa Aquarius 2520 95 Mine/Source Country Ownership Production in kg Chromite Tailings Retreatment Plant South Africa Aquarius (50%); GB Mining and Exploration Ltd. (Johannesburg) (25%); Sylvania South Africa Ltd. (West Perth, Australia) (25%) 140 Nkomati Mine**** South Africa African Rainbow Minerals (50%); MMC Norilsk Nickel (50%) 1430 Crocodile River# South Africa Eastern Platinum 3480 Elandsfontein Mine## South Africa Xstrata 1110 Total operation Russia MMC Norilsk Nickel 22600 Lac des Iles Mine Canada North American Palladium  Ltd 760 Sudbury (Xstrata)### Canada Xstrata 1000 Sudbury (Vale Inco) Canada Vale Inco 4354 Tati Nickel Mine Botswana MMC Norilsk Nickel 280 Stillwater Mine USA Stillwater Mining Co. 2640 East Boulder Mine USA Stillwater Mining Co. 1210 Global recycling (excl. North America) 9700 North America recycling 18000 Japan (from imported Nickel Ore [likely from Australia]) 770 Other / unaccounted 17543 Total 212000 Source: United States Geological Survey (2008); company reports * mine under development ** African Rainbow Minerals reported 3860 kg production in 2007 (124,121 oz), Anglo Platinum reported 3660 kg (117,700 oz) *** Mine operated under "Pool and Share Agreement", production data represents only the share of the respective owner, not total production for the mine **** Nkomati is a Chrome and Nickel mine, which produces platinum as a by-product. Reports only indicate a production of 1430 kg of PGM, all of which was attributed to platinum in this table # reported as 1740 kg for six month ending Dec 31 2007, value doubled for approximate full year result ## The USGS reported Xstrata's combined production for Elandsfontein and Mototolo mines as 2590 kg of PGM. The value for Elandsfontein was derived by deducting XK Platinum Partnership's share from the Xstrata's total value for South Africa ### By-product of Nickel, no PGM data released, data are estimated based on total production for Canada and production at other two Canadian mines In the next step total production values at mine stage for each firm or production values for other sources were derived from the production values of individual mines or sources (see Table 4.3). If a mine was a joint venture or operated under a pool and share agreement, the production value was proportionately attributed to each owner according to the share in the joint venture. Platinum production was dominated by Anglo Platinum, 96 which accounted for almost one third of global production, and the three largest firms accounted for more than half.14  Table 4.3: Global platinum production by company Owner Total Production control in kg As % of total production Anglo Platinum 68544 32.33 MMC Norilsk Nickel 23595 11.13 Lonmin plc 20230 9.54 Recycling North America 18000 8.49 Other / unaccounted 17543 8.28 Aquarius 13740 6.48 Global recycling (excl. North America) 9700 4.58 Impala 8519 4.02 Northam Platinum 6610 3.12 Vale Inco 4354 2.05 African Rainbow Minerals Ltd. 3949 1.86 Stillwater Mining Co. 3850 1.82 Eastern Platinum 3480 1.64 Xstrata 3205 1.51 Royal Bafokeng Resources (Pty.) Ltd. 3000 1.42 Bakgatla-Ba-Kgafela Traditional Community 1440 0.68 Japan 770 0.36 North American Palladium  Ltd 760 0.36 Kagiso Platinum Venture 385 0.18 Modikwa Communities 311 0.15 Mvelaphanda Resources 179 0.08 Bapo Ba Mogale tribe 179 0.08 GB Mining and Exploration Ltd. 35 0.02 Sylvania South Africa Ltd. 35 0.02 Total* 212411 100.19 * error due to rounding Source: Author's analysis of United States Geological Survey (2008) data and company reports. The data in Table 4.3 can be broken down to attribute the global platinum production to shareholders in these firms. The production derived from recycling and unprocessed ore cannot be attributed to a specific firm and thus have to be added to the category "other/unaccounted". However, this means that 78 percent of global platinum production can still be attributed to a specific firm or organisation, because the majority is produced through mining. 14 According to Ericsson (2010) the three largest PGM producers account for 65 percent of production. The discrepancy is likely a result of Ericsson using 2008 data for all PGM, while the data used here are for Platinum only and for 2007. 97 Attributing each firm's production share proportionately to its owners allows for the establishment of a list of cities, where the owners of the capital that produces this particular natural resource are located. If a mining firm that produces 1000 kg of platinum is with five percent owned by an investment fund in New York control over 50 kg (5 % of 1000 kg) of platinum production is attributed to New York. If there are multiple owners in one city or one owner has stakes in multiple firms, the values are added up to arrive at a total value for a city. Usually the office locations or residences of the registered shareholders were used as the place to which to attribute control. However, in some cases the registered owners were joint stock companies and in turn owned by other firms, in which case the locations of owners was further broken down to a second level. For example, Anglo Platinum was owned with 79.64 percent by Anglo American. This means Anglo Platinum is practically a direct subsidiary of Anglo American, which has a registered head office in London. However, a large proportion of Anglo American's shareholders are located in South Africa and other parts of the world. Many owners in South Africa have transferred their voting rights to a nominee in Johannesburg. This means a large proportion of control is attributed to Johannesburg. There is obviously some ambiguity as to where decisions regarding a firm's operations are made and how much influence shareholders have. In general, if a firm has a majority (50 % plus one vote) shareholder, practically all important decisions are made by that shareholder. If a firm has one or more dominant shareholders (more than 25 percent, but no more than 50 %), this (these) shareholder(s) will have a significant say in any decision, but do not solely control the direction of the company. If there are no large shareholders in a company, significant decision making power usually lies with the board of directors, whereas the expectations on profit by shareholders still need to be met (see e.g. Herman, 1981). As a general rule control of a company was allocated to the next level of ownership, if the share was 50 percent or less. For example, Xstrata (headquarterd in Zug, Switzerland) holds a 24.9 percent share in Lonmin. This means that 24.9 percent of control over Lonmin's total production share was allocated to Zug, because decisions as to how to vote in Lonmin's shareholder meetings are likely made by the Xstrata board and management. At the same time, there is no known shareholder of Xstrata in Zug and subsequently no control over Xstrata's own production was actually allocated to Zug. Xstrata's main shareholder in 2007 was Glencore International AG in 98 Baar, Switzerland, with 34.5 percent and proportional control was allocated to Baar15. In the case of Anglo Platinum, which is held to approximately 80 percent by Anglo American and practically a direct subsidiary of Anglo American, control was not solely allocated to London, but proportionately attributed to Anglo American's owners. Table 4.4 shows the cities of command and control of the platinum industry. These are the world platinum cities. A little bit more than one third of control over platinum production could not be allocated to a specific place. This one third includes production the share of production listed as "other / unaccounted" in Table 4.2 (the share of global production for which the source could not be identified at all), any share attributed to recycling and production from imported unprocessed ore as well as well as mine production for which ownership could not be established. This means in turn that about two thirds of the locations of ownership over platinum production are known. The largest share of control is located in Johannesburg, accounting for almost one fifth of production, followed by London with about twelve percent and Moscow with almost ten percent. 15 The town centres of Zug and Baar are approximately two kilometres apart and within less than 50 kilometres of Zurich. The three towns/cities can be considered a functional entity. 99 Table 4.4: Urban control over platinum production for 2007 City Country Control in % Other / unknown Other 34.09 Johannesburg South Africa 19.42 London UK 12.4 Moscow Russia 9.61 General South Africa 6.7 New York USA 3.95 Zug Switzerland 2.71 Sao Paulo Brazil 2.05 Sydney Australia 1.52 Pretoria South Africa 1.3 General USA 0.76 Pilanesburg South Africa 0.68 Zurich Switzerland 0.64 General UK 0.54 Minneapolis USA 0.53 Baar Switzerland 0.52 Newport Beach USA 0.51 Los Angeles USA 0.47 Wayne, PA USA 0.44 Toronto Canada 0.36 Cape Town South Africa 0.20 General Luxembourg 0.20 Kagiso Communities South Africa 0.18 Modikwa Communities South Africa 0.15 Paris France 0.06 Boston USA 0.06 General Bermuda 0.04 General France 0.04 Austin USA 0.02 Perth Australia 0.02 Bellville, Western Cape South Africa 0.01 Source: Author's analysis of United States Geological Survey (2008) data and company reports. Two types of cities emerge from the list in Table 4.4. First, there are those cities that are close to places of production or represent an administrative centre in the country of production. These cities include the South African cities and communities, with Johannesburg dominating the landscape, Moscow and Toronto. Second, there are those cities that are far away from platinum reserves. Cities in this group are fairly diverse in size and location. The cities with the largest share in this group are London and New York, often referred to as the leading duo of world cities (Derudder and Taylor, 2005). Both represent primarily portfolio investment. New York and London are followed by Zug, Sao Paulo and Sydney, which mainly represent the headquarters of large mining firms that originally focused on other mining activities and now invest in the platinum 100 industry. Most other cities in the United States and Europe represent small portfolio investment. About one fifth of the control over the platinum industry is concentrated in Johannesburg, while all South African cities combined control more than one quarter, about 28 percent. On the one hand South Africa holds roughly three quarters of all Platinum reserves and South African cities may be considered under-represented in relation to the reserve locations. On the other hand, controlling more than a quarter of the production is substantial and indicates that capital is not completely de-nationalised and global. The dominance of Johannesburg within South Africa is likely overstated, because voting rights of individual and small shareholders at the JSE are now transferred by default to a nominee in Johannesburg, while the actual owners can be anywhere in South Africa or outside. Similarly, control over Russian production remains almost entirely in Moscow. The locus of control and command Establishing locational control over the platinum industry is a complex issue. The list of cities in Table 4.4 suggests a hybrid of control in places that control direct investment as well as portfolio investment. Moreover, control will be shaped by complex relationships between investors and the board and management of a company. Similarly, management structures within firms vary and relationships between management and labour will affect the locus of power. But, the way capital is commanded is in all likelihood not a matter left to the producer services in a dominant way. This section will analyse the views, perceptions and experiences of executives in the mining industry, mining consultants, unionists, activists, government officials and academics. These views were expressed in interviews conducted during 2009 in Johannesburg16. The focus of these interviews was the platinum industry. On October 22nd, 2009 Lonmin Plc released the following statement (RNS, 2009): Relocation of Operational Headquarters 16 The interviews were conducted from 13 May to 7 June, 2009 and 21 November to 6 December 2009. 101 Lonmin Plc ("Lonmin" or "the Company") announces that it intends to relocate its operational  headquarters  from  London  to  Johannesburg.  This  change  is designed to place the executive management team in a single location close to its  mining  operations  and  to  enhance  day-to-day  management  and communications. It will also enable the Company to engage more effectively with its South African stakeholders. As the first phase of this process the Company's CEO, Ian Farmer, will be based in Johannesburg from 1 January 2010 and will spend the significant majority of his time in South Africa from that point onwards. [...] The Board of Lonmin will remain based in the UK. To support the Board and the primary listing of the Company's shares on the London Stock Exchange a small office will be retained in London. Ian Farmer, CEO of Lonmin, said: "As part of Lonmin's continuing transformation, now is the right time to start the process of consolidating the executive management team in South Africa. This move  will  bring  many  benefits  as  we  continue  to  deliver  operational improvements and build on the firm foundations put in place over the last year." Roger Phillimore, Chairman of Lonmin, added: "This move demonstrates Lonmin's commitment to South Africa where it intends to continue to play a full part in the national transformation process. However I want to emphasise that Lonmin greatly values its UK domicile and primary listing in London and will continue to maintain a Board with the blend of backgrounds, skills and experience required to provide effective leadership appropriate for a FTSE100  company.  The  Board's  key  functions  of  management  oversight, strategic  direction,  decision  making  and  corporate  governance  will  remain  in London." The statement points to a very delicate issue in intra corporate power relations. On the one hand, a company requires close control over its operations to produce efficiently and satisfy shareholder demands for running a profitable company. As one executive of one of the large platinum producers put it with respect to running the day-to-day operations in his company: the decision making process is collaborative between London and Johannesburg, but "no matter how good your investments are, you have a balance sheet to attend to" and that comes down to the operations [Interview 20]. On the other hand, a company's management needs to maintain close relationships with its major investors and these are rarely located in close proximity to the mines, but at least somewhere else in a country and in the case of mining often in different countries and continents. Moreover, the location of a company can be considered an asset in itself, as a financial analyst for the platinum industry remarks with respect to Lonmin's registered 102 head office location: "The listing in London and the London base gave you a first class rating." He continued regarding the question, if the London base is it almost symbolic, considering the previous statement? "It is, but it is also cash, because of your rating" [Interview 19]. The delicacy of this relationship between investors and operations is clearly of concern to Lonmin's executive and especially clear in Roger Phillimore's statement above. There is strong sense within the mining industry that operations and corporate strategy are separate but at the same time two sides of the same coin. Operations have to be efficient to produce a profit for shareholders in form of dividends and share price appreciation. At the same time companies are almost constantly working to attract more capital through the flotation of new shares or other investments to turn money into capital and at a later stage into more money. There seems to be no generally accepted view of how the spatial division between operations and corporate strategy should be organised, but some proximity to the operations is clearly required. As another financial analyst for the industry remarked shortly after the announcement of the relocation of Lonmin's operational headquarters to Johannesburg: "That model with head office in London was never going to work" [Interview 23]. There is clearly a spatial division of labour and control within the industry and specific firms, which is partly shaped by corporate and national cultures. There is a disproportionate wealth of mineral resources under South African soil and this mineral wealth has in many ways shaped the country and in particular the city of Johannesburg, which only came into existence as a mining camp once gold was found on the Witwatersrand (see e.g. Beavon, 2004, p. xvii). The relationship between the mineral wealth and the capital required to access it, seems ambiguous, at times fragile and constantly shifting. One analyst remarked: There has always been "an uncomfortable relationship with capital from overseas" [Interview 19]. An independent consultant in the South African mining industry explained that because of the economic isolation and limited investment opportunities under apartheid, "an intense network of elites" had developed in South Africa [Interview 18]. Industrial mining activities, especially in South Africa with many very deep deposits, are almost always capital intensive and historically most of the capital in South Africa comes from Europe, especially Britain, since the beginnings of industrial mining 103 ([Interview 19]; see also Feinstein, 2005), although the availability of cheap labour provided by a racially segregated state before, during and after the apartheid years has certainly aided this type of industrial development. One executive in his 70s, who has been working in the South African mining industry for close to half a century, worked for a large London based company that split up into smaller companies in the mid 1960s. With reference to London, he said: "They called the shots". He continued to explain, why the company was split up and why South African assets were put under more local management: "When you are out of Africa, you are used to quite a bit of mayhem and these were British Gentlemen" [Interview 2]. Another mining executive confirmed with respect to international linkages that to this day most foreign directors of mining companies are British [Interview 5]. There seems almost general consensus that management of operation from a distance is problematic from a corporate management perspective. In response to a question about external control exercised from far-away places one executive, referring to former Lonmin CEO Brad Mills, who was located in London, remarked  that he caused the company an "enormous trauma", "he did not have the slightest clue of what was going on" and "he nearly killed the company" [Interview 2]. Another analyst with insider experience from Lonmin's Zimbabwe operations said that people who are in close control are not on the ground. Referring to the former Lonrho, Lonmin's predecessor company, chief executive, he said "Tiny Rowland was based in London and that's where the decisions are made" [Interview 22]. In addition to Lonmin, South Africa's second largest and the world's third largest PGM producer, the question regarding external control was often a subject of conversations regarding Anglo Platinum, the world's largest PGM producer. Again, at the centre of discussions was the relationship between London and Johannesburg. Anglo Platinum has a registered head office in downtown Johannesburg, but it is owned to approximately 80 percent by Anglo American, which was founded in South Africa in 1917. Anglo American moved its headquarters to London in 1999 after it merged completely with Luxembourg based Minoroco (Anglo American, 2011), in which it previously had a stake and which was considered Anglo American's overseas division during apartheid. The international ambitions of companies like Anglo American, were severely restricted during the apartheid era because of embargoes and exchange control. 104 Changes came with the advent of the first democratically elected government. An analyst views the situation as follows: 1994 broadened the international shareholder base. With reference to the apartheid years, he continued: "We had our own little world to take care of" [Interview 19]. Generally two views emerged during the interviews regarding the relationship between Anglo Platinum and Anglo American. The first view is that Anglo Platinum is under full control of Anglo American and all decisions are made in London. This view emerged primarily from people who were not affiliated with Anglo Platinum or Anglo American but had, in several cases, specific knowledge about the company. The other view was that there is a close cooperation between the two entities and that there is considerable autonomy in Johannesburg, with some guidance from London. This view was expressed mainly by people affiliated with or close to Anglo American or Anglo Platinum at the time the interview took place. There is also a sense amongst some that there has been a shift towards more centralised control: "Anglo Platinum has always been treated as a separate company, until recently" [Interview 18]. The same interviewee suggested that "all decisions are here, Jo'burg always the commercial capital", when he was talking about the historical context in a national framework at the beginning of the interview. The gold just kept coming and did not run out, so wealth remained in the city and hence decision making remained as well. However, when talking about current corporate structures, he made the following statement: "No major decision regarding the future of Anglo [American] or BHP is made here, none!" [Interview 18]. His view of Anglo Platinum was confirmed by a mining executive: "all decisions are taken in London", it's a division of Anglo American. It is a "definite migration" to London [Interview 5]. Another executive had a more critical view towards the management of AngloPlatinum: It is treated as a division of AngloAmerican, which is now led by "a lady from Kansas"17, who does not really understand the local conditions. He gave the following example: She demanded that the accident rate had to be decreased quickly, but that requires behavioural change, it is not a technical problem and thus a long term question [Interview 2]. Two statements of people close to Anglo Platinum suggest the relationship between Anglo American and Anglo Platinum is much more cooperative: Anglo 17 A reference to Anglo American's CEO Cynthia Carroll 105 American owns 80 % of Anglo Platinum, so when difficult decisions (e.g. retrenchments) are to be made, there is now a cooperation between London and Johannesburg [Interview 20]. Similarly, an executive with specific knowledge about Anglo Platinum said: "All our JVs [joint ventures] were negotiated in Johannesburg. Naturally we had to get approval from our major shareholder who is based in London. At the end of the day it is the board of directors who have the final say and in our case this is Johannesburg based" [Interview 21]. The spatial division of decision making processes also depends considerably on the culture within individual corporations. There are cultural differences within firms, states an analyst. For example, Impala Platinum has a small head office and one big dominant mine18. For meetings "the headquarter people" would go to the mine. At Anglo Platinum on the other hand there is a big head office in Johannesburg and for meetings people from the mines would come to the head office [Interview 18]. There is some indication that firms are moving towards smaller head offices. As a labour union representative remarked, the Harmony Gold headquarters have "maybe 35 people or so. [...] The bureaucracy from the past is gone" [Interview 24]. Similarly, an executive gave the example of Xstrata, which has a small "centre" (i.e. headquarters) and "Anglo" with lots of direction from headquarters [Interview 20]. This is a view confirmed by another analyst: "London will tell Johannesburg that we need this kind of production and productivity level and then Johannesburg will decide how to go about it." An exception to this would be Xstrata: "They empower their mine managers" [Interview 23]. In addition to corporate culture, individuals will influence the spatial division of decision making power. One observation states: Decision making in a company depends on the personality of the CEO and the board of directors [Interview 18]. In spite of the international nature of investment flows, national cultures seem to influence decision making processes. For example, American corporations expect total obedience from subsidiaries, says one executive [Interview 2]. While Platinum production is highly dominated by South Africa, an executive for marketing of platinum said that there is always something special about "the Russians"at PGM conventions. He further remarked: "There is always a mystique how Norilsk moves its product into the 18 In 2007 Impala had interests in two large platinum mines in South Africa and two in Zimbabwe. The statement may refer to an earlier situation. 106 market". This is a situation also influenced by national policies, according to the executive: Russia has large stockpiles [of platinum and other minerals], there is lots of curiosity. "You do your best to get a clear picture of how it works in Moscow" and Norilsk makes some effort to become more transparent [Interview 15]. Regardless of cultural aspects, the ownership of companies is of great influence, especially, if investors are located overseas. An analyst remarked in this context: The investors are in London and the decisions are made where the investors are, but the "Chinese have taken over quite a few mining properties" and BEE (black economic empowerment) requires the transfer of ownership to previously disadvantaged South Africans to 26 percent by 2014. However, "come 2014 that figure is probably revised downwards" [Interview 22]. International capital remains important. For example, Northam Platinum was "looking for investment" from London and Europe for its Booysendaal mine [Interview 14]. And, since the end of apartheid, "Anglo moved very much to the international side", but still has about 35-40 percent South African base (ownership) [Interview 19]. This trend is prevalent not only in the mining sector, says a mining executive: Large traditionally South African companies are now internationally owned, e.g. Standard Bank and Sasol are now maybe 40 percent South African owned [Interview 20]. The location of a stock exchange listing seems to still matter to some degree, as an executive said: It is a strategic decision where one is listed [Interview 20]. However, international capital can also be attracted through a single listing. Referring to the example of Anglo Platinum, which used to have an IDR [International Depository Receipt] listing in Luxembourg, it is now only JSE listed, although its shareholders are not all in South Africa. The fourth largest shareholder is Wellington in the US. [Interview 20]. There appears to be a general consensus that Johannesburg has a good financial system that is on a par with high-income economies. One executive explained that "we used to be listed at the LSE, but after I gave it a very hard look of what it cost us" the company is no longer listed at the LSE, only the JSE. Since there are improvements in communications, multiple listings really are not necessary any longer [Interview 2]. Decision making power also depends on the distribution of ownership. As outlined above, companies with a majority shareholder are under the control of this shareholder, while those with a dominant shareholder will be under considerable 107 influence of one or more dominant shareholders. There is generally a lot of direction from the board of directors in companies that have exclusively shareholders with stakes smaller than 25 percent. With reference to Anglo American, which does not have any majority or dominant shareholders, an executive remarked: the board "calls the shots", power lies with the board and the CEO. "Something really has to go wrong" for shareholders to vote against the direction of the board, it must be very serious issue (e.g. executive numeration), otherwise you cannot unite the shareholders [Interview 5]. One analyst sees very little involvement from shareholders in the running of the company: The shareholders show no interest. "The shareholders here are all passive". He further remarked that the STRATE (Share Transactions Totally Electronic) system in place at the JSE transfers the voting rights of shareholders by default to banks and brokers (i.e. nominees), which "de-materialises" the shares [Interview 18]. Another analyst finds that the actual operations of a company are not of much interest to shareholders. In response to the question, if there is any direct connection between shareholders and the mine, he replied: "If I am a portfolio manager in London, I am quite removed." For analysts and portfolio managers it will not go down to the mine level [Interview 19]. While the view that there is a lack of engagement between shareholders and the operations was expressed most often during interviews, one executive said: Shareholders used to be passive, but now there is a "medium level" of involvement [Interview 20]. There was also a lot of agreement that for most shareholders the financial result was the top priority in the running of a company and that companies cater primarily to their shareholders: With respect to the PGM industry in general, an analyst said: "Each time when they announce results, they have to go overseas" (from Johannesburg) [Interview 23]. A labour union representative took a critical view towards foreign investment: "Overseas investors will only stay around as long as there is money to be made" [Interview 24]. One executive had a similar view: Here in South Africa we rely a lot on portfolio investment and they are out in a second [Interview 2]. However, another executive sees a change in the industry: In the last few years lots of investors went only for the dividend, now there is more long-term investment [Interview 14]. Another labour union representative views the relationship between Foreign Direct Investment and Portfolio investment as follows: South Africa has not seen any sudden pull-out yet, but 108 the experience in Argentina and other countries shows, it is a real possibility. The reality is that South Africa is not experiencing real FDI, at least not at a scale to create jobs for "our" people [Interview 6]. Trade and consumption of platinum There is certainly a complex relationship between various places that constitute a web of control over the platinum industry. The focus in this chapter so far has been on the places of ownership of the mines and the mines themselves. The refining, consumption and trade of the metal plays a role as well, however. Most large mining companies of PGM have their own refineries, often located close to the mines. Some smaller firms, however, rely on the services of third parties. In such cases, a PGM matte (the concentrated metals from the ores) is being shipped to a refinery, which then separates the constituent metals and returns pure metal products, i.e. platinum, palladium etc. as well as other by-products, e.g., nickel or cobalt. In 1996 PGM matte from ore mined in Montana was exported to Belgium for refining and separation (United States Geological Survey, 2010b). Northam Platinum in South Africa used to refine its PGM matte at W.C. Heraeus in Hanau (near Frankfurt), Germany, but South African legislation required the company to have more beneficiation of mining products in South Africa and Northam and Heraeus have now entered into a partnership to refine the PGM matte in South Africa [Interview 15]. In 2009 demand for Platinum from jewellery applications exceeded demand from auto catalyst applications for the first time. Demand from jewellery was 3.01 million troy ounces (42.7 percent of total demand), while 2.23 million troy ounces (31.7 percent) were required in the auto catalyst industry. In 2008 the auto catalyst industry accounted for 3.655 million troy ounces (45.7 percent) and jewellery for 2.06 million troy ounces (25.8 percent) in demand. Other applications include the chemical, electrical, glass, medical, biomedical and petroleum sectors as well as demand from investment. Although platinum is mainly used in industrial application, demand for platinum as an investment has increased 44 fold from 15,000 troy ounces in 2005 to 660,000 troy ounces in 2009, while demand from all other sectors remained comparatively stable (Jollie, 2010). 109 China became the largest consumer of platinum in 2009 with 2.21. million troy ounces (1.41 oz. in 2008), followed by Japan with 1.25 million troy ounces (1.735 oz. in 2008), Europe with 1.83 million troy ounces (2.61 oz. in 2008) and North America with 0.86 million troy ounces (1.145 oz. in 2008). Demand from the rest of the word was 0.89 million troy ounces (1.09 oz. in 2008). While demand from auto catalyst applications decreased in all regions between 2008 and 2009, it still accounted for the largest share in demand in Europe, North America and the rest of the world. In Japan jewellery became the sector with the largest demand in 2009 after it has been slightly higher or lower than the auto catalyst sector between 2005 and 2008. In China jewellery was already the sector with by far greatest demand and it almost doubled from 1.06 million troy ounces in 2008 to 2.08 million troy ounces in 2009 (Jollie, 2010). Demand undoubtedly influences the platinum industry and many producers maintain sales offices in Europe, North America and Asia, but the focus in this chapter is on the ownership of productive assets. Although investment accounts for a relatively small proportion of demand, there is increased activity in this sector. Platinum is becoming an investment and considerable trade is conducted in London and Zurich (Jollie, 2010). Although relatively little platinum is sold on the spot market, the London Platinum and Palladium Market (LPPM) facilitates this type of trading and holds a fixing twice a day. Trading takes place in London and Zurich and delivery of platinum and palladium will be done to the vaults of the members of the LPPM, unless otherwise agreed (London Platinum and Palladium Market, 2009a, 2009b).  Interdependencies: operational and strategic control This chapter applies a mix of quantitative and qualitative data to trace the connections between the world cities of the platinum industry and the sites of capital investment in the periphery. The case of the platinum industry is of course unique and perhaps unusual. The natural deposits of PGM are highly concentrated in the South African Bushveld Complex and a small number of other locations. Moreover, the value of the product compared to its volume is higher than almost any other product that is produced on earth. It is roughly double the value of gold. Using the platinum industry as 110 an empirical example raises important questions regarding the capital commanding power of the producer services. It points towards a much more complex relationship between capital and those who command it than previously suggested by world city research. Although some APS are clearly involved in investments in the platinum industry, mainly through portfolio investments, a very large share of the industry is owned and controlled by dominant individuals, small scale private investors, other non-APS firms and governments, which can function fairly independently without substantial input from the APS. Furthermore, platinum prices are fairly dependent on the use value of the metal for industrial and jewellery production. Speculation plays only a minor role. It is a prime example of a natural resource produced under monopolistic conditions and this chapter identifies the cities where the control over this monopoly is located. By applying bottom-up world city research, Johannesburg, London and Moscow were identified as the dominant cities controlling the platinum industry by tracing the ownership of productive assets. While on the one hand Johannesburg, other South African cities and Moscow as administrative centres in platinum producing countries feature highly as places of control over the industry, places typically identified as world cities, mainly London, New York and the Swiss cities and towns Zug, Zurich and Baar, which can safely be considered part of the same city region, show also a strong presence. Ownership of more than half the industry can be traced to just five cities (including those collectively identified as 'General' in South Africa) and these are either in the same country, i.e. the same political territory, as the resource or identified as world cities by using producer services as an indicator. Such a pattern suggests a specific nexus between places of resource production and the network of world cities. Ownership is located either relatively near the resource, bounded by state territory or in the core of the world-system connected to the resource through the world city network. The only notable exceptions may be Sao Paulo and perhaps Sydney, which are in this control network over platinum production, because of their position in the natural resource industry in general, possibly articulating a sub-network of the resource industry that works relatively independently from the world city network articulated by producer services firms. Both cities are, however, also often included in the lower echelons of the world city hierarchy. The argument that the strength of a city is the multiplicity and robustness of its connections to other cities (e.g. Taylor, 2001, 2004a) explains to some degree the 111 presence of London, New York and Zurich in this list, even though they are not the most dominant cities in terms of ownership in the platinum industry. Assuming that every industry has its own sub-network of cities, such as the platinum industry, one would expect another set of cities to feature in these sub-networks, but if the leading world cities are really as well connected as previously suggested, and they may well be, these would appear in most lists for other industries together with those cities holding considerable power over whatever specific industry the list is based on. However, the point here is that cities like Johannesburg and Moscow are of great influence in an industry that is essential to a certain sector of the global economy, such as the automobile industry, some medical and other high-tech applications as well as jewellery. While Johannesburg and Moscow are not the cities where everything comes together, they are dominant in a certain sector and world cities like London and New York are dependent on their input and functioning to maintain their strength. While the world city network is certainly an expression of the international division of labour, the data emphasise that it is also an expression of the international division of control. The qualitative analysis in this chapter suggests that there is a spatial division between operational and strategic control within the largest companies in the platinum industry. Interviews conducted in Johannesburg pointed to a special relationship between Johannesburg and London and it often emerged that the strategic control over the two largest South African Platinum producers lies in London, more so than Johannesburg. Nonetheless, the need for a strong presence in Johannesburg, especially with respect to operational control, was evident for both Anglo Platinum as well as Lonmin, which recently relocated its operational headquarters to Johannesburg. Emphasis on the importance of London to gain access to capital in some companies suggests that London is needed as a switchboard of capital investments. Money is coming together in London, but it needs to be funneled into capital investments to turn it into more money, which is then returning to London. At the same time smaller, but still significant firms, concentrate both strategic and operational control in Johannesburg and have in fact unlisted from the London Stock Exchange (LSE) and are now only listed in Johannesburg. Also, with respect to the Platinum industry there is a certain dependence of London on Johannesburg to oversee, control and maintain the capital that was invested. This is indeed a two-way relationship. 112 In many ways Johannesburg functions as what has often been described as a gateway city (see Sassen, 2001, p. 173; Short, Breitbach, Buckman and Essex, 2000). However, the city is more than just a gateway, where money is funnelled into the resource hinterland (in this case the Platinum mines of the Bushveld Complex and the Zimbabwean Great Dike) through the city gate from one side and more money is leaving the gate after surplus has been extracted from the labour and environment in the opposite direction sometime later. Johannesburg itself is a centre of control and a switchboard of processes. It is a linking city, holding multiple connections and being a source of capital (measured here in terms of ownership) itself. Taylor, Walker, Catalano and Hoyler (2002) have identified seven different ways of measuring power differentials and Johannesburg emerged as both a minor regional command centre and a fairly low scoring high connectivity gateway. This taxonomy and categorisation of Johannesburg is perhaps an accurate description of its position in the network, but the network may be privileged in their analysis, which is entirely based on locational connections between the offices of producer services firms. The power Johannesburg enjoys is partly bestowed on it by the bounded territory of the South African state and the ownership in the mining industry held in the city, rather than exclusively the network. The producer services network does not pay adequate attention to ownership, which is highly concentrated in cities not considered at the top of the world city hierarchy. Chapters 3 and 4 identified a number of semi-peripheral cities as major centres of ownership control over non-fuel mineral resources in general and platinum specifically. The network power is at least to some degree dependent on the input from these cities. Taylor, Walker, Catalano and Hoyler (2002, p. 239) argue that "we have to be careful not to give precedence to the transparent – 'command and control centres' – over more opaque network power", but the network power may have been given unjustified precedence in recent years. It has paid virtually no attention to actual production and privileged the sectors concerned with servicing (sometimes fictitious) capital. There is a particular interrelationship between a territory (the state), a place (the city), scale and the network, constituting power in commanding capital (see Jessop, Brenner and Jones, 2008 for a discussion of territory, place, scale and network (TPSN) as analytical frameworks for geographic research). 113 In his description of imperial London Anthony King (1990, pp. 6-7) takes a historical route to analysing urban development. While the influence of the colonisers on colonial cities is well known, he argues that “[i]t was from among this colonial urban system [cities, especially externally oriented port cities, in the colonies] that a majority of today's 'world cities' were to develop”. In other words, imperial London was built on the riches of other cities, which London accessed through its external relations or in Taylor's (2004a, p. 1) words, its "second nature". Thus development in the cities of the west or the metropolitan core cannot be understood outside the context of the colonial periphery. More recently, urban development in world cities around the globe needs to be understood in this context. The data in this chapter have identified considerable ownership power in Johannesburg. At the same time, it is conceivable that cities like London and New York still derive their riches from peripheral cities or in Andre Gunder Frank's (1972) words from their satellites. While there is clearly a two-way relationship between London and Johannesburg, the multiplicity and diversity of connections that cities like London and New York maintain give them the flexibility to maintain a powerful status should a specific economic venture fail while others are economically vibrant. World city research, using producer services, probably identified these power relations correctly, but perhaps overstated the network power and did not adequately explore the interdependencies between places. Moreover, in focusing on the producer services the world cities literature has paid too little attention to the power of territory, place and scale. States often remain in tight control of their territory and there are powerful connections at smaller scales, as for example smaller independent networks within the resource economy or other economic sectors and spheres of life. Identifying interdependencies through vertical world city research Using a bottom-up approach to researching command and control functions in a world city network and tracing vertical connections, made it possible to identify direct control over capital rather than using the proxy of the producer services. The data identified a high concentration of ownership power over platinum production in cities in 114 the semi-periphery, which suggests a different distribution of command and control functions in the administration of the platinum monopoly compared to a world city network articulated through producer services. A very similar pattern has been identified in chapter three. The pattern identified through the empirical example of the platinum industry suggests that wealth and control functions are now relatively dispersed across the globe and the world is no longer strictly divided into a metropolitan core and a resource periphery. Cities like Johannesburg and Sao Paulo hold considerable wealth and control over natural resources. Similar studies in other sectors are likely to identify Mumbai and Delhi or Shanghai and Beijing as holding considerable ownership control, for example in various manufacturing sectors, including the automobile industry, and software development. World cities like London or New York derive their power position through the diversity of sectors they are involved in and their high levels of connectivity. They have the "power to" service global capital (see Tayor, Walker, Catalano and Hoyler, 2002), but just as imperial London was dependent on its colonial ties, the contemporary world city derives its power through inputs from elsewhere. Vertical world city research provides the means to identify where the input comes from and identify those cities that have ownership power or "power over" specific productive assets. Moreover, it also provides the means to identify, where the production of wealth originates through the extraction of surplus value from labour and environment and how these places are linked into the world city network.  The control over the platinum industry is quite concentrated in a few places, although these are by no means all in the global north as might have been expected, and production is under monopolistic control of a few companies, which generate profits by extracting surplus value from labour and the earth. Although the example of the platinum industry is a very specific one, some generalisations are possible, assuming that other sectors, especially those under monopolistic control, show similar patterns. Further bottom-up approaches and vertical world city research of other sectors can contribute to identifying a pattern of control in the world-system. Moreover, the distribution of ownership control in the non-fuel mineral industry considered in chapter 3 and the platinum industry suggests, as Amin (1994, p. 88) notes, that we are moving towards a world that is closer to that of Marx’ times than that of the Keynesian / Fordist era, i.e., 115 polarisation is occurring at a global scale but is dispersed to a multiplicity of places. The global spatial division of labour and control is likely no longer a matter of metropolitan core and resource periphery divided into large territories, but a matter of a highly connected network of cities with ownership control on one side and peripheral surroundings on the other. The pattern points to a world shaped more like a mosaic with peripheral and core functions literally adjacent to each other in cities around the world, rather than a two-sided division of territory. 116 Part III 117 Chapter 5: The Wealth of World Cities: Nation States Make Cities The Anglo  American  Corporation  struck  a  clever  blow for  capitalism when  it became the first to break the taboo against talks with the ANC. In September 1985 [...]  its chairman met with Congress leaders deep in the Zambian bush. Anglo thus legitimized contact with the movement, which had been shunned by the  South  African  political  and  corporate  establishments  as  well  as  by  most Western governments. The meeting proved a major diplomatic coup for the ANC. After waiting for decades for salvation from Eastern Europe, the ANC began to espy allies amongst their long-term enemies, the capitalists. (Waldmeir, 1997, p. 73) Perhaps it is nowhere more intriguing to study the entanglement of politics and economics than in South Africa. The support of white owned industry through artificially cheap black labour during the apartheid era in South Africa is quite well known and an obvious form of support for industry by a racist government. The post liberation entanglements are certainly not as obvious or black and white. Johannesburg has affirmed its position as the leading city in southern Africa and perhaps all of Africa south of the Sahara since the end of apartheid. It is dominant within South Africa. It is home to 74 per cent of national corporate headquarters and 55 percent of the total office space in the country. Some specialised sectors cluster outside Johannesburg, as for example the oil industry in Cape Town, but there is little doubt on the overall leading position of Johannesburg in the country and the region (Mabin, 2007, p. 49-50). While Johannesburg may be dominant in southern Africa it also retains a peculiar and perhaps dependent position in relation to other world cities, particularly London. There are two aspects of world city research that I examine in this chapter with a particular focus on the city of Johannesburg. First, the relationship between the nation state and the creation of world cities is analysed. Following Brenner (2004, p. 3), the argument is that  nation states have a profound interest in positioning their cities as world cities and that states actively implement policies that position their urban economies optimally in global circuits of capital. I am arguing that in the case of Johannesburg this process began before the South African state existed in its current form, through the grooming of the liberation movement and ANC leadership for leading a capitalist state and ensuring the abandonment of policies of nationalisation of the most important national economic 118 sectors. Subsequent liberalisation policies have allowed Johannesburg to assume its current position as a leading urban centre in southern Africa.  Second, Johannesburg still functions as a satellite or subsidiary city to other world cities, especially London. Moreover, by tracing the development of two leading South African mining houses, I will argue that much of the wealth that produced London and its world city status was created in the global periphery and channelled through Johannesburg to the British capital. Johannesburg's positionality in relationship to London as well as the mines gives it the status of a gateway (e.g. Sassen, 2001, p. 173) or linking city. This chapter will begin with a very brief overview of the political economy of the transition period in South Africa (for a detailed account of the transition see e.g. Bond, 2000b), followed by a selective description of specific events that led to the transition to majority rule in South Africa from the mid-1980s onwards. As the above quote suggests a close link between the South African business community and the ANC was started some time before the end of apartheid. This link remained strong throughout the transition to democracy and beyond. The next section describes the relocation of two of the largest South African firms to London, and I subsequently analyse this move in the context of the world city concept, before completing the chapter with a brief conclusion. Transition in South Africa Racial segregation in South Africa precedes the formal establishment of apartheid in 1948. This system was at least partly shaped by the mines' need for cheap, relatively unskilled labour in deep level mining. Rural black people were forced into wage labour due to taxation by the state and there was a system of labour migration to the mines. The apartheid system formalised existing segregation and put it under close and brutally enforced state control, but even the Truth and Reconciliation Commission concluded that the 'blueprint' for apartheid was provided by the mines. It was not an 'Afrikaner state innovation' (Hamann, 2009, 436; see also Feinstein, 2005). Although the system was sustained for many years, by the 1980s apartheid had run aground as massive protest and unrest, especially in the townships, and international 119 pressure brought the system to its knees. At the same time, South African businesses could no longer grow substantially within the system that had sustained them comfortably for so many years. Because of sanctions and exchange controls South African firms could not re-invest their profits overseas, which severely depressed their profit margins and thus shareholder value. The market capitalisation of large firms in most countries exceeds the value of their net assets. In 1995 Anglo American, which was at the time still a South African company, had a market capitalisation 22 percent lower than the value of its net assets (Carmody, 2002, p. 263). Similarly, while by the 1990s US firms earned 30 percent of their profits overseas, South African firms earned only seven percent overseas at the same time. As a result large corporations in South Africa increasingly pressed for a negotiated settlement with the liberation movement, the ANC (Carmody, 2002, p. 262). In addition to the internal pressure and relatively newfound opposition to apartheid by western powers, some major events took place on the international political stage that were certainly not inconsequential, even if not solely responsible for the transition to democracy in South Africa. The collapse of communism in Eastern Europe and the demise of the Soviet Union "crucially affected political leaders' domestic choices and thereby conditioned the nature of societal transitions" (Habib, Pillay & Desai, 1998, p. 110). The ANC had received support from the Soviet Union during the Cold War, but the end of the existence of the communist super power changed the political considerations for both the apartheid regime and the liberation movement. On the one hand the fear of South Africa being "overrun" by communists ceased within the South African establishment, while on the other hand the ANC did not need to showcase its allegiance to the Soviet Union any longer and in fact needed to orient itself towards new partners. The business community inside South Africa also got involved in attempts to stabilise the state, especially after the June 1976 Soweto uprising. Under the leadership of Harry Oppenheimer of Anglo American and Anton Rupert of the Rembrandt group, corporations established the Urban Foundation, which provided urban development initiatives, especially in the townships (Hamann, 2009, p. 436). Business thus undertook typical state functions outside any democratic control, because the state was not providing any of these. It also pacified a population that had to live with the consequences of the corporations' prosperity, because these people were forced into providing their labour at artificially low rates. 120 Businesses, especially the handful of large South African conglomerates that dominated the economy, had gained considerable political leverage with both the National Party government and the liberation movement. The four largest conglomerates controlled 83 percent of the Johannesburg Stock Exchnage (JSE) in 1988 (Fine and Rustomjee, 1996, p. 103; see also Carmody, 2002, p. 256).When the ANC came to power in 1994, it had little power over the civil service and the security forces, which were still staffed by apartheid appointees (Jordan, 2004, p. 206). So, the ANC was interested in keeping relative stability in South Africa, by catering to the elites that had much to lose. Civil servants appointed under apartheid remained in office. Perhaps more importantly though, the ANC carefully catered to the business community and adopted a neo-liberal line. A signal toward the local corporate elite and international investors was the re- appointment of Derek Keys as finance minister, who had already served in the last National Party government (Jordan, 2004, p. 206). The relationship between the ANC and the business community, foremost Anglo American, had started several years earlier, however. Sowing the seeds of the government-business nexus By the 1980s South African capital was at the brink of bursting out of its confines. Due to exchange controls internally and sanctions externally, South African firms had few places to go other than the domestic market and so firms like Anglo American invested in businesses unrelated to their core interests, such as supermarkets or newspapers. Anglo's web of companies accounted for 30 percent of South Africa's gross domestic product (Jamieson, 1990, p.2) and it directly or indirectly controlled about 50 percent of the Johannesburg Stock Exchange (Lieberfeld, 2002, p. 360; Cox and Rogerson, 1985, p. 223) during this time. At the same time a sense that apartheid could not be sustained took hold in large sections of white South Africa and the ANC seems to have shifted its strategic direction towards negotiations rather than guerilla warfare (see Waldmeir, 1997 for a documentation of the negotiations leading to the end of apartheid). When a small group of white business people, led by Anglo American chairman Gavin Relly, travelled to Zambia in September 1985 to meet a group of the most senior 121 ANC leadership, including ANC president Oliver Tambo and his political secretary and speech writer Thabo Mbeki, it may have looked liked an odd arrangement, but the meeting was likely foundational -- albeit not instrumental, perhaps -- for many neo- liberal arrangements in place in present day South Africa. This mutual exploration of the other side by the two groups was facilitated by Zambian president Kenneth Kaunda at his presidential lodge in Mfuwe. The place was certainly no world city, but for a brief moment a place where the future of South African capital was carefully considered and the future role of policy makers tested by the owners of capital. Although Anglo American patriarch Harry Oppenheimer did not support the meeting, the businessmen travelled with journalists of The Sunday Times, which Anglo part owned, and the meeting was well publicised inside South Africa (Waldmeir, 1997, p. 73; Lieberfeld, 2002, p.360). The meeting took place just a few weeks after South African President P.W. Botha indicated a continuation of a hard line and warned external forces to stay out of South African affairs in his so called Rubicon Speech, which was widely expected to signal a change of policy in South Africa, rather than the opposite. In hindsight, the meeting in Mfuwe was a logical step for both sides. On the one hand, business leaders understood that they could not continue under the current arrangements of relative isolation in a world of increasing global capital, financial and commodity flows.19 On the other hand, the ANC leadership had lived in exile for decades and many of their comrades were imprisoned in South Africa, so that they had an interest in negotiating to resolve the stalemate. The business people were in a position of power inside South Africa, but outside the National Party government. Gavin Relly described his impression of the ANC side as follows: "They were just a bunch of homesick South Africans". (Waldmeir, 1997, p. 74). Gavin Relly was familiar with Kaunda from when he directed Anglo's operations in Zambia, which were partly nationalised around the same time. Another member of the group of business people, Zac de Beer, also knew Kaunda from his time as head of Anglo American in Zambia (Lieberfeld, 2002, p. 360). Facilitated by Kaunda, who was instrumental in nationalising Anglo's assets in Zambia, the business people tested the 19 Although the isolation was never complete, for example one of the participants in the meeting represented the company Premier Milling (also partly owned by Anglo American), which exported flour and cooking oil to Zambia, a Frontline state (Lieberfeld, 2002, p. 360). 122 ground with the ANC on its economic policy and the question, if it would nationalise the mines in a future ANC led South Africa (Lieberfeld, 2002, p. 361). The relationship between the ANC, which by the 1980s had its headquarters in the Zambian capital Lusaka, and the Zambia office of Anglo American was already tight. The organisation used Anglo's post box for sensitive mail and had parties at the home of Anglo's managing director in Zambia (Waldmeir, 1997, p. 74). However, the meeting in Mfuwe was likely the most instrumental event in starting the relationship between business people and the liberation movement. In many ways liberation and liberalisation were closely related interests as the National Party apartheid government stood in the way of both. The meeting was risky for both sides. Oliver Tambo almost did not attend, because he feared a backlash from the South African Communist Party (SACP) and the military wing Umkhonto we Sizwe (MK) camps within the ANC. He feared accusations that he was selling out to business and was only persuaded by Kaunda to attend. Tambo had received a limited mandate to negotiate from an ANC congress three months before the Mfuwe meeting. However, it turns out that the communists' concerns from within the ANC that the main objective of the business people was to secure functioning of capitalism in a post-apartheid South Africa was right on the mark. For the business people attendance was risky, because there was no assurance that they would not be ostracised from the white establishment. In fact President Botha had successfully persuaded representatives of the Rembrandt Group, Barclays, Barlow Rand and Sanlam not to participate and in the end only Anglo American group firms were represented. Nonetheless, their position of power allowed them to engage in talks with the ANC leadership and the government did not actively intervene (Waldmeir, 1997, p. 73; Lieberfeld, 2002, p. 361; Pfister, 2003, p. 6). Anglo American had demonstrated with its move the flexibility of capital to do what is best for capital. In spite of clear differences and expectations to the contrary at least by the businessmen, the Mfuwe meeting took place in a very friendly, almost comradely atmosphere, with the use of first names and consumption of alcoholic drinks (Waldmeir, 1997, p. 74; Lieberfeld, 2002, p. 361-362). Several other important meetings between the ANC and a wide cross-section of white South African society followed the September 1985 encounter (see e.g. Lieberfeld, 2002; Pfister, 2003), but the first and perhaps most ground breaking step from inside South Africa was made by business. 123 First grow big, then get out: the case of Anglo American The days of apartheid were indeed numbered by September 1985 and the pressure for reform on the regime grew. By the beginning of 1990 P.W. Botha had left office and the ANC was unbanned. At the same time businesses expected a re-integration of South Africa into the world economy. Large firms prepared to streamline their businesses and some of the biggest players were ambitious to expand beyond the national confines of South Africa. Generally, businesses would divest their non-core interests and try to acquire new assets that match their core portfolios, often overseas. By the year 2000, some of the largest companies managed to shift their primary listings to stock exchanges outside of South Africa, in fact all to the London Stock Exchange. South African Breweries, the insurance company Old Mutual, the networking and communications company Dimension Data and the mining houses Gencor (through its subsidiary Billiton) and the Anglo American Corporation received permission from the South African Reserve Bank (SARB) to list overseas. All five firms retain a secondary listing at the JSE (Walters and Prinsloo, 2002, pp. 62-63), but only Dimension Data kept its head office in Johannesburg, while the other firms moved to London (Padayachee, 2003, p. 197). I will focus in this section on the two mining houses and their transition from national firms with diverse interests to global diversified mining conglomerates. To this day the impact of Anglo American on the South African economy is hard to ignore, although it has certainly lost the dominant grip on the economy that it had throughout the 1980s. The company no longer has stakes in supermarkets, newspapers, food and beverage industries and other non-mining related businesses, but its geographic reach stretches far beyond South Africa. It has dramatically expanded its international business, which it had started in the 1950s through a complicated network of companies (Innes, 1984, pp. 233-236; see also chapter 3). Anglo American was founded relatively late in comparison to other large South African mining houses, most of which date their origins back to the late 19th century. Initially much of the required capital came from overseas, especially from Europe and Britain in particular, but with considerable German financing as well. However, the profits made from South African mining, especially 124 diamonds and gold were so huge that subsequent capital needs could be met by internal company sources. Anglo American founder Ernest Oppenheimer, a German Jew, arrived in South Africa in 1903 to work in the diamond industry. Together with the US bank Morgan Guaranty, the US firm Newmont Mining and the National Bank (South Africa), he established Anglo American in 1917 (Padayachee, 2003, pp. 183-186). Given his background and American involvement in Anglo American, Oppenheimer had to continuously pledge his loyalty to South Africa. He became a member of Prime Minister Jan Smuts's South Africa Party and won a seat in parliament for the party in 1921. By the late 1930s Oppenheimer had expanded his operations significantly across southern Africa. He managed to take complete control of the diamond mining company De Beers and the London-based diamond syndicate (Padayachee, 2003, p. 186). Oppenheimer also tried to gain full control over Zambian (northern Rhodesia) copper production in the 1920s, but failed because copper prices plummeted during the great depression (Jamieson, 1990, p. 114). Anglo American failed in another attempt to establish a monopoly and it had to diversify its interests across the mining business and far beyond. However, it maintained monopolistic control over diamonds through De Beers, which remained a separate company, but under full control of Anglo. A significant part of Anglo's profits and basis for expansion were thus derived from monopoly capitalism (see Innes, 1984 for an extensive discussion of monopoly capitalism in the context of Anglo American and South Africa). From the 1920s to the 1940s Anglo American built up new investments in southern Africa, but in the 1950s the company focused mainly on the development of gold mines inside South Africa. In 1958, however, Anglo managed to take over a major mining company in South Africa through the American based Engelhard Group, which it controlled. This gave Anglo American a strong presence in America and money flowed from South Africa to the United States. By 1964 Anglo merged some of its group companies into the London based firm Charter Consolidated (Charter), a major base from which it grew its international interests (Innes, 1984, 233-235). The other major basing point to gain an international foothold came from an unlikely source: the Zambian government. In spite of its failed attempt to gain monopoly control over Zambian copper, Anglo American continued to hold large interests in the industry through Zambia Anglo American (ZAMANGLO). In the process of 125 nationalisation after independence, the Zambian government of Kenneth Kaunda bought 51 percent of ZAMANGLO and paid Anglo American for it in full in American dollars. The proceeds from this deal were the basis for the Minerals and Resource corporation (Minorco), which was registered in Hamilton, Bermuda (Innes, 1984, p. 235; Jamieson, 1990. p. 114). Initially Minorco lost Anglo American considerable amounts of money. It was registered in Bermuda, but controlled from Anglo's headquarters in Johannesburg. By 1987 Minorco had become a financially viable company, in large part through the buying and selling of shares in the American bank Salomon Brothers. In the same year Minorco moved its head office to Luxembourg, partly because of political pressures in Bermuda (Jamieson, 1990, pp. 115-117). While Charter concentrated on businesses in Africa, Europe, Asia and Australia, Minorco was most active in the Americas. By the early 1980s, it became the largest foreign investor in the United States by revenue (US$ 26 570 million), ahead of major corporations such as Royal Dutch Shell Group (US$ 19 830 million) and BP (US$ 11 023 million) (Innes, 1984, p. 235-236). One of the seemingly odd characteristics of the gigantic web of firms within the Anglo American Empire20 was that there were few controlling shareholdings amongst the ten largest firms in the group. For example Anglo American itself held only 34 percent in Charter and 45 percent in Minorco, but Charter held another 23 percent in Minorco and -- most interestingly -- ten percent in its part owner Anglo American. A similar construct existed with group firms such as De Beers, the Anglo American Investment Trust and Anglo American Gold (Innes, 1984, p. 274). The web, the flexibility inherent in it and the diversity were the strength of the company and it was controlled from downtown Johannesburg. In the 1990s Anglo American started to focus its interests once again on the mining business and to disentangle its multiple cross ownerships within the group. In order to do this minorities in some of the core companies in the group, including Minorco, were bought out. Moreover, De Beers and the Oppenheimer family sold their shares in the most important mining and industrial companies in the group to Anglo American, so that Anglo American became both an operating mining company and a holding company, but the complex web of cross ownership linkages were disentangled 20 Innes (1984) lists 656 companies in the Anglo American Group for 1976. The list is still not comprehensive. 126 (Gleason and Levin, 1999a). In October 1998 the decision to merge the Anglo American Corporation and Minorco was announced. The merger created Anglo American plc, which would have its primary listing and head office in London from 1999 on. The relocation was approved by the South African Reserve Bank (Walters and Prinsloo, 2002, p. 62). Anglo American is now a multinational corporation registered in the United Kingdom, with an approximately 40 percent South African shareholder base (see chapters 3 and 4). The story of Anglo American is a peculiar one. As Innes (1984, p. 236) writes, the history of Anglo's subsidiary Charter being the largest investor in the United States suggests "dire implications for classical theories of imperialism, which assumes a one- way flow of investment (export of capital) from the capitalist centre to the periphery". The case of Anglo American, which by the 1980s had exported substantial amounts of capital to the heartland of the global economic core, the United States, is of course contrary to the theory. However, what Innes could not foresee at the time of writing in 1984 is that the story would change dramatically in the late 1990s when Anglo American, a symbol of South African capitalism, would switch its stock exchange listing and registered head office to London, UK. The Anglo American empire was once controlled from downtown Johannesburg, but once the protection and isolation of the South African state disappeared it left Johannesburg for the seemingly greener pastures of the world city London. The rise and exit of Afrikaner capital from South Africa The second case that warrants careful attention is the company that was known for a period of about 20 years as Gencor. It, too, was one of the large South African mining houses albeit not nearly as dominant as Anglo American. Its origins go back to the 1950s when the National Party government actively supported the creation of Afrikaner controlled capital to counteract the British South African economic dominance in the country. Gencor was incorporated in 1980 through the merger of the General Mining and Finance corporation and Union Corporation (Fine and Rustomjee, 1996, p. 127 173; Padayachee, 2003, p. 195).21 Similarly to Anglo American it was a large conglomerate of companies that expanded far beyond its core mining interests. At the advent of democracy in South Africa Gencor decided to divest its non-core interests and sold assets worth about one third of its market capitalisation to shrink its value from 21 billion South African Rand to 14 billion. Amongst the divested companies were some very large firms, including Sappi (forest products, 38.6 %), Engen (oil refining and distribution, 49.9 %), Genbel (investments and finance, 49.9 %) and Malbak (consumer conglomerate, 49 %). Gencor retained all its interests in the mining business (Gleason and Levin, 1999b). Gencor's management had plans to expand the company outside of South Africa, but focusing on mining. Around the same time Royal Dutch Shell was interested in selling its mineral assets subsidiary Billiton in order to focus on its oil business. From early 1993 until late 1994 Gencor and Royal Dutch Shell negotiated a deal that allowed Gencor to buy Billiton, although it did not actually have the cash reserves to do so. The deal was a complicated financial arrangement. Due to capital control Gencor was limited in its ability to take money out of South Africa, but it had about 300 million US dollars in overseas assets through TransAtlantic Holding, a company dating back to the 1950s, when Union Corporation was active in the London property market. However, Royal Dutch Shell asked for 1.8 billion US dollars and an attempt to partner with a Canadian mining company in the takeover failed. In the end Gencor bought Billiton from Shell for 1.14 billion US dollars, financing the purchase through its overseas cash reserves, bank loans and bonds from Shell. The South African mining house Gencor with its origins as an Afrikaner capital company now had multinational interests with operations in 15 countries (Jones, 1995, pp. 238-239). Between 1994 and 1997 the company consolidated some of its interests in the South African mining industry, while selling other assets. The most important step in the transformation of Gencor was when the company was split into two companies again: Gencor and Billiton, with identical shareholders at the time of the initial split (Gleason and Levin, 1999b). Most of Gencor's non-precious metal interests (in South Africa, Australia, Mozambique, Columbia, Brazil, Suriname, Canada and the United States) were 21The company was officially named General Mining Union Corporation Limited (Gencor) in 1980, but changed its name to just Gencor Ltd in 1989 (Jones, 1995, pp. 262-23). 128 transferred to Billiton and the company was listed on the London Stock Exchange in 1997 with a secondary listing in Johannesburg (Padayachee, 2003, p. 195; Walters and Prinsloo, 2002, p. 62). What had been a European subsidiary of a South African company for three years, was suddenly a European company again. This company had a large South African shareholder base, however. After the split in 1997 Gencor's only major assets were precious metals interests. The company merged its gold interests with Gold Fields of South Africa in 1998. At the same time it sold its platinum rights to Impala Platinum in exchange for Impala shares. As a result, Gencor went from being a major mining house to a holding company with a 47.4 percent stake in Impala and an 18.7 percent stake in Gold Fields. Its staff was reduced to five (Gleason and Levin, 1999b). In 2003 Gencor's only asset was its stake in Impala, which it transferred to its shareholders. It was deleted from the FTSE100 index in May that year (FTSE, 2003) and practically ceased to exist. In 2001 Billiton merged with the Australian natural resources company BHP to form BHP Billiton, one of the world's largest diversified mining companies. Although the UK listed arm of BHP Billiton retains a large shareholder base in South Africa (at least 24.4 %, see Table 3.3) and has a secondary listing at the JSE, its story is almost identical to that of Anglo American in terms of growing under the protection of the South African apartheid state and then leaving its domicile in Johannesburg to move to London. Fattening the world city from afar It is one of the main arguments of this dissertation that conceptions of power in world city research need to be re-examined in light of the current dominance of producer services firms as world city network articulators. Chapters 3 and 4 explore the control hierarchy of the natural resources industry and identify some cities not usually identified as top-ranking world cities as the main control centres of natural resources. The cases of Anglo American's move to London and Gencor's self-induced dismantling and transfer of assets overseas provide some background in explaining why Johannesburg is identified as a major place of control over natural resources, while at the same time it usually only appears in the lower rungs of world city rankings. Because Anglo American and BHP 129 Billiton have their origin, at least partly, in South Africa, many shareholders continue to have their domicile in Johannesburg. However, control over monopolistic resource production within the territorial confines of South Africa and to some degree the larger region of southern Africa, is shared between Johannesburg, London and other cities. There is increasingly monopolistic control over the global resource base by a few companies, whose operations truly span the globe. Their business is focused, but their operations are flexible. Bearing in mind Braudel's assertion (see Taylor, 2000, pp. 7-10) that the most successful capitalists are those who remain flexible and can switch their assets quickly, if new opportunities for bigger profits and greater accumulation occur, the spatial flexibility and geographic spread of assets (i.e. mines and mineral rights), gives these large mining companies a powerful standing in the world economy. At the same time the assets of these companies are elementary to the functioning of the global economy: The construction and high-tech sectors would fail without them. So would the automobile and shipping industries. The places from which these conglomerates are controlled are, by definition, command and control centres -- even if the network may appear more complicated than the more familiar London-New York-Tokyo triad. There is little doubt that the relocation to London of South African companies has transferred considerable decision making power over productive assets from Johannesburg to the imperial centre of the British empire and a contemporary world city. What is perhaps most important for the argument brought forward in this thesis and understanding contemporary world cities is that the foundation of the wealth, which is now controlled in London, was created in the South African mines under the protection of a racist state. The two cases of Anglo and Gencor/Billiton are peculiar to South Africa, but some generalisations can be drawn from them. Padayachee (2003), building on Sklair and Robbins (2002 [cited as Sklaar and Robbins, 2002]), points to a large gap in the literature regarding the characteristics of third world transnational corporations. Most large transnational corporations are registered in countries of the first world, but a small number of firms are domiciled in third world countries. In the years from 1957 to 2001 between one and eleven percent of all Fortune Global Corporations were located in the third world. The 1980s recorded the highest proportion of third world corporations in this list. While until 1990 more than two 130 thirds of third world corporations in the Fortune list were oil, mining and metals companies, the proportion had dropped to around one third from 1995 onwards (Sklair and Robbins, 2002, pp. 86, 89). While even eleven percent is a small proportion the role of large transnational corporations in the third world may not be insignificant in the formation of a world city network that facilitates control over productive assets around the world. My primary concern is not to find evidence that would suggest that cities in the global south are economically as powerful, strong and well connected as the world cities of the global north, such as London, New York and Tokyo. My concern is more about establishing and systematically investigating the links from the control and command centres of the world economy to the places of production, where the wealth is created that is -- presumably -- controlled through the network of world cities. My argument is that large cities in the global south facilitate the extraction of surplus value from labour and environment in the global south, while maintaining some control and command functions. At the same time, many decisions regarding productive assets in the global south are made elsewhere and control through ownership over these assets is located in distant places that are connected through the network of world cities. This network also facilitates uneven development through the transfer of profits (paid out as dividends to shareholders). The role of large transnational corporations in the global south is certainly not insignificant in this context. Building on Frank's (1969, p. 6) conceptualisation of a metropolis-satellite structure, I characterise the city of Johannesburg and some other large cities in the global south as linking cities. They are satellites of sorts, but the control structure of the world economy has shifted from a comparatively simple, hierarchical structure of imperial domination to a more complex, multipolar network of world cities that are dispersed across all continents. This network spans the globe, operates 24 hours a day and extracts surplus from the territories around its nodes. The linking cities cannot be considered as being in sole control, but crucial in facilitating the extraction of surplus value. Parnreiter's (2002) study clearly demonstrates such a role for Mexico City in facilitating access to the regions of low cost production and articulating the interactions between a national economic space and global economic actors (see also Sassen, 2002). 131 Innes (1984, p. 236) argues that Anglo American in the 1980s had to be "understood in no other way than as a multinational -- and one of the world's more important multinationals at that". He also argues that it would be foolish to suggest that Anglo American's role as the largest foreign investor in the United States would suggest that the US are being colonised by South African capitalism. However, the economic might of South African business has certainly left its mark on the sub-continent with South African firms, first of all Anglo American, extracting profits from around southern Africa (Innes, 1984, p. 238). Similarly, Bond (2004) argues 20 years later that South Africa's relationship with the rest of southern Africa is a sub-imperialist one, facilitated by the ANC government and endorsed by major corporations (see also Gibb, 2006). My argument is that there are two specific features that suggest contemporary world cities have not become world cities out of their own strength. First, the wealth that world cities are built on was - at least partly - created in distant places. As Anthony King (1990, pp. 6-7) argues, contemporary world cities cannot be understood outside the context of the colonial periphery. Many large corporations active in the South African mining industry had their headquarters in London throughout their existence. London was the place where capital was raised and dividends returned. South Africa merely hosted the operations and provided the minerals. Anglo American and Gencor, however, were different in that their domicile was South Africa, headquartered in downtown Johannesburg. Their moves to London can only be understood as building the strength of Britain's foremost world city from afar. Second, national governments have actively enabled some cities to become world cities with all their control and command functions. Johannesburg remains in the lower echelons of the world city network and it can certainly be characterised as a linking city, facilitating the connections between capital and production. Johannesburg's role in the network was enabled by a socio-spatial transformation actively supported by a new constellation of political power. Capital globalised when Anglo American and Gencor moved to London. As discussed in chapter 1, this form of globalisation is not a self- propelled process (see Yeung, 2002). It is the result of deliberate steps taken by the companies and allowed by the South African Reserve Bank under an ANC government. As Swyngedouw (2004, p. 42) suggests, "the spaces of the circulation of capital have been upscaled" in what he characterises as 'glocalisation'. When Anglo American 132 chairman Gavin Relly and the small group of business men travelled to Zambia in September 1985 to meet ANC president Oliver Tambo, Thabo Mbeki and other ANC leaders, they likely knew little of what the exact consequences of this meeting would be. They sought assurances that Anglo American and South African capitalism would continue to exist beyond the transition to majority rule. They got what they wanted. The results, although by no means carved in stone, would have been unthinkable even ten years earlier, when the Soviet Union's influence in the region had been strengthened by Angola's and Mozambique's transition to independence. The relationship between the ANC and the business community certainly evolved, but it was no doubt close. In 2002 Nelson Mandela publicly said that in 1990 the ANC had received one million South African Rand from former Anglo chairman Harry Oppenheimer for the relocation of its 20,000 exiles (Padayachee, 2003, p. 192). He also said that "before the ANC came to power in 1994, the organization had realised that it would lose the support of the business community with its then policy of nationalization of the mines and monopolistic industries" (cited in Padayachee, 2003, p. 192). The state-economy nexus and the rise of the world city network Understanding the world city in John Friedmann's (1986, pp. 69, 71) words as the "basing points" for global capital, Johannesburg is part of the world city network, even if it is only a minor node. The city is the administrative centre that facilitates the efficient resource extraction for much of southern Africa. While considerable ownership over the resource industry is concentrated in Johannesburg, the city is also a node through which surplus is channelled to distant shareholders. Contrary to ANC policy of nationalising monopoly industries during the years of the struggle, and thus somewhat ironically, the liberalisation policy of the ANC government has facilitated Johannesburg's role as a linking city. It allowed large South African companies to leave the country and liberated its capital owning class from the confines and exchange controls that the apartheid government had maintained while in power. Political liberation quickly culminated in economic liberalisation and the network of South African elites was able to join the transnational capitalist class (for a 133 characterisation of this class see Sklair, 2001). Johannesburg has become just one basing point for the diverse activities of two of the world's largest mining companies, Anglo American and BHP Billiton. Gencor chairman Derek Keys, who temporarily served as Finance Minister in the last years of the National Party government, became chairman of Billiton in 1995 (Bond, 2000b, p. 25; Jones, 1995, p. 264) and former Gencor director of finance Mick Davis is now Chief Executive of BHP Billiton. In the early 20th century General Jan Smuts, Prime Minister of the South African Union from 1919 until 1924 and from 1939 until 1948, had "reservations about the Americans moving in for a fat profit" (quoted in Padayachee, 2003, p. 185). In the 1990s it was the Afrikaners (as well as English dominated capital) that moved out for fat profits. This move was facilitated under the majority rule of the ANC and the South African state. South African capitalists have gained flexibility, but their flexibility is characterised by the geographic spread of their assets rather than the diversity of their portfolio. In industries that provide basic inputs into the global economy, such as natural resources, spatial flexibility may be more important than sectoral flexibility as there will always be long term prospects for basic industries. The spatial diversification and increased sectoral focus of South African conglomerates is an indication of the power that these firms maintain. Spatial flexibility within a specific monopoly is perhaps more powerful than flexibility across sectors. There is therefore a strong argument to be made to expand world city research beyond the producer services. Moreover, this chapter has shown that world cities need states. Had the South African government not lifted exchange controls in the mid-1990s, the largest South African firms would not have been able to shift their headquarters and listings to London and utilise the world city network in the same way. Most importantly though for this thesis, the production of London is achieved to a large degree through the input from the colonies with ever mutating forms of capitalism. At least two of the world's largest mining firms today have grown up under the protection of South African isolation and with favourable conditions of cheap black labour, before breaking out of their national confines and taking their fortunes to London and elsewhere overseas. Once again the world city London has grown substantially, mainly through the input of black people's sweat and labour and the African environment, while later claiming the benefits and the power and the glory. 134 Chapter 6: Vertical Control in the World City Network We, the poorest of the poor are requesting that you make an improvement from the current offers of 6,5 %, 8 and 10 % that you left at the table earlier this week. Food prices are five times the prices they were when you last raised our wages and electricity prices have gone up 31,33 %. Our children live in dusty squatter camps  and  in  the  deep  rural  villages  where  they  share  water  with  animals. Together we can no longer afford food and the meagre wages you give us are making us sick. Today, they are making us to feel [sic] tired and next week we are going to be sick and tired! Today, we are asking you Mr or Madam Chamber to raise our wages but next week we are going to DEMAND. [...] As you drink a glass of  expensive whisky somewhere overseas,  our children have no drinking water because your other brothers have put water meters which force them to go to the streams where animals drink. Darkness is the order of the day as electricity prices has [sic] skyrocketed. Food is a luxury. House is shack [sic]. [...] (Seshoka, 2009) On July 16th, 2009, Lesiba Seshoka, spokesperson of the National Union of Mineworkers (NUM), wrote on behalf of 150,000 members of his union to the South African Chamber of Mines. The quote above represents excerpts from this letter addressing the owners of goldmines inside and outside of South Africa. Two aspects of the quote are of particular relevance to this chapter. First, it describes the conditions of the mine workers from the point of view of the labour organisation, which sees a world of difference22 between the workers and the owners. Second, it points to a specific spatiality in this world of difference, pointing to the consumption of luxury goods by the owners outside of South Africa in contrast to the absence of basic goods and services for the workers inside the country. The letter also addresses an entanglement between the state and the owners of the mines, when it refers to "your other brothers" in connection with the metering and commodification of drinking water, a form of class entanglement between the transnational ownership of mines and the local bureaucracy and law enforcement. When the ANC came to power in 1994, it had a delicate balancing act to fulfil. On the one hand it was under pressure to right past wrongs; on the other, it felt the need to preserve economic stability and cater to the business community inside the country and 22 The phrase 'world of difference' is borrowed from the book title A world of difference: encountering and contesting development (Sheppard, Porter, Faust and Nagar, 2009). 135 international investors. For many observers on the left South Africa is a neo-liberal state, but there is some evidence that it has maintained considerable elements of more social- democratic policies promoted by the South African Communist Party (SACP) and the Congress of South African Trade Unions (COSATU) (D. Pillay, 2008, p. 47; D. Pillay, 2004, pp. 167-168). The current South African Constitution allows for a wide range of development trajectories and initially the ANC largely adopted a state-oriented social- democratic approach with its Reconstruction and Development Programme (RDP) announced in 1994 (D. Pillay, 2004, p. 167). The Keynesian social-democratic leanings of the government were, however, quickly pushed aside for the neo-liberal oriented Growth, Employment and Redistribution (GEAR) programme, which "was arguably one of the most inflexible and intransigent interpretations of neo-liberal orthodoxy in the world, and produced little growth at first (it started to push into the 4–5 per cent per annum range only after 2001 and has fallen back since 2007)" (Padayachee, 2009, p. 15). Both currents - the Keynesian social-democratic one and the orthodox neo-liberal one - continue to push against each other within the South African government. Shortly after its ascent to political power the ANC also started to explicitly circulate the idea of a South African developmental state, but there is much doubt that this will materialise in any meaningful sense, given the continuing emphasis on neo-liberal policies (D. Pillay, 2008, p. 47). The competing currents are also articulated through GEAR and the reduction of tariffs by even greater amounts than required by the World Trade Organisation (WTO), on one hand, and tightening of labour relations through the Labour Relations Act on the other (Carmody, 2002, p. 258). Regardless of which interpretation of the South African political economy prevails, there is growing inequality and many social indicators show negative development. The Gini coefficient (by income), for example, rose from 0.6 in 1994 to 0.72 in 200623 and unemployment rose, while the housing, education and health sectors performed poorly (Bond, 2008, p. 11). There was a core group of people in the ANC that pushed for neo-liberal reforms and since 1997 Thabo Mbeki has acted as de facto president since an ageing Nelson Mandela slowly withdrew from active politics (Jordan, 2004, p. 209). Trevor Manuel, 23 The United Nations Development Programme states a Gini coefficient for South Africa of a considerably lower 0.58 in its 2009 Human Development Report. This is still amongst highest in the world, ahead only of Namibia, Comoros, Botswana, Haiti, Angola, Colombia and Bolivia (United Nations Development Programme, 2009, pp. 195-197). 136 who has been a government minister from 1994 on and headed the finance ministry from 1996 - 2008, has been at the forefront of reforms and ended the exchange controls to allow the free flow of capital in and out of South Africa. Even former unionists and SACP members started to preach the orthodoxy of fiscal discipline to COSATU members in 1994 and worked on consent for orthodox policies (Habib, Pillay and Desai, 1998, p. 106). In spite of these macro-economic measures the South African economy has been characterised by low growth and high unemployment (Seidman Makgetla, 2004). There has been little growth, negative employment trends and no redistribution under GEAR (Bond, 2008). In addition to these macro-economic policies, the government also actively supported the creation of a black capitalist class by legislating the sale of white owned assets to formerly disadvantaged South Africans. While this may seem like a heavy handed interference into economic liberties by the South African government contrary to neo-liberal theory, it can also be understood as a deliberate minor market intervention that merely corrects unfair practices of the past and signals a strong belief in the market as a corrective force to previous racial discrimination. The macro-economic policies since the advent of democracy in 1994 have considerably shaped South African society and had considerable influence on Johannesburg's integration into the world economy. This chapter will start with a review of Johannesburg's position in the world city network and conceptualisations of the world city paradigm that are specifically relevant to the city. It will then analyse the policy of Black Economic Empowerment (BEE) in South Africa and its role in ensuring the functioning of a capitalist economy, followed by a section on the spatial concentration of BEE firms in Johannesburg and the associated spatial concentration of control and command functions. The last two sections analyse the vertical chain of command and control functions from the cities of strategic control through the places of operational control and the implementation of decisions on the ground. These two sections are based on an analysis of labour relations in the mining industry as well as the relations between mining companies and local communities. I will continue to examine the topic through the world city lens and return to the particular relationship of cities and states, including how certain world cities provide important links between economic and political power. I will also pay particular attention to the role of Johannesburg as a regional business centre and how it facilitates the 137 efficient flow of control over investments from the investors to the place of production. In short, I will examine how the policies of the national South African government have contributed to the production of Johannesburg as a world city and how command and control functions are exercised through a network of cities reaching small communities and individuals in rural South Africa. The world city as a neo-liberal city Neo-liberal policies are by no means confined to the South African national state. Municipalities across the country are either forced to or actively embrace adapting their policies to the fiscal realities of balanced books and austerity. Neo-liberal urban policy in South Africa has been pushed since the end of apartheid by institutions such as the Urban Foundation (the private sector initiative founded by Harry Oppenheimer and Anton Rupert after the 1976 Soweto uprising, see Chapter 5) or the World Bank (Todes, 2000; see also Harrison, Huchzermeyer and Mayekiso, 2003; U. Pillay, 2008). There has been a large literature on neo-liberal urbanism (e.g. Peck and Theodore, 2010; Brenner and Theodore, 2002; Jessop, 2002), but there are few explicit links between ideas of neo- liberal urbanisation and the world city concept. Similarly, Murray (2004) contrasts the global city concept, which applies urban political economy as its starting point, with 'post modern' urbanism, which analyses the cityscape as contested terrain, where local struggles manifest. However, as mentioned in Chapter 2, there are a number of studies applying ideas from the world city concept to studies of cities as bounded entities that grapple with the influences and stresses imposed on them by the processes of globalisation. One of the most explicit works that link neo-liberal urbanisation to the world city concept is David McDonald's (2008) book World City Syndrome: neoliberalism and inequality in Cape Town. While he recognises that the world city hypothesis is problematic in certain ways, he also states that "it nevertheless offers enormously useful insights into contemporary urbanization" (McDonald, 2008, p. 5). He emphasises that the homogenizing and hegemonic influences of capitalist development are inescapable for cities around the world, even if some -- such as Pyongyang, Havana and Teheran -- are not as closely incorporated into the dominant urban flow. For 138 McDonald, Cape Town is a world city, because it is a capitalist and a neo-liberal city (McDonald, 2008, p. 6). McDonald's conceptualisation of the world city, applied to the case of Cape Town, is remarkably close to Friedmann and Wolff's work. They write (as quoted in Chapter 2): "Our basic argument is that the character of the urbanizing processes -- economic, social and spatial -- which define life in these 'cities' reflect, to a considerable extent, the mode of their integration into the world economy" (Friedmann and Wolff, 1982, p. 309). There is an inherent recognition that the level of integration into global processes shapes life in cities, or in more theoretical terms that the system-world has a profound effect on the life-world. This mode of integration will undoubtedly be different in Johannesburg as opposed to New York or London or even Cape Town. The specifics are unique, but the world city concept can help explain some processes of urbanisation in cities across the world. Rogerson (2005), for example, examines the changes of post-apartheid Johannesburg in the context of South Africa's re-integration into the global economy. He notes that Johannesburg's experience was different from most other major cities in southern Africa, because South Africa's liberalisation of its economy was largely a result of domestic policy, while that of other countries (affecting life in their cities) was usually the result of externally enforced structural adjustment policies. Regardless of the terms of global integration, such changes affect life in cities and the world city concept provides a useful analytical framework in this regard. What is important to note here, however, is that it is often national policy that has the strongest influence on a city's mode of integration into the global economy rather than local, municipal urban policies. This is the case regardless if this policy is superimposed by some supra-national organisation like the International Monetary Fund (IMF) and the World Bank or a domestically conceived and grown policy. The city of Johannesburg is both a global city and a globalising city. It is embedded in the network of world cities with hierarchical relationships and as a place of commerce it fulfils specific functions within the global economy. There is a very specific place in the structure of the world urban system for Johannesburg. At the same time there are myriad actors in the city that shape the processes of globalisation. These are actors who articulate economic flows, social agendas, political events and cultural perceptions. 139 Johannesburg is often considered a special case in world city research. It was referred to as “the most isolated world city” and having “a continent to itself and no clear similarities with other world cities” (Taylor and Walker, 2001, p. 42; see also Rogerson, 2005) or an “aspirant global city” (Mabin, 2007), considered “an abnormal base” for understanding globalization (Beavon, 1998) and it was the only world city on the African continent in John Friedmann's first classification of world cities in 1986 (see Friedmann, 1986). The city is also one of the most divided cities in the world with a highly uneven distribution of wealth. Parts of the city offer services and environments that are similar to wealthy neighbourhoods and suburbs in cities around the world, while unemployment is at very high levels and many neighbourhoods lack basic services (see e.g. Bond, 2000a; Beall, Crankshaw, & Parnell, 2000, 2002; Crankshaw and Parnell, 2004; Mabin, 2007; Parnell, 2007). The city of Johannesburg is the centre of a large metropolitan area and both the city proper and the metropolitan region are dominant within South Africa and the entire sub-region. The city itself (after the extension of its boundaries in 2000) has a population of about 3.2 million (Beavon, 2006, p. 50) while the extended metropolitan area counts about nine million people (Mabin, 2007, p. 39). More significantly in assessing the city's economic significance, however, is its substantial contribution to the national economy accounting for almost 16 per cent of South Africa's Gross Domestic Product (GDP), while the population share is only 7.1 per cent (Rogerson, 2005, p. 18). The city is also the major financial centre for the entire sub-region of southern Africa. The city's economy is dominated by the service sector, most notably financial and business services, which accounted for 22.6 per cent of total formal employment and 31.7 per cent of its gross geographic product (Rogerson, 2005, p. 20). The city is also home to 74 per cent of national corporate headquarters and 55 percent of the total office space in the country. Some specialised sectors cluster outside Johannesburg, as for example the oil industry in Cape Town, but there is little doubt on the overall leading position of Johannesburg in the country and the region (Mabin, 2007, p. 49-50). The city has a long history as a cosmopolitan centre albeit being a colonial outpost. Johannesburg already enjoyed “something of a world city status when it was only forty years old” (Beavon, 2006, p. 50). During the 1920s the majority of the population was white, i.e. not particularly cosmopolitan, but it was a major destination 140 for immigrants and comparable to populations in other white settler cities, like those in Canada or Australia (Mabin, 2007, p. 38) and Johannesburg became certainly the centre of the political struggle of the black majority prior to and during the apartheid years. Although the common story is that Johannesburg was built on the basis of its gold deposits, Mabin (2007, p. 38) suggests that it was never purely a mining economy, but was a diverse financial (and mining) centre, fulfilling the characteristics of a contemporary world city. After the end of the gold standard in South Africa in late 1932, investment poured into the country through the city, with a visible effect on urban form, as skyscrapers for offices and apartments were erected in the city centre (Robinson, 2003a, pp. 260-261) and the city hosted the 1936 Empire Exhibition (see Robinson, 2003b). There is relatively little doubt that Johannesburg is a globally connected city, even if a number of world city rankings exclude it from the status of being a world city. It is certainly “a prism through which to view the world” (Mabin, 2007, p. 59), in spite of being or especially because it is often considered a special case when included as a world city (see Taylor and Walker, 2001). But as Rogerson (2005) explains, Johannesburg's current status is a reflection of South Africa's re-integration into the global economy after the end of apartheid. The role Johannesburg plays in both the national as well as the global economy has thus been substantially shaped by national policies. Had the ANC pursued its longstanding policy of nationalisation of the mining industry once it was in government, the role of Johannesburg as a linking city to global capital as well as a command and control centre would have almost certainly been very different, although its role would not have been insignificant. The next two sections explore the policy of BEE and some of the changes in mining policy after the end of apartheid as well as the concentration of BEE ownership in Johannesburg. The subsequent two sections will explore how control and command functions are exercised from the offices in Johannesburg and how these decisions affect the communities in which the investments are taking place. 141 Black Economic Empowerment and the new mining charter When the ANC finally came to power in 1994, it was under pressure to improve living standards for the black majority as well as to avoid a panic amongst white capital owners and international investors. One of the policies that took shape in the subsequent decade or two was BEE. Establishing BEE legislation in South Africa was a long and drawn out process. While the overall macro-economic policies were taking shape through GEAR by 1996, it was not until almost ten years into majority rule that it was clear what BEE should look like. Although BEE has long been in the public sphere and assets have been transferred to BEE companies since 1994, the National Black Economic Empowerment Act No. 53 did not come into effect until 2003 (Hamann, 2009, p. 449). The empowerment act specified that each sector of the economy had to publish a transformation charter that would advance the objectives of the act and was developed by major stakeholders in that sector. The mining charter, which set out specific terms for the mining industry, was already published a few months before the empowerment act was signed into law in October 2002 (Hamann, Khagram and Rohan, 2008, p. 30). Radical approaches to redistribution and black empowerment were in contradiction to the ANC's emphasis on reconciliation and it meant that any BEE legislation had to be non- threatening to white business interests (Southall, 2004, p. 313), which explains the lengthy process due to competing political interests even within the ruling party. When drafts of the new mining charter -- which suggested that 51 percent of the industry should be controlled by blacks (and other previously disadvantaged South Africans) within ten years -- were leaked to the public in July 2002, the markets reacted with a massive sell- off of shares (Hamann, Khagram and Rohan, 2008, p. 29). The sell-off re-enforced the fear of economic and political instability within the ruling party, should the government take too drastic actions. Interpretations of BEE vary significantly. Some consider it a relatively balanced form of state involvement and corporate social responsibility (e.g.  Hamann, Khagram and Rohan, 2008), while others see it as a class project with the aim of creating a black bourgeoisie, accepting continued high inequality and ensuring the smooth functioning of capitalism (e.g. Southall, 2004; Bond, 2000b, pp. 39-46; Beall, Gelb & Hassim, 2005, 693-694). There has been a very tight entanglement between political activists and the 142 new black capitalist class, including several of those imprisoned on Robben Island during the apartheid years, most prominently perhaps former secretary-general of the NUM and the ANC, Cyril Ramaphosa, and former ANC Gauteng premier and current minister in Jacob Zuma's cabinet Tokyo Sexwale (Southall, 2004). One of the most successful black business people is South Africa's first black dollar billionaire and founder of ARM Gold (African Rainbow Minerals) Patrice Motsepe, who is regularly involved in big BEE deals (Newmarch, 2007). The main focus of the BEE strategy in the public media is the transfer of ownership to historically disadvantaged South Africans (HDSA) (Hamann, 2009, p. 447). The mining charter defines HDSA as follows: "The term Historically Disadvantaged South Africans (HDSA) refers to any person, category of persons or community, disadvantaged by unfair discrimination before the Constitution of the Republic of South Africa, 1993 (Act No. 200 of 1993) came into operation" (Government of South Africa, 2002). Although the term HDSA clearly includes practically everyone, who was not classified as white by the apartheid government, the policy has been widely criticised for its narrow reach. In practice, it reaches a very small number of the previously disadvantaged South Africans, although the 2003 act was supposed to widen the scope of BEE: "The transfer of shares by large corporations has clearly accelerated since early 2003, but the involvement of the same handful of black business leaders in deal after deal across multiple economic sectors has simply emphasised the narrowness of the process (in both racial and gender terms) and limited its stabilising impact." (Beall, Gelb and Hassim, 2005, p. 694). The South African government has gone further than most governments in legislating corporate social responsibility including BEE, which goes far beyond voluntary standards that are often characteristic of corporate social responsibility initiatives. The government of South Africa emphasises that BEE is indeed broad based. In order to enforce compliance by companies with the legislation, the government introduced a broad-based black economic empowerment score card, covering the following categories in which companies have to take initiatives (percentage in parentheses indicate the weight of each category) (Hamann, 2009, pp. 447-450). - Ownership (20 %) - Management Control (10 %) 143 - Employment equity (10 %) - Skills development (20 %) - Preferential procurement (20 %) - Enterprise development (10 %) - Socio-economic development initiatives (10 %)  The above list shows that ownership transfer accounts for only 20 percent of the Broad-based Black Economic Empowerment strategy under which companies get evaluated on a score card according to a number of categories. The other categories are laudable initiatives, but aim more at long term social transformations, while ownership transfer has a more immediate effect in transferring wealth. In contrast, employment equity and skills development, for example, are initiatives that will take a generation or two to show any substantial social transformation. Hence, although ownership transfer accounts for only 20 percent on a company's scorecard on which it is evaluated regarding BEE compliance, the focus on this particular aspect of BEE in the public media is not surprising, because ownership is the primary mechanism of control over a company, or in explicitly Marxist terms it is the control over the means of production. Generally, companies have chosen two mechanisms to comply with the ownership transfer of assets requirement. They either sell assets to BEE companies and create completely separate entities or they create joint ventures with BEE companies. The scorecard is also taken into consideration when companies apply for an extension of mining rights under new legislation (Hamann, 2009, p. 450). While BEE legislation applies to all sectors, my focus here will be on the mining industry with particular emphasis on platinum mining and the specifics for this sector set out in the 2002 Broad-based Socio-Economic Empowerment Charter for the South African Mining Industry, commonly referred to as the Mining Charter, and its 2010 amendment. Under the mining charter companies in the industry are required to achieve 26 percent ownership by HDSA with a practical date for completion set for 2014 (Government of South Africa, 2002). This target was confirmed in the 2010 amendment to the mining charter (Department of Mineral Resources, 2010), although the Mining Charter Impact Assessment Report of 2009 (Department of Mineral Resources, 2009) stated that the industry was at that point far from achieving the goal set in the mining 144 charter. One of the major challenges in arranging the transfer of assets for market prices is the availability of credit. As one executive in the mining industry stated: "The single biggest problem is trying to sell to someone who does not have money." There are some guaranteed loans from banks or the state to historically disadvantaged South Africans, but that included a fairly small number of beneficiaries. The scheme depends on the continuous growth of the market, explained the executive, stating that Tokyo Sexwale, for example, had good timing, as he bought low, sold high and "the wealth was created for him" [Interview 5]. One strategy to finance BEE deals were the creation of Special Purpose Vehicles (SPVs), which are companies formed by a BEE company and a financier solely for the purpose of investing into another company, usually one that is actively looking for a BEE partner. Under such an arrangement, the BEE company would hold 100 percent of ordinary shares in the SPV and thus have full voting rights. The BEE partner would raise funds partly through a loan from the financier and partly by issuing preference shares in the SPV to them. After about three to five years the financier would redeem the preference shares. The details of SPV arrangements vary, but they generally rely on rising share prices for the BEE partner to repay its loans (Engdahl and Hauki, 2001). This is clearly a strategy that has often not worked out, especially during times of stock market volatility. When the mining charter came into effect, the industry agreed to support BEE firms in securing finance up to R 100 billion for the first five years. In order to achieve the goal of 26 percent HDSA ownership the transfer of more than these R 100 billion were required and subsequent transfers would be on a willing seller willing buyer basis, at fair market value. In 2010 the South African mining industry was estimated to be worth R 2 trillion, which means that R 100 billion industry fund accounts for only about five percent. Hence, future transfers would rely on the continuous growth of the industry to finance the transfer of the additional 21 percent. Although the interim target for HDSA ownership in 2009 was 15 percent, the report states the "aggregated BEE ownership of the mining industry has, at best, reached 9 percent" (Department of Mineral Resources, 2009, p. 17). BEE ownership of the top 100 firms listed at the JSE is even slightly lower at around 8 percent. Moreover, the Black Management Forum pointed out that the actual black ownership will be considerably lower than even 8 percent, because BEE firms can 145 be partnerships with significant non-HDSA ownership, and that there is no progress. The president of the Black Management Forum was quoted by the Mail and Guardian: "The study has unfortunately confirmed that transformation is in reverse gear" (Mail and Guardian, 2010). After the financial crisis of 2008, BEE deals were in decline and roughly half the value in 2009 compared to 2008 (Rumney, 2009). And, although BEE was intended to be broad-based, it clearly has not achieved this goal in terms of ownership. The Mining Charter Impact Assessment Report (Department of Mineral Resources, 2009, p. 17) states: "Regrettably, the reported level of BEE ownership is concentrated in the hands of anchor partners and SPV’s [Special Purpose Vehicles], representing a handful of black beneficiaries, contrary to the spirit and aspiration of both the Freedom Charter and the Mining Charter." In an interesting twist the report still describes the mining charter as "a deliberate intervention" that is supposed to be in line with the 1955 Freedom Charter, which stated that the "mineral wealth beneath the soil… shall be transferred to the ownership of the people as a whole” (Department of Mineral Resources, 2009, p. 16). Political commentators on the left in South Africa have been highly critical of BEE, amongst them the brother of former president Thabo Mbeki and economic analyst, Moeletsi Mbeki. When addressing the current president Jacob Zuma, he said that BEE "contributes to accelerate economic inequalities within the black population without contributing to economic growth and job creation" (Ferreira, 2011). BEE in its current form has also been criticised by labour unions as "elitist". COSATU complained that millions of shares were transferred to a few individuals, while the overwhelming black majority was left disempowered. In a statement regarding a BEE deal involving ArcelorMittal, COSATU said: “Instead of making a rich elite minority even richer, BEE should benefit the workers, including the unemployed and poor communities” (Sowetan, 2010). Regardless of the overall success or failure of BEE in South Africa, perhaps the most important element of this policy is that it has not affected the functioning of capitalism in any substantial way. To the contrary, capitalism has been successfully protected by the South African state, now dominated by an alliance of leaders in the former liberation movement including those imprisoned on Robben Island and the liberal business community. By the late 1980s inequalities within South Africa could no longer 146 be sustained through racial laws and apartheid was replaced with a liberal democracy. But two decades later stark inequalities remain, while the country overall is politically stable, with the overwhelming majority of people still voting for a government that maintains this inequality in a different guise. Black but not broad-based Following McDonald's (2008) analysis that Cape Town is a world city, because it is a capitalist and neo-liberal city, I argue that the same remains true for Johannesburg, perhaps even more explicitly. BEE is a neo-liberal policy: The political goal of redistribution is assumed to be best achieved through market mechanisms with minor regulatory interventions by the state. This type of policy symbolises a remarkable turnaround of political values and a shift to realpolitik amongst the ANC. Patrick Bond (2000b, p. 15) illustrates this with Nelson Mandela's words, when he was still serving his last month in prison in early 1990. At the time Mandela insisted that the demand from the Freedom Charter for "the nationalisation of mines, banks and monopoly industries is the policy of the ANC and a change or modification of our views in this regard is inconceivable". Although the statement initially caused some concern in the (white) business community in South Africa, any such notion was completely off the table just a few years later. Instead of nationalising mines, banks and other monopoly industries, redistribution was intended to be undertaken within the capitalist system, partly through the mining charter, which was in an almost cynical twist described as "in line with clause (3) of the freedom charter" (Department of Mineral Resources, 2009, p. 16). BEE is hardly broad-based in terms of redistribution of wealth amongst the people. Moreover, it is not geographically broad-based and has contributed to Johannesburg's leading role as a business centre in South Africa and a world city. Table 6.1 shows all BEE companies in the Platinum industry for 2007. The data are based on Table 4.3. BEE companies accounted for approximately 7.5 percent of global platinum production in 2007. Only about ten percent of this production can be attributed to owners with a registered head office outside Johannesburg, close to the 147 respective site of production and mineral deposits. The remaining 90 percent of economic entities have their offices in Johannesburg and about two thirds of the production can be attributed to the two largest BEE producers in the sector, Northam and African Rainbow Minerals. Both of these companies have BEE status, because their respective owners are HDSA. However, the ownership of neither company is broad-based, but in the hands of a single owner or family. On June 30th, 2008 Northam was owned to 21.7 percent by Mvelaphanda Resources, while another 22.2 percent were owned by Anglo Platinum. Mvelaphanda Resources became the controlling shareholder the following year and by June 30th, 2009 it owned 62.8 percent of the company, while Anglo Platinum sold all its shares. Mvelaphanda Resources was founded by Tokyo Sexwale and maintained its BEE status. African Rainbow Minerals was founded in 1997 by Patrice Motsepe and became the BEE partner for the gold miner Harmony Gold. A private company (African Rainbow Minerals & Exploration Investments (Pty) Limited), which is controlled by Motsepe's family trust, continued to hold more than 40 percent of African Rainbow Mierals. Patrice Motsepe's personal wealth is estimated at 2.15 billion US Dollars (Newmarch, 2007). Table 6.1: Platinum production attributed to BEE companies, 2007 Owner Head office location Total Production control in kg As % of BEE production As % of global production Northam Platinum Johannesburg 6610 41.18 3.12 African Rainbow Minerals Ltd. Johannesburg 3949 24.60 1.86 Royal Bafokeng Resources (Pty.) Ltd. Johannesburg 3000 18.69 1.42 Bakgatla-Ba-Kgafela Traditional Community Saulspoort (North West Province) 1440 8.97 0.68 Kagiso Platinum Venture Johannesburg 385 2.40 0.18 Modikwa Communities Johannesburg 311 1.94 0.15 Mvelaphanda Resources Johannesburg 179 1.12 0.08 Bapo Ba Mogale Tribe North West Province 179 1.12 0.08 Total 16053 100.00 7.57 Source: Author's analysis of United States Geological Survey (2008) data and company reports. Commentators in the popular press have noted the parallels between current developments in the South African mining industry and previous ones in the apartheid era. The main difference is that the capital owning class in South Africa is not longer exclusively white, but the number of black people included is very small. Not unlike 148 Ernest Oppenheimer in the 1920s, the new faces in the industry are all very well connected, including political ties to the governing party. The South African Mail and Guardian, for example, titled a story about Patrice Motsepe "Billion-dollar brother" and noted that he is besides Nicky Oppenheimer and Anton Rupert the only South African on the Forbes Magazine list of the world's billionaires (Newmarch, 2007). Tokyo Sexwale was Premier of Gauteng Province before focusing on his business Mvelaphanda, and was appointed as a cabinet minister, when Jacob Zuma became President of South Africa. BEE has created a black capitalist class. The high level of ownership in few hands, even amongst the BEE companies, means of course that little control over the industry is exercised by people away from the business centre, Johannesburg. Some of the decisions by smaller owners with offices in Johannesburg may be made in locations closer to the mining communities and the Johannesburg office functions mainly as a location for interacting with other players in the industry. Nevertheless, these are fairly small producers and in some cases, as for example Royal Bafokeng Resources (Pty.) Ltd., there are explicit charges of corruption and undemocratic behaviour, where individuals act as representative of a community without an actual mandate [Interview 10]. Most importantly, however, there is no broad- based redistribution of ownership in the South African platinum mining industry and the industry continues to function within a capitalist framework, where there is a need for centralised business services and thus control functions in a major business centre. Moreover, BEE firms are for profit organisations and the further transfer of assets to HDSA relies on further growth and higher profit margins, of which one component is access to cheap labour, accelerating accumulation in the hands of a few. For example, expectations between employers and the union were vastly different during wage negotiations at the Two Rivers Platinum Mine, jointly owned by Impala and African Rainbow Minerals. NUM spokesperson Lesiba Seshoka commented with respect to the wage negotiations: “We thought the ARM [African Rainbow Minerals] leadership would understand the dynamics of where the workers are coming from, but they are acting as if they don’t know about poverty” (Modimoeng, 2009). Even if the mines would have been nationalised, it is conceivable that Johannesburg would have continued to play a major role, just as Santiago does in the administration of Codelco (see Chapter 3). However, the locus of decision making would 149 have shifted to some extent. There is a sense that BEE "created a drive for more South African" control [Interview 5], but this was largely concentrated in one city. The role of Johannesburg as a capitalist city and as such a world city was to a considerable degree shaped by national level neo-liberal policies and a considerably weakened approach to the redistribution of wealth from whites to blacks. Labour relations and Johannesburg's "long arm" In most world city rankings Johannesburg appears as a third or fourth level world city, with a relatively small presence of producer services firms and weak connections compared to cities like New York, London and Hong Kong (e.g. Taylor, Catalano & Walker, 2002). Often Johannesburg is identified as a special case due to its geographic distance to the world's other major cities (e.g. Taylor and Walker, 2001, p. 42), but the city's role and specific functions in the vertical chain of command and control has rarely been analysed. In Chapter 4 I identified Johannesburg as a major player in the control over the platinum industry. While the strategic control over the world's biggest and third biggest platinum mining firms is exercised to a large degree in London, there is strong evidence that Johannesburg exercises most of the operational control over the platinum industry as well as considerable strategic control, especially through the smaller Platinum mining firms. If strategic and operational control are separated in different cities, there is a two way dependent relationship between these cities. This section will investigate in more detail, how this operational control is exercised and how it links the sites of the actual capital investment, in this case the platinum mines, to the sites of the investors, i.e. those holding shares in platinum mining firms, through a network of cities. Typically this level of control would be attributed to cities identified as gateway cities, but perhaps a more appropriate term may be linking cities, because such cities link places across the globe, even if they are functioning as gateways into national economies. For the two biggest platinum miners in South Africa, the spatial division of control can roughly be divided into strategic control in London and operational control in Johannesburg. In the next two sections I will further investigate the relationship between 150 control functions located in Johannesburg and their relationships to mining towns and communities across South Africa. The focus will be on labour relations in the industry and the relationships between mining firms and local communities more broadly. These sections will rely on qualitative information derived from interviews, news coverage as well as reports from civil society organisations and mining firms. While the examples used here are exclusively from South Africa, because of the nature of the Platinum industry and locational constraints in conducting interviews, the relationships described will be similar across southern Africa, regardless of national boundaries, due to Johannesburg's functioning as the major business centre in southern Africa. There is generally a strong sense that the day-to-day operational decisions regarding mines in South Africa are made in South Africa. As one analyst (quoted above) stated: "London will tell Johannesburg that we need this kind of production and productivity level and then Johannesburg will decide how to go about it." [Interview 23]. There exists, however, an additional spatial division of control and command functions between the mine level and a firm's headquarters, usually located in Johannesburg. There is thus a three layered hierarchy of control and command functions in the mining industry. First, there is the strategic control level at a firm's headquarters that is usually close to the location of investors or in a highly connected city. Second, there is the operational control layer, which will be a regional or national office located in the same country or region as the operations. Third, the level of implementation is located at the site of production, in this study a mine. This three layered relationship becomes particularly apparent with respect to labour relations. A COSATU analyst expresses the distant strategic control level as quite removed from the operations at the mine: "Firms like AngloAmerican consider South Africa just as one of those investment opportunities" [Interview 6]. He further expressed concern that the mining industry in South Africa is in serious decline and that through stock listings oversees and headquarters outside South Africa, firms have become very footloose and can easily withdraw from their South African investments, aided by very loose government regulation [Interview 6]. With reference to the structure of the extractive industry in South Africa, he further expressed concern that South Africa primarily exports raw materials rather than processed products with additional value 151 added: "Unfortunately our economy is still structured along those colonial lines" [Interview 6]. While the structural constraints of the industry are recognised by union representatives, labour relations are by and large negotiated within South Africa. There is, however, a shift toward negotiations within individual firms rather than between all mining firms represented by the Chamber of Mines in downtown Johannesburg and the unions, most dominantly the NUM. As one mining union representative explained: The role of the Chamber of Mines has changed, it used to be one voice, but some Platinum, coal and other businesses withdrew from the chamber and now the chamber is mainly for gold mining. This is a big problem in labour relations, he continued, because of disparities between chamber affiliated mines and unaffiliated mines. "The only reason why the Chamber [of Mines] still exists is because of the NUM [National Union of Mine Workers]. It keeps parity between bigger mining houses" [Interview 24]. While there was a shift toward negotiating labour contracts with individual firms rather than the Chamber of Mines, these negotiations take place within a national South African context. According to the same labour representative there is not much direction from headquarters and not much has changed from a South African perspective on the ground. The emphasis, in his words, has moved a little bit towards the investor side as there is some concern by management about "what will the investor say?", e.g. when the union publishes a press release that is not favourable to the company. This he said, would give a bit more leverage to the union, but overall change is small. Direct contact with overseas offices does not happen in labour relations: "We never negotiated with a CEO, we always negotiated with HR [human resources] managers and they are always here [in Johannesburg/South Africa]" [Interview 24]. Mining communities and world cities Another aspect of the mining industry in which the three layered hierarchy becomes particularly evident is the relationship between mining companies and local communities living on potential mining land or near existing and potential mines. In March 2008 the international non-governmental organisation Action Aid released a report 152 titled "Precious Metal: The impact of Anglo Platinum on poor communities in Limpopo, South Africa". The report documented and analysed the impact that the mining activities of Anglo Platinum has on local communities in the northern province of Limpopo in South Africa. Pressing issues identified by Action Aid were the loss of agricultural land, water contamination, questionable legal practices for relocation agreements, the impact of intrusive mining activities on communities such as blasting, police brutality and continuation of mining activities in densely populated rural areas leading to further dislocation (Curtis, 2008). Anglo Platinum issued a report in response to the Action Aid report, in which it accused the non-governmental organisation of one-sided reporting, representing only a small minority within the affected communities and rebutted some of the allegations (Anglo Platinum, 2008). The London Mining Network (LMN), an alliance of human rights, development and environmental groups, extensively quoted the Action Aid report and documented similar human rights and environmental allegations against Anglo Platinum's parent company Anglo American in other countries (Mohideen, 2010). Action Aid recommended in its report that the South African Human Rights Commission (SAHRC) should open a full inquiry into the matter. The SAHRC accepted this recommendation and reported on its findings in the same year with a report titled: "Mining-related observations and recommendations: Anglo Platinum, affected communities and other stakeholders, in and around the PPL Mine, Limpopo" (South African Human Rights Commission, 2008). Even if some of the allegations against Anglo Platinum remain disputed, the company did not explicitly rebut all of the allegations, but pointed mainly to methodological issues and "sampling". The process of relocation of communities from future mining land documented by Action Aid provides considerable insights into the chain of command and control from those providing strategic control, operational control and on-the-ground implementation, i.e. a three layered hierarchy or relationship within corporate control. It also provides insights into the role of the state in enabling corporations in their endeavours through specific legislation and enforcing the law through police action. Interview data support the conclusion that there is a layered chain of command and control within corporate decision making. Moreover, the case of Anglo Platinum in Limpopo is not the only one where communities are in conflict with mining companies over land right issues and other environmental impacts. Recent documented 153 cases include a conflict between the Mudimeli community in Limpopo and the mining company Coal of Africa (Makana, 2009) as well as the Richtersveld community in the Northern Cape and the mining firm Alexkor (Merten, 2005). In one instance of community relocation for the expansion of the Anglo Platinum owned Potgietersrus Platinums Limited mine in 2003, villagers of Ga-Pila received R 5000  (approximately US$ 650 at the time) compensation plus a replacement home in the new village built by Anglo Platinum. Most importantly, however, the community as a whole lost 1800 hectares of agricultural land that it had occupied for generations. In the new village people only have small garden plots surrounding their houses (Curtis, 2008, pp. 14-15). Perhaps the most important aspect of this process is the implementation of the resettlement, which Action Aid described as "in practice a forced removal" based on interviews with villagers. A core challenge in the process of relocating communities is the legitimate representation of villagers. Under South African company law, so called section 21 companies were set up by Anglo Platinum in communities where it works. These are not-for-profit companies that Anglo Platinum considers legitimate representatives of the community. Action Aid claims that there are other community groups with much broader support and that the section 21 companies are financially closely entangled with Anglo Platinum and that some of these companies are undemocratic and lack accountability (Curtis, 2008, , pp. 15-16). Anglo Platinum by contrast describes its involvement in setting up and financing section 21 companies as "a principled position" that covered the costs of the relocation process (Anglo Platinum, 2008, p. 22). Perhaps the most important aspect in the process of establishing new mines and possible negative impacts on communities is the framework provided by national law. In South Africa consent by communities for mining to take place on their land is not required, however, they have to be consulted if resettlement is required. The establishment of mines and functioning of the mining industry is clearly a priority for the South African state. Action Aid quotes a senior official in the Department of Minerals and Energy as saying: "The problem is that one village might say yes and the three other villages might say no. Then there’d never be a mine there. We’re waiting for an agreement but we wouldn’t get it. But they do have to consult. People should be informed about the activity that will take place in their areas" (Curtis, 2008, p. 20). The London 154 (UK) based advocacy non-governmental organisation War on Want reported that when villagers protested the presence of a drilling team on the community's land near Anglo Platinum's Modikwa mine, police arrested community leaders and ordered the crowd to disperse. Subsequently the "crowd became restive and the police opened fire. Some 26 people were reportedly taken to local hospitals, eight of whom had rubber bullet wounds and one who had been hit in the arm by live ammunition" (Curtis, 2007, p. 8). The reports by Action Aid and War on Want suggest a strong entanglement between the state and companies, especially through the lack of land rights by communities and the powerful enforcement of the law by the South African Police. An activist in the mining industry with connections to Action Aid confirmed this view during an interview: The DME colludes with the big mining companies, the police and the magistrate. This constitutes an "incredible power", there is a "complete nexus between the state and companies" [Interview 10]. He considered it very difficult to resist mining on community land. With respect to South Africa's new mineral rights policy and the possibilities to prevent mining on community land he said: "As soon as the drilling machines are in the area, one has to react" [Interview 10]. In terms of where decisions are being made, a consultant in the industry assessed the locational level of control as follows: The "fight", e.g. regarding water rights, is never done at the local level. Communities knew they had to take the issue to the head office. That's where the engagement has to happen and it often includes going to the press. Involving the press is more effective for communities when dealing with larger firms, because they care more about bad publicity. A medium sized company would not care [Interview 18]. What is perhaps most revealing, however, is the role of third party actors in the downwards chain of command and control. Often firms like Anglo Platinum hire specialised consultants for the community engagement process. It is not the company itself talking to the community, for example when it comes to access to water: "The fuck- up is done by the on-the-ground contractor". [Interview 18]. 155 The state, cities and vertical control in a network of cities and communities In this chapter I examined two central aspects of a network of cities that efficiently facilitates the flow of capital from investors to the sites of bricks and mortar investments, where surplus value is produced. First, the role of the South African state is essential in the production of Johannesburg as a world city. Second, there is a layered hierarchy of control that is facilitated by a network of cities and stretches vertically from the location of the investors to the sites of investment. Following McDonald (2008) I characterise Johannesburg as a world city, because it is a capitalist and a neo-liberal city. The city has perhaps been shaped as a neo-liberal city more by national policy than by its own endeavours. South Africa's economic policies required the existence of a major business centre that links investors to investment opportunities. These investment opportunities are both within the South African state territorial boundaries and across the larger region southern Africa. As Brenner (2004, p. 3) argues, cities that try to position themselves optimally in global circuits of capital are the result of national, regional and local state strategies. Johannesburg is a node for overseas and domestic investment in the country and presumably the region. Most importantly, the South African government has not challenged capitalism as the principal mechanism to organise its economy and thus created or maintained the need for a central place of control to organise this capitalist economy. Johannesburg has taken on this role. Building on findings from Chapter 4 this case study of Johannesburg also suggests that the reach of the world city network can be characterised as a three layered hierarchy of control. The three levels can be identified as, first the level of strategic control, second the level of operational control and third the level of implementation of commands from higher levels. Previous studies of world cities have focused primarily on the interconnectedness of world cities through an interlocking network articulated by producer services firms (e.g. Taylor, Catalano & Walker, 2002; Taylor, 2004a). While these studies provide considerable insight into the position of a fairly small number of world cities relative to each other, there is practically no analysis of how these cities interact with the periphery of the world economy, which to a large degree is essential for generating the profits that sustain the world city network. Moreover, networks of world cities have been constructed primarily around producer services firms, because these 156 economic entities are considered essential in organising highly decentralised global production processes of goods and services (Sassen, 2001, p. 11). Sassen identifies insurance, banking, financial services, real estate, legal services, accounting and professional associations as producer services (Sassen, 2001, p.90). The producer services will likely locate in the same places as the corporate headquarters of transnational firms in the primary and secondary sectors, because these firms are the major clients for producer services firms. However, there is little evidence that producer services exercise substantial strategic control. Considering the types of ownership of the world's largest mining firms (see Table 3.4), only banking or financial services exercise direct control and command functions at the strategic level of control through fund managers and nominee accounts. The other producer services provide services to other firms, but do not exercise considerable strategic control unless they are shareholders of another corporation. I therefore argue that command and control functions assumed to be exercised by producer services firms are actually exercised through a layered multi-tiered chain of specialised corporate functions in various locations within those corporations that engage in the production of material goods. The primary concern of these corporations is to produce a profit for their shareholders that can be paid out as dividends and is shown through share price increases. Strategic decisions as to how to achieve this goal are generally made at the corporate headquarters. The decisions are communicated to and usually approved by shareholders at an annual general meeting and subsequently negotiated downwards to other levels of the organisation. While producer services are likely a suitable indicator for a specific type of network that connects cities from where producer services firms can easily reach their corporate clients, i.e. cities with a high level of centrality and therefore world cities needed for organising "the world for the efficient extraction of surplus" (Friedmann and Wolff, 1982, p. 309), I argue that command and control functions are somewhat more dispersed than indicated through these networks of producer services firms. First, decisions regarding the extraction of surplus, i.e. generating profits for shareholders, have to be negotiated across several levels of decision making, which I have identified as a three level hierarchy, of which the strategic level at the top is clearly the most powerful one. Second, this "efficient extraction of surplus" can only be achieved within an 157 appropriate political framework, which in turn has to be negotiated in the centres of political control, not usually captured in networks of producer services. This chapter has applied the example of the platinum industry to analyse the vertical chain of command and control and thus the connections from the peripheral locations of production to the world cities. Within the platinum industry London and Johannesburg are central nodes of strategic command functions, with Johannesburg exercising most of the operational control. While the findings in this chapter are built on the example of this specific industry, I want to cautiously suggest that some generalisations are appropriate. For example, much of the manufacturing industry could be organised along similar lines of control and command. Thus, the world city network is likely a lot more complex than identified through the producer services. 158 Chapter 7: Conclusion I started this dissertation by outlining two basic premises that underlie my main argument. First, the global economic system is based on a system of unequal exchange between the core economies and peripheral economies, as theorised by world-systems theorists. Second, capitalism is inherently anti-market and should be considered an additional layer above the market. Based on these two basic premises my aim is to re- ground world city research in a world-systems perspective. When considering world- systems theory, the emphasis of my analysis is on the characterisation of core and periphery as modes of economic production rather than geographic territories. This is important, because the rough categorisation of the world into core and periphery (with the addition of the semi-periphery at a later stage) implied a binary view of the world that was divided in First World and Third World, Global North and South, developed and developing world. Other concepts like polarisation at a global scale made similar contributions to a black-and-white-perception of the world, although authors from the world-systems school have always stressed the interconnectedness within the system. Yet, I wish to maintain a world-systems perspective in world city research. The binary view of world geography derived from world-systems theory is well outdated (if it was ever correct). However, the conceptualisation of core economic production processes as those yielding high profits through monopolistic control and peripheral production processes as taking place in highly competitive environments with low profit margins, remains a powerful explanatory concept for globally uneven development. The origins of world city research can be traced to a theoretical grounding in world-systems theory. Friedmann and Wolff (1982, p. 309) have characterised world cities in their seminal paper "World city formation: an agenda for research and action" as "vast, highly urbanized - and urbanizing - regions [that] play a vital part in the great capitalist undertaking to organize the world for the efficient extraction of surplus". While this perspective has not completely disappeared from later world city research, one of my main critiques is that it has paid far too little attention to actual production processes and investigated world cities and the network that these world cities form almost exclusively through the lens of the APS, which were taken as a proxy for command and control 159 functions. But in order to create value that can be extracted and to create sustained economic growth, people need to produce real products, like food, raw materials, manufactured goods or indeed services that make the production of such goods more efficient. The connection to these processes of production has almost disappeared from contemporary world city research with its focus on a narrow band of economic activity. I am therefore arguing for a world city research that bases its analysis on a considerably wider range of economic activities than the APS and investigates the connections between the places of control over economic activities and the locations of surplus production. Previous world city research's reliance on the APS did not comprehensively capture the command and control structures of the global economy and the role of these sectors in steering and organising a post-Fordist highly decentralised global economy. Sassen argued that there is a need for centralised control as production of goods and services became more scattered around the globe to seek the cheapest and thus most profitable sites for this production. APS would provide the necessary input for firms to "manage and control global networks of factories, service outlets, and branch offices" (Sassen, 2001, p. 11). Hence, those cities that hosted these APS, including insurance, banking, financial services, real estate, legal services, accounting and professional associations (Sassen, 2001, p.90), were considered the global cities. Although Sassen clearly points to the connection of the APS to the places of production, i.e. the "global networks of factories, service outlets, and branch offices", there has been no systematic investigation into this matter. The degree of reliance of firms on APS for the organisation of their production remains largely unknown. Later in the evolution of world city research the emphasis in constructing world city networks and hierarchies shifted from place based attribute data to indicators of connectivity, but the link between APS and the global network of productive activity remained largely unquestioned. The emphasis has been on 'advanced' and 'services', almost completely obscuring 'producer'. Following Giovanni Arrighi (1994, pp. 4-5), I have revisited Marx's formula of capital MCM', where M = money capital, i.e. liquidity, flexibility and freedom of choice; C = commodity capital, i.e. capital invested in a particular venture with a specific input- output combination and a view to make a profit; M' = expanded money capital, i.e. expanded liquidity, flexibility and freedom of choice. The APS are represented by M and 160 M' in this formula, but there is no sustained analysis of C in world city research. The argument here is that the role of the APS as a proxy for capital commanding functions is overstated. One of the main questions that emerges, then, is what constitutes a command and control function? More broadly speaking, what is power in a world urban system? Following the tradition of world-systems theory, I argue that the most powerful cities are those that are home to the people and institutions controlling global production monopolies. Amin (1997, pp. 3- 5) identifies five monopolies that the core countries utilise to extract surplus value from the periphery. These are the technological monopoly, the monopoly over financial control of worldwide financial markets, monopolistic access to the planet's natural resources, media and communication monopolies, and monopolies over weapons of mass destruction. Taylor (2000, pp. 7-10) considers APS a new production monopoly and contemporary world city research may well capture the monopoly over financial control of worldwide financial markets with its focus on APS, but the connections to other monopolies are rather weak, especially since all of these monopolies depend strongly on the political support of a variety of nation states and national elites. Power in world city research has been largely conceptualised as 'power to' rather than 'power over'. 'Power to' is characterised as a medium, such as the APS, to service global capital and thus direct this capital's specific input-output configuration. This 'power to' exhibits a considerable flexibility, depends largely on portfolio investment and follows Braudel's concept of power, according to which the most successful capitalists are the ones maintaining the highest degree of flexibility in order to switch their asset base as soon as more profitable input-output configurations arise (Taylor, Walker, Catalano & Hoyler, 2002). On the other hand 'power over' is characterised as the capacity exercised by the direct ownership power of specific assets, as for example a mine. While this type of power has more direct decision making power over very specific assets, it has significantly less flexibility and fewer options to switch its capital investment to more profitable options, should conditions change. In Friedmann's (1986; Friedmann and Wolff, 1982) characterisation of the world city, power as capacity (or 'power over') is the dominant conceptualisation of power, while Taylor's (e.g. 2004a) analysis of the world city relies almost exclusively on 'power to' as a concept for determining a city's position 161 in the world. This conceptualisation is less hierarchical and more networked, as power is derived from the links to multiple nodes in the network, each fulfilling a specific function in the organisation of capital (Taylor, Walker, Catalano & Hoyler, 2002, p. 232). . In this dissertation, I have returned to Friedmann's concept of power and investigated the direct ownership power over one of the world's most important monopolies, the monopoly over the extraction of natural resources. I have thus considerably diverted from the currently dominant direction in world city research. I am taking one of Friedmann's main points in his conceptualisation of world cities, the idea that world cities are basing points for capital, and thus trace the connections from the world city to the periphery and back. This is by no means to say that the network is obsolete. In fact, this dissertation has identified intersections between the two conceptualisations of power and New York and London have emerged as some of the best connected cities in the control network of natural resources. These two cities, so often identified as the leading duo in the world city network, represent ownership of the largest number of the ten largest non-fuel mineral mining companies, but their strength is their diversity in ownership. In terms of absolute direct control the cities of Santiago and Johannesburg show the strongest direct control, but articulated through ownership in only one and two firms respectively. Moreover, New York and London are also the places where the largest proportion of the ten companies can be traded at a stock exchange. Ownership of the largest non-fuel mineral producers can be categorised into five different types. One category consist entirely of APS firms, i.e. pension and investment funds as well as nominee accounts, and if a firm is partly owned by another public corporation, these may include APS firms. However, a large share of these ten mining companies is owned by small individual shareholders, who have not transferred their vote to a fund manager or nominee, large individual shareholders, who are in some cases in very close control of the corporations, other non-APS firms including other mining companies as well as governments. These owners can exercise their control and command functions as voters in the companies' general assemblies, largely outside the realm of the producer services. If such a shareholder holds more than a silent interest in the company, he/she or the company will have considerable direct control. Braudel's conception of power through flexibility has an explicitly spatial element to it, because much flexibility is derived from seeking new locations for the production of 162 goods and services under monopolistic conditions. While portfolio investment clearly exhibits this characteristic of spatial flexibility, it is rarely in control, owning relatively small shares of a specific enterprise and thus remains dependent on the successful management of a company by its executive. Investors can withdraw by selling shares, but rarely is there enough collective influence to change the course of action of a company. I have therefore attributed power in corporate decision making to a combination of ownership power and the power of the board room. In terms of investigating the resource industry in light of the world city concept, two points stand out. First, companies involved in the resource business are part of the larger monopoly over natural resources and these remain an essential ingredient for the global economy. Second, large diversified mining companies exhibit considerable spatial flexibility with mining and mineral rights assets stretching across all continents. This means these firms have the spatial flexibility to switch assets while at the same time exercising ownership and direct decision making power over these assets within a monopolistically or more correctly oligopolistically organised sector. When Anglo American and Gencor (now part of BHP Billiton) had the opportunity to break out of the territorial confines of the South African state, both decided to focus on the mining business, selling non-core business subsidiaries to diversify and expand spatially. They are now the second and third largest mining companies in the world controlled from Melbourne and London. In order to analyse how capital is actually commanded, I investigated the case of the platinum industry and traced the world's entire platinum production to its ownership, broken down to the shareholders of each of the companies owning the mines. While I undertook this analysis within the theoretical framework of world city research, I consider it 'bottom-up' world city research, starting at the mine level and tracing the command and control functions over the assets from the place of production to the place of control, rather than assuming some abstract control mechanisms. I call this vertical world city research (see Smith, 2001 for a similar approach). Control over platinum production is concentrated in Johannesburg, London and Moscow, with other cities in South Africa, New York, Zug/Baar (Switzerland), Sao Paulo, Sydney and Pretoria playing important roles as well. This is a diverse list of cities stretched over five continents, many of which are often included in the lower echelons of 163 world city research, but the list shows only three places in the traditional core countries and more in so called emerging markets, middle-income countries or the third and former second world. Looking at this list of cities, the question arises, is there a new spatial order in the world economy emerging? It would clearly be far too ambitious to claim the identification of a new spatial order in the world economy, based on these data from the platinum mining industry. However, this list is an indication of what could be a shifting geography of control functions over the global economy. The lists of cities produced in Chapters 3 and 4 include a diverse set of cities with typical world cities, including New York, London, Tokyo and Paris mixed in with other small and large places across the world. The other places included in both lists represent to a very large degree the so called BRICS (Brazil, Russia, India, China and South Africa).24 There are also cities in Chile, Australia, Canada, the USA, Mexico and Switzerland included, but there is no indication of any substantial control being held in low-income countries. The emergence of substantial ownership control in cities in middle-income countries, points to an important shift in global economic might. These countries retained relatively high growth rates in the wake of the 2008 financial crisis, while the traditional core countries show much slower recovery. The firms included to produce some of the most influential world city rankings25 are grouped into the six categories accountancy, advertising, banking/finance, insurance, law and management/consultancy. These are all APS firms that provide services to other firms, but these producer services firms likely service some types of companies more than others. I am suggesting that these APS firms are more likely to be directly involved in managing financial assets as opposed to material assets such as mines or factories. As suggested earlier, control over large resource firms can largely be exercised outside the influence of the APS. By conducting vertical or 'bottom-up' world city research, it was possible to identify the cities or places of direct control over specific assets, rather than relying on the proxy of APS firms to establish which places exercise command and control 24 Often this list does not include South Africa (then referred to as BRIC), but this focus on the resource industry clearly warrants the inclusion of South Africa. No city in India is actually included either in Chapter 3 or 4, but if the focus were on a different industry, e.g. certain high tech areas or steel production, India would likely be included. 25 These are the firms used in data set 11 of the Globalization and World City Research Network (GaWC). 164 functions. In one case the starting point was the ten world's largest non-fuel mineral producers, with a diverse set of assets distributed around the globe. In the other case the focus was on the much more closely defined platinum industry, but the starting point was the mine level, which allowed for more detailed assessment of ownership power. These two constructions of an urban hierarchy represent the command and control functions over a specific monopoly that Amin (1997, pp. 3- 5) identifies as one of five monopolies exercised by the core countries. The geographic distribution of these cities does not, however, represent Amin's spatiality of "polarisation at a global scale", which is basically a division of the capitalist world (now practically the entire globe) into two. Instead, the cities are truly spread out across the globe, even if some areas show slightly stronger representation than others, but it is not a division into the Global South and the Global North or the Third World and the First World. What this list of cities depicts though is a dotted landscape of urban control points over resource extraction across the world. These dots are located in middle and high income countries and distributed around the world and they are not clustered in the typical North America - Western Europe - Asia-Pacific landscape. While this landscape still shows a pattern of polarisation, it is not a large territory cut in two. Instead it is a landscape with the centres of control spread out but networked on the one side and the spaces in between, where surplus value is extracted, on the other. This is a global landscape that resembles the world of Marx's times more than that of Lenin's times (see Amin, 1994, p. 88). At the time of Marx's major writing, the property owning class extracted surplus value through continuous pressure on wages around the world. From around the 1880s onwards, however, wages roughly kept pace with or exceeded economic growth in the industrialised world facilitated by the historical compromise between capital and labour, while large surpluses were still extracted from labour in the third world by continuous pressure on wages. Thus the benefits received by the working class in the core countries were - at least partly - the result of third world people's labour. It was an unequal exchange (e.g. Wallerstein, 2004) or polarisation at a global scale (Amin, 1974). The landscape of ownership control over resource extraction identified here suggests that there is a re-ordering of polarisation under way, one that is polarised between the wealthy pockets in cities around the world and the peripheral location of resource extraction and other surplus value production. 165 There are remarkable similarities to Frank's (1969, p. 6) metropolis-satellite structure, but the metropolis is dispersed amongst urban centres around the world and is no longer comprised of the large territorial space of North America and Western Europe. Surely, there remain income inequalities at the global scale, but the income inequalities are also rising within the traditional core countries, which were characterised for a long time by relative equality facilitated by the Keynesian-Fordist post war World War II compromise and dismantled through neo-liberal reforms from the late 1970s onwards. And, as previous world city research indicates, cities around the world are now networked to form what can be conceptualised as a single space of flows, albeit with a lot of variation within this space. The world city network is the new metropolis, but my argument is that this world city network spans further than the networks constructed on the basis of APS firms suggest and that the basing points for capital, like Johannesburg and Sao Paulo (important in the resource industry) or other cities (for other sectors), have considerable control and command functions. Nonetheless, this structure still includes what Frank termed the satellites. These are smaller more peripheral cities, perhaps hosting regional headquarters or national offices of large firms and further down the chain smaller offices, which are often responsible for the implementation of strategic decisions, like labour relations or local land acquisitions. There are different levels of control. One level can be characterised as strategic control, another as operational control and a third one as the level of implementation. The example of Johannesburg in this dissertation shows that cities can of course also accommodate a combination of these functions. Moreover, the linkages to the world city network, the new networked metropolis, can hardly be escaped anywhere. It is a world spanning system and farming communities, occupying land in the platinum rich Bushveld Complex in South Africa have felt this, as I have shown in Chapter 6. The connection to London through the ownership of Anglo- Platinum by Anglo American and back to Johannesburg through a large (about 35 - 40 percent) shareholder base in South Africa is inescapable. But, what facilitates these connections to no small degree is not only the network of world cities, it is also the mosaic of nation states. States have a profound interest, although not always explicitly, to position at least one of their cities as a central node in the global circuit of capital, making it a world city. The desperation for foreign direct investment has reduced state power in 166 some places, but it did not render the nation state obsolete. Johannesburg's development as a  world city, for instance, was largely directed by the mode of South Africa's reintegration into the global economy with a focus on neo-liberal growth trajectories like GEAR and the retraction from the nationalisation policies of the freedom charter, as for example the focus on market solutions for BEE. Another contribution that vertical world city research makes to the development of world city research in general is that it provides a way of tracing the production of wealth from the place of surplus extraction, in this study a mine, to the place where the dividends from a firm's profit are received. The platinum industry derives much of its wealth from South Africa's Bushveld Complex. This is where the mines and minerals are and it is where the labour by mine workers is performed. Other sites in the platinum industry include parts of Russia, the Great Dike in Zimbabwe and a few other locations, where platinum is often mined as a by-product of other minerals. The sites where the control over the mines is exercised are Johannesburg, London, Moscow, New York, Zug, Sao Paulo, Sydney, Pretoria, Pilanesburg and Zurich to name the ten most influential control centres. Only one city, Pilanesburg, can be considered directly within a mining community, the others are scattered around the globe, although the place that assembles the largest share of control, Johannesburg, is within South Africa. However, although mine and city are in the same country, there is likely such an enormous social distance between the mines and the owners, they could as well be in different countries. This list includes mainly world cities, even if not usually identified as at the top of the hierarchy, but the spread of wealth may well go beyond these cities. Often owners are identified as investment and pension funds or nominee accounts. In those cases, the decision making power likely lies in the cities, but the receivers of the dividend may well be somewhere else, including small cities or rural areas. For this to happen, the world city may simply be a switchboard of processes and a redistributor of wealth. Interestingly, much of the wealth created by Anglo American is still produced in South Africa, where the company has most of its assets. It is then channelled through the head office in London and redistributed to its shareholders, of which approximately 40 percent are in South Africa. In this sense, the world city never ends. It is not bounded. It is a place, where many strings come together and wealth gets redistributed - from the poor to the rich that is - by efficiently organising the extraction of surplus value. The world city thus 167 depends on the labour and material input from the periphery to be what it is. It is a space of flows. In this sense all cities are ordinary and extraordinary at the same time, but there are power differentials between places and vertical world city research provides a tool to identify these power differentials. The decision making power concentrated in world cites puts these cities into a central position in the command and control structure over the global economy. But just because a business deal is concluded at a fine Italian restaurant, does not mean that the restaurant is in control of the deal; likewise, just because business is conducted in certain cities, does not mean that these cities are in control of the global economy, they facilitate control and thus have some leverage. However, collectively the actors getting together in these cities have considerable control. At the same time world cities are utterly dependent on the input from their peripheries. Just as Richard Smith's (2007) example of the CEO, who would be powerless, if all his assisting devices and human assistants would be stripped away from him, world cities would not be the places of power without their relations to the periphery, but the structures of uneven development (e.g. Smith, 1984) maintain the power - at least for now. However, asymmetries are produced through specific spatial and social relations and given the world city's dependency on the productive input from the periphery, there is considerable collective power in the territories between the urban control points doting the globe. The idea that all cities should follow the development trajectories of current world cities is nonsense and given the uneven relationships and flows of wealth it is also simply not possible. Moreover, establishing "some cities as exemplar and others as imitators" (Robinson, 2006, p. 94) is not what world city research set out to do. World city research can help identify uneven development trajectories at a global scale and analyse the effects such relationships have on places around the world. While communities around the world may struggle with the power of large corporations that want to extract resources from their land and labour, decision makers in world cities struggle to maintain their power position in the network by constantly repositioning their cities at specific intersections of global capital flows. This dissertation focused on the control over global production processes through unequal exchange and global monopolies maintained by the core countries. It provided an empirical analysis of the non-fuel mineral sector with a particular focus on the platinum 168 industry. While this focus analyses an important monopoly, it is nonetheless a partial picture. An analysis of other monopolies and monopolistic processes of economic, political, social and cultural control will be required to arrive at a picture that is closer to completeness. However, the spatial pattern of control identified here points at a different power configuration from that identified on the basis of producer services and suggests that there is a new spatial order in the world economy emerging. This spatial order is not a world divided into two large territories, such as a core and a periphery, a centre or global metropolis and the resource hinterland or simply the First and Third World. Today the world cities are the central locations, where the surplus produced elsewhere is received and redistributed and these world cities are scattered around the globe, including the middle income countries. But in some ways the world city never ends, because some important actors reside far from world cities, but use them as switchboards for the distribution of economic surplus and to disseminate their decisions on investments. American billionaire investor Warren Buffet, for example, lives and works in Omaha, Nebraska, even though his trading and investments go through Wall Street. On the other hand mining communities in South Africa's Limpopo province that want to register their protest against the mining activities on their land have to go through the same network of cities to reach the decision makers. At the same time the world cities themselves are home to both core and periphery type economic processes that often occur literally next door to each other. These are new socio-spatial configurations, but they are still based on an old division of labour, where some have to sell their surplus labour and others receive the profits. 169 Bibliography Afshar, F. (1998). Balancing global city with global village. Habitat International, 22(4), 375-387. Alderson, A.S. & Beckfield, J. (2004). Power and position in the world city system. American Journal of Sociology, 109(4), 811-851. Amen, M., Aecher, K. & Bosman, M. (Eds.). (2006). Relocating global cities: from the center to the margins. Lanham: Rowman & Littlefield. Amin, A. & Graham, S. (1997). The ordinary city. Transactions of the Institute of British Geographers, 22(4), 411-429. Amin, S. (1997). Capitalism in the age of globalization: the management of contemporary society. London: Zed Books. Amin, S. (1994). Re-Reading the postwar period: an intellectual itinerary (Transl. M. Wolfers). New York: Monthly Review Press. Amin, S. (1974). Accumulation on a world scale: a critique of the theory of underdevelopment (transl. Brian Pearce). New York: Monthly Review Press. Anglo American. (2010). Anglo American AGM 2010: notice of annual general meeting and shareholder information. [on-line]. Available: m/anglo_nom.pdf . [last accessed: 2010 December 24]. Anglo American. (2011). Where we operate. [on-line]. Available: . [last accessed: 2011 January 28]. Anglo Platinum. (2008, March). Anglo Platinum’s response to ActionAid allegations. Johannesburg: Anglo Platinum. Armstrong, W. & McGee, T.G. (1985). Theatres of accumulation: studies in Asian and Latin American urbanization. London: Methuen. Arrighi, G. (1994). The long twentieth century: money, power and the origins of our times. New York: Verso. Axford, B. (2005). Critical globalization studies and a network perspective on global civil society. In: R.P. Appelbaum & W.I. Robinson (eds.). Critical globalization studies (pp. 187-196). New York: Routledge. 170 Baran, P. & Sweezy, P. (1966). Monopoly capital: an essay on the American economic and social order. New York: Monthly Review Press. Barnes, T.J. (2001) Retheorizing Economic Geography: From the Quantitative Revolution to the “Cultural Turn”. Annals of the Association of American Geographers, 91(3), 546-565. Beall, J., Crankshaw, O. & Parnell, S. (2002). Uniting a divided city: governance and social exclusion in Johannesburg. London: Earthscan Publications. Beall, J., Crankshaw, O. & Parnell, S. (2000). Local government, poverty reduction and inequality in Johannesburg. Environment and Urbanization, 12(1), 107-122. Beall, J., Gelb, S. & Hassim, S. (2005). Fragile stability: state and society in democratic South Africa. Journal of Southern African Studies, 31(4), 681-700. Beaverstock, J.V. (1996). Migration, knowledge and social interaction: expatriate labour within investment banks. Area, 28, 459–470. Beaverstock, J.V. (1990). A new international division of professional and managerial labour: some explorations. Unpublished PhD thesis, University of Bristol. Beaverstock, J.V., Doel, M.A., Hubbard, P.J. & Taylor, P.J. (2002). Attending to the world: competition, cooperation and connectivity in the World City network. Global Networks, 2(2), 111-132. Beaverstock, J.V., Smith, R.G. (1996). ‘Lending jobs to global cities’: skilled international labour migration, investment banking and the City of London. Urban Studies, 33, 1377–1394. Beaverstock, J.V., Smith, R.G. &Taylor, P.J. (2000). World-city network: a new metageography? Annals of the Association of American Geographers, 90(1), 123-134. Beaverstock, J.V., Taylor, P.J. & Smith, R.G. (1999). A roster of world cities. Cities, 16(6), 445-458. Beaverstock, J.V., Smith, R.G., Taylor, P.J., Walker, D.R.F. & Lorimer, H. (2000). Globalization and world cities: some measurement methodologies. Applied Geography, 20(1), 46-63. 171 Beavon, K. (2006). Johannesburg 1986-2030: a quest to regain world status. In M.M. Amen, K. Archer & M.M. Bosman (Eds.). Relocating Global Cities: from the center to the margins (pp. 49-73). Lanham: Rowman & Littlefield. Beavon, K. (2004). Johannesburg: The Making and Shaping of the City. Pretoria: UNISA Press. Beavon, K. (1998). Johannesburg: Coming to grips with globalization from an abnormal base. In F.-C. Lo & Y.-M. Yeung (Eds.). Globalization and the world of large cities (pp. 352-388). Tokyo: United Nations University Press. Berle, A.A. & Means, G.C. (1933). The modern corporation and private property. New York: MacMillan. Blaut, J. (1976). Where was capitalism born? Antipode, 8(2), 1-11. Bond, P. (2008). South Africa's 'developmental state' distraction. Mediations, 24(1), 8-27. Bond, P. (2004). The ANC's 'Left Turn' & South African sub-imperialism. Review of African Political Economy, 31(102), 599 - 616. Bond, P. (2000a). Cities of gold, townships of coal: Essays on South Africa’s New Urban Crisis. Trenton, NJ and Asmara, Eritrea: Africa World Press. Bond, P. (2000b). Elite transition: from apartheid to neoliberalism in South Africa. London: Pluto Press. Braudel, F. (1992 [1984]). Civilisation and Capitalism, 15th - 18th century. Berkeley: University of California Press. Brenner, N. (2004). New state spaces: urban governance and the rescaling of statehood. Oxford: Oxford University Press. Brenner, N. (1998). Global cities, glocal states: global city formation and state territorial restructuring in contemporary Europe. Review of International Political Economy, 5(1), 1-37. Brenner, N. & Theodore, N. (2002). Cities and the geographies of “Actually Existing Neoliberalism”. Antipode, 34(3), 349–379. Brown, E., Derudder, B., Parnreiter, C., Pelupessy, W., Taylor, P.J. & Witlox, F. (2010). World City Networks and Global Commodity Chains: towards a world-systems' integration . Global Networks 10(1), 12-34. 172 Brown, T.J., Bide, T., Hannis, S.D., Idoine, N.E., Hetherington, L.E., Shaw, R.A., Walters, A.S., Lusty, P.A.J. & Kendall, R. (2010). World Mineral Production 2004 - 2008. Keyworth, Nottingham: British Geological Survey. Burawoy, M. (2000). Introduction: Reaching for the Global. In: M. Burawoy et al. Global ethnography: forces, connections, and imaginations in a postmodern world. Berkeley: University of California Press, pp. 1-40. Burawoy, M. (1991). Ethnography unbound: power and resistance in the modern metropolis. Berkeley: University of California Press. Burt, R.S. (1983). Corporate profits and cooperation. New York: Academic Press. Cairncross, F. (1997). The Death of distance: how the communications revolution will change our Lives. Boston, Massachusetts: Harvard Business School Press. Carmody, P. (2002). Between globalisation and (post) apartheid: the political economy of restructuring in South Africa. Journal of Southern African Studies, 28(2), 255-275. Carmody, P.R. & Owusu, F.W. (2007). Competing hegemons? Chinese versus American geoeconomic strategies in Africa. Political Geography, 26(5), 504-524. Castells, M. (2001). The Internet Galaxy: Reflections on the Internet, Business, and Society. Oxford: Oxford university Press. Castells, M. (1997). The power of identity. Malden, MA: Blackwell. Castells, M. (1996). The rise of the network society. Cambridge, MA: Blackwell. Castells, M. (1989). The informational city: information technology, economic restructuring, and the urban-regional process. Oxford, UK: Blackwell. Codelco. (2010a). The Corporation: History. [on-line]. Available: . [last accessed: 2010 December 16]. 173 Codelco. (2010b, November 15). The Corporation: FAQ. [on-line]. Available: Last accessed: 2010, November 15. Cowen, M.P. & Shenton, R.W. (1996). Doctrines of Development. London: Routledge. Cox, B.A. & Rogerson, C.M. (1985). The corporate power élite in South Africa: interlocking directorships among large enterprises. Political Geography Quarterly, 4(3), 219-234. Crang, P. (1997). Cultural turns and the (re)constitution of economic geography: Introduction to section one. In R. Lee and J. Wills (eds.). Geographies of Economies. London: Arnold, pp. 3–15. Crankshaw, O. & Parnell, S. (2004). Johannesburg: race, inequality, and urbanization. In: J. Gugler (ed.). World Cities beyond the west: globalization, development, and inequality. Cambridge: Cambridge University Press, pp. 348-370. Curtis, M. (2008, March). Precious Metal: The impact of Anglo Platinum on poor communities in Limpopo, South Africa. Johannesburg: Action Aid. Curtis, M. (2007, August). Anglo American: The alternative report. London: War on Want. Daniels, P. W. (1993). Service industries in the world economy. Oxford: Blackwell. Davis, D.E. (2005). Cities in global context: a brief intellectual history. International Journal of Urban and Regional Research, 29(1), 92-109. de Angelis, M. (2001, September 2). Marx and primitive accumulation: the continuous character of capital’s ‘enclosures’. The Commoner , 2. [on- line]. Available: . [last accessed: 2011 October 17]. Department of Mineral Resources. (2010, September). Amendment of the Broad-based Socio-economic Empowerment Charter for the South African Mining and Minerals Industry. Pretoria: Republic of South Africa. Department of Mineral Resources. (2009, October). Mining Charter Impact Assessment Report. Pretoria: Republic of South Africa. 174 Derudder, B. (2008). Mapping global urban networks: a decade of empirical world cities research. Geography Compass, 2(2), 559–574. Derudder, B. & Taylor, P. (2005). The cliquishness of world cities. Global Networks, 5(1), 71–91. Derudder, B, Taylor, P.J., Witlox, F. & Catalano, G. (2003). Hierarchical Tendencies and Regional Patterns in the World City Network: A Global Urban Analysis of 234 Cities. Regional Studies, 37(9), 875–886. Derudder, B. & Witlox, F. (2010). Commodity chains and world cities. Chichester: Wiley-Blackwell. Derudder, B., Witlox, F. & Taylor, P.J. (2007). U.S. cities in the world city network: comparing their positions using global origins and destinations of airline passengers. Urban Geography, 28(1), 74-91. Dicken, P., Kelly, P., Olds, K. & Yeung, H. W.-C. (2001). Chains and networks, territories and scales: towards a relational framework for analysing the global economy. Global Networks, 1(2), 89-112. Duménil, G. & Lévy, D. (2004a). Capital resurgent: roots of the neoliberal revolution (Trans. D. Jeffers). Cambridge, Mass.: Harvard University Press. Duménil, G. & Lévy, D. (2004b). Neo-liberal dynamics – towards a new phase? In: K. van der Pijl, L. Assassi & D. Wigan. (Eds.). Global regulation: managing crises after the imperial turn. Hampshire, UK: PalgraveMacMillan, pp. 28 – 42. Emmanuel, A. (1972). Unequal exchange: a study of imperialism of trade. New York: Monthly Review Press. En+ Group. (2008). About Us. [on-line]. Available: . [last accessed: 2010 December 16]. Engdahl, C. & Hauki, H. (2001). "Black Economic Empowerment: an introduction for non-South African businesses. Unpublished Master Thesis, Department of Law, School of Economics and Commercial Law, Gothenburg University, Sweden. Ericsson, M. (2010). Global mining towards 2030: Food for thought for the Finnish mineral policy process 2010, Report of Investigation 187. Espoo: Geological Survey of Finland. 175 Farlex, Inc. (2010). Capitalization. In The Free Dictionary. [on-line]. Available: http://financial- Last accessed: 2010 November 12. Feinstein, C.H. (2005). An economic history of South Africa: conquest, discrimination and development. Cambridge: Cambridge University Press. Ferguson, J. (2006). Global shadows: Africa in the neoliberal world order. Durham: Duke University Press. Ferreira, E. (2011, February 21). Moeletsi Mbeki: 'Ideologically, the ANC is stuck'. Mail & Guardian Online. Fine, B. & Rustomjee, Z. (1996). The political economy of South Africa: from minerals-energy complex to industrialisation. London: Hurst & Company. Frank, A.G. (1998). ReORIENT: global economy in the Asian age. Berkeley: University of California Press. Frank, A.G. (1972). The development of underdevelopment. In J.D. Cockcroft, A.G. Frank and D.L. Johnson (Eds.), Dependence and underdevelopment (pp. 3-17). Garden City, N.Y.: Anchor Books. Frank, A.G. (1969). Latin America, underdevelopment or revolution: essays on the development of underdevelopment and the immediate enemy. New York: Monthly Review Press. Frank, A.G. & Gills, B. (Eds.). (1993). The world system: five hundred years or five thousand? London: Rotledge. Friedman, T.L. (2007). The world is flat: a brief history of the twenty-first century. Vancouver: Douglas and McIntyre. Friedmann, J. (1995). Where we stand: a decade of world city research. In P.L. Knox & P.J. Taylor (eds.). World Cities in a World System. Cambridge: Cambridge University Press, pp. 21 - 47. Friedmann, J. (1986) The world city hypothesis. Development and Change 17, 69–83. Friedmann, J. & Wolff, G. (1982). World city formation: an agenda for research and action. International Journal of Urban and Regional Research, 6(3), 309-344. 176 FTSE. (2003, May 12). Gencor (South Africa) unbundling of shareholding in Impala Platinum (South Africa) Changes In FTSE Indices. [on-line]. Available: runbundlingofImpala.jsp;jsessionid=D80B3B6468382BA860FA652DF E286C7D.tc1 . [last accessed: 2011 March 10]. Geddes, P. (1915). Cities in evolution: an introduction to the town planning movement and to the study of civics. London: Benn. Gibb, R. (2006). The new Southern African Customs Union Agreement: dependence with democracy. Journal of Southern African Studies, 32(3), 583-603. Glassman, J. (2006). Primitive accumulation, accumulation by dispossession, accumulation by ‘extra-economic’ means. Progress in Human Geography 30(5), 608–625. Glassman, J. (2003a). Structural power, agency and national liberation: the case of East Timor. Transactions of the Institute of British Geographers, 28(3), 264 – 280. Glassman, J. (2003b). Rethinking Overdetermination, Structural Power, and Social Change: A Critique of Gibson-Graham, Resnick, and Wolff. Antipode, 35(4), 678 – 698. Gleason, D. & Levin, A. (1999a, February 26). Breaking out of wedlock - Unbundling 5: Anglo American Corp SA. Special Report. Financial Mail. [on-line]. Available: [last accessed: 2011 March 8]. Gleason, D. & Levin, A. (1999b, February 26). Breaking out of wedlock - Unbundling 2: Gencor. Special Report. Financial Mail. [on-line]. Available: [last accessed: 2011 March 8]. Glencore International AG. (2010). Glencore Fact Sheet. [on-line]. Available: [last accessed: 2010 December 17]. Godfrey, B. & Zhou, Y. (1999). Ranking world cities: multinational corporations and the global urban hierarchy. Urban Geography, 20(3), 268-281. 177 Gordon, I. (2006). How should we write about London? The working capital view. City - analysis of urban trends, culture, theory, policy, action, 10(2), 185 - 196). Government of South Africa. (2002, October 11). Broad-Based Socio- Economic Empowerment Charter for the South African Mining Industry. [on-line]. Available: . [last accessed: 2011 August 26]. Graham, S. & Marvin, S. (2001). Splintering Urbanism: networked infrastructures, technological mobilities and the urban condition. London: Routledge. Grant, R. & Nijman, J. (2002). Globalization and the Corporate Geography of Cities in the Less-Developed World. Annals of the Association of American Geographers, 92(2), 320–340. Gugler, J. (Ed.). (2004). World cities beyond the west: globalization, development and inequality. Cambridge: Cambridge University Press. Habib, A., Pillay, D. & Desai, A. (1998). South Africa and the global order: The structural conditioning of a transition to democracy. Journal of Contemporary African Studies, 16(1), 95-115. Hall, P. (1979 [1966]). World Cities (1st McGraw-Hill paperback edition). New York: McGraw-Hill. Hamann, R. (2009). South Africa: the role of history, governement, and local context. In S.O. Idowu & W.L. Filho. (Eds.). Global practices of corporate social responsibility. Berlin, Heidelberg: Springer. Hamann, R., Khagram, S. & Rohan, S. (2008). South Africa's Charter Approach to Post-Apartheid Economic Transformation: Collaborative Governance or Hardball Bargaining? Journal of Southern African Studies, 34(1), 21-37. Hannerz, U. (1996). Transnational connections: culture, people, places. London: Routledge. Hannerz, U. (1992). The global ecumene as a network of networks. In: A. Kuper (ed.). Conceptualizing society. London: Routledge. Harrison, P., Huchzermeyer, M., & Mayekiso, M. (Eds.) (2003). Confronting fragmentation: Housing and urban development in a democratising society. Cape Town: UCT Press. 178 Hart, G. (2002). Disabling globalization: places of power in post-apartheid South Africa. Berkeley: University of California Press. Harvey, D. (2006). Spaces of global capitalism: towards a theory of uneven geographical development. London: Verso. Harvey, D. (2005). A brief history of neoliberalism. Oxford: Oxford University Press. Harvey, D. (1989a). The condition of postmodernity: an enquiry into the origins of cultural change. Oxford: Blackwell. Harvey, D. (1989b). From manegiralism to entrepreneurialism: the transformation in urban governance in late capitalism. Geografiska Annaler, 71B(1), 3-17. Harvey, D. (1982). The limits of capital. Oxford: Basil Blackwell. Held, D. (2000). Regulating Globalization? The Reinvention of Politics. International Sociology 2000, 15(2), 394-408. Herman, E.S. (1981). Corporate control, corporate power : a Twentieth Century Fund study. Cambridge: Cambridge University Press. Hopkins, T.K. and Wallerstein, I. (1986). The age of transition. London: Pluto Press. Innes, D. (1984). Anglo American and the rise of modern South Africa. London: Heinemann. International Metalworkers' Federation. (2010). Trade Unions Ask Norwegian Pension Funds to Divest of Grupo Mexico Shares. [on-line]. Available: . [last accessed: 2010 December 16]. Jamieson, B. (1990). Goldstrike! The Oppenheimer empire in crisis. London: Hutchinson Business Books. Jessop, B. (2002). Liberalism, Neoliberalism, and Urban Governance: A State-Theoretical Perspective. Antipode, 34(3), 452-472. Jessop, B., Brenner, N. & Jones, M. (2008). Theorizing sociospatial relations. Environment and Planning D: Society and Space, 26(3), 389-401. Jollie, D. (2010, May). Platinum 2010. Hertfordshire: Johnson Matthey. 179 Jones, J.D.F. (1995). Through fortress and rock: the story of Gencor 1895- 1995. Johannesburg: Jonathan Ball. Jordan, P. (2004). The African National Congress: from illegality to the corridors of power. Review of African Political Economy, 31(100), 203- 212. Katz, C. (2001). On the grounds of globalization: a topography for feminist political engagement. Signs: Journal of Women in Culture and Society, 26(41), 1213-1234. King, A. (1990). Urbanism, colonialism, and the world-economy: cultural and spatial foundations of the world urban system. London: Routledge. Korff, R. (1987). The world city hypothesis: a critique. Development and Change, 18, 483-495. Laclau, E. (1971). Feudalism and Capitalism in Latin America. New Left Review, I/67(May-June), 19-38. Lagos, G. (1997). Developing National Mining Policies in Chile: 1974-96. Resources Policy, 23(1/2), 51-69. Lai, K.P.Y. (2009). New spatial logics in global cities research: networks, flows and political spaces. Geography Compass, 3(3), 997-1012. Larner, W. & Le Heron, R. (2002). From economic globalisation to globalising economic processes: towards post-structural political economies. Geoforum, 33(4), 415–419. Lieberfeld, D. (2002). Evaluatingt he Contributionso f Track-Two Diplomacy to Conflict Termination in South Africa, 1984-90. Journal of Peace Research, 39(3), 355-372. London Platinum and Palladium Market. (2009a). About The Fixings. [on- line]. Available: . [last accessed: 2011, February 7]. London Platinum and Palladium Market. (2009b). The Market. [on-line]. Available: . [last accessed: 2011, February 7]. Lynch, J., & Meyer, D. R. (1992). Dynamics of the US system of cities, 1950 to 1980: the impact of the large corporate law firm. Urban Affairs Quarterly, 28, 38–68. 180 Mabin, A. (2007). Johannesburg: (South) Africa's aspirant global city. In: K. Seghers (Ed.). The making of global city regions: Johannesburg, Mumbai/Bombay, São Paulo, and Shanghai. Baltimore: Johns Hopkins University Press, pp. 32-63. Mail and Guardian. (2010, September 14). JSE's BEE study is misleading, says BMF. Mail & Guardian Online. Makana, C. (2009 October 21). Communities in bid to stop coal mining scheme. The Sowetan. Marcuse, P. (2005). ‘The city’ as perverse metaphor. City - analysis of urban trends, culture, theory, policy, action, 9(2), 247-254. Marcuse, P. & van Kempen, R. (Eds.). (2000). Globalizing cities: a new spatial order? Oxford: Blackwell. Mawdsley, E. (2007). China and Africa: Emerging Challenges to the Geographies of Power. Geography Compass 1(3), 405–421. McDonald, D.A. (2008). World City syndrome: neoliberalism and inequality in Cape Town. New York: Routledge. McFarlane, C. (2010). The comparative city: knowledge, learning, urbanism. International Journal of Urban and Regional Research, 34(4), 725-742. McFarlane, C. (2008). Urban shadows: materiality, the ‘southern city’ and urban theory. Geography Compass, 2(2), 340-358. Merten, M. (2005, May 4). No right to rights, says govt. Mail & Guardian Online. Meyer, D.R. (2003). The challenges of research on the global network of cities. Urban Geography, 24(4), 301-313. Ministry of Finance. (2010). Companies Excluded from the Investment Universe. [on-line]. Available: pension-fund/responsible-investments/companies-excluded-from-the- investment-u.html?id=447122 . [last accessed: 2010 December 16]. Mitchell, J.C. (1987). Cities, societies and social perception: a Central African perspective. Oxford: Clarendon. Modimoeng, K. (2009, August 5). NUM offended by apartheid treatment. The Sowetan. 181 Mohideen, K. (2010, May). London Mining Network briefing: An introduction to Anglo American plc. [on-line]. Available: content/uploads/2010/05/Introduction-to-Anglo-American.doc . [last accessed: 2011 September 26]. Murdoch, J. (1998). The spaces of actor-network theory. Geoforum, 29, 357- 374. Murdoch, J. (1997a). Towards a geography of heterogeneous associations. Progress in Human Geography, 21(3), 321-337. Murdoch, J. (1997b). Inhuman/non human/human: actor-network theory and the prospects for a non-dualistic and symmetrical perspective on nature and society. Environment and Planning D: Society and Space, 15, 731- 756. Murray, M.J. (2004). The evolving spatial form of cities in a globalizing world economy: Johannesburg and Sao Paulo. Cape Town: HSRC Publishers. Newman, P. & Thornley, A. (2005). Planning world cities: globalization and urban politics. Hampshire, UK: Palgrave MacMillan. Newmarch, J. (2007, September 10). Billion-dollar brother. Mail and Guardian. Norilsk Nickel. (2010a, May 27). Annual Report 2009: Focus on Efficiency. Moscow: Norilsk Nickel. Norilsk Nickel. (2010b). History. [on-line]. Available: . [last accessed: 2010 December 22]. Norlund, C. (2004). A critical comment on the Taylor approach for measuring world city interlock linkages. Geographical Analysis, 36(3), 290-296. Öncü, A. & Weyland, P. (Eds.). (1997). Space, culture and power: new identities in globalizing cities. London: Zed Books. Padayachee, V. (2009). Development issues for the future: A South African perspective. European Journal of Development Research 21, 15–20. Padayachee, V. (2003). Capital, the state and globalization: South African Conglomerates at the end of the nineteenth and twentieth centuries. In: K.S. Jomo & Khoo Khay Jin (Eds.). Globalization and its discontents, revisted. New Delhi: Tulika Books. 182 Parnell, S. (2007). Politics of transformation: defining the city strategy in Johannesburg. In: K. Seghers (Ed.). The making of global city regions: Johannesburg, Mumbai/Bombay, São Paulo, and Shanghai. Baltimore: Johns Hopkins University Press, pp. 139-167. Parnreiter, C. (2002). Mexico City: the making of a global city?. In S. Sassen (Ed.), Global networks, linked cities (pp. 145-182). New York: Routledge. Peck, J. (2002). Political Economies of Scale: Fast Policy, Interscalar Relations, and Neoliberal Workfare. Economic Geography, 78(3), 331- 360. Peck, J. & Theodore, N. (2010). Recombinant workfare, across the Americas: Transnationalizing "fast" social policy. Geoforum, 41(2), 195-208. Pfister, R. (2003). Gateway to international victory: the diplomacy of the African National Congress in Africa, 1960-1994. Journal of Modern African Studies, 41(1), 51-73. Pillay, D. (2008). Globalization and the informalization of labour: the case of South Africa. In: A. Bieler, I. Lindberg & D. Pillay (Eds.). Labour and the challenges of globalization: What prospects for transnational solidarity? London: Pluto Press, pp. 45-64. Pillay, D. (2004). Media diversity and the contested character of the post- apartheid state. Politikon, 31(2), 167–184. Pillay, U. (2008). Urban Policy in Post-Apartheid South Africa: Context, Evolution and Future Directions Urban Forum, 19(2), 109-132. Poon, J. (2003). Hierarchical tendencies of capital markets among international financial centres. Growth and Change, 34(2), 135-156. Reuters. (2008, February 1). FACTBOX - The world's biggest mining companies. [on-line]. Available: pageNumber=2&virtualBrandChannel=0&sp=true . [last accessed: 2009 September 21]. RNS. (2009, October 22). Relocation of Operational Headquarters. RNS - the company news service from the London Stock Exchange. Robinson, J. (2006). Ordinary cities: between modernity and development. London: Routledge. 183 Robinson, J. (2004). In the tracks of comparative urbanism: difference, urban modernity and the primitive. Urban Geography, 25(8), 709-723. Robinson, J. (2003a). Johannesburg's futures: beyond developmentalism and global success. In: R. Tomlinson, R.A. Beauregard, L. Bremner & X. Mangcu (Eds.). Emerging Johannesburg: perspectives on the postapartheid city. New York: Routledge, pp. 259-280. Robinson, J. (2003b). Johannesburg’s 1936 Empire Exhibition: Interaction, Segregation and Modernity in a South African City. Journal of Southern African Studies, 29(3), 759-789. Robinson, J. (2002). Global and world cities: a view from off the map. International Journal of Urban and Regional Research 26(3), 531-554. Rogerson, C.M. (2005). Globalization, economic restructuring and local response in Johannesburg - the most isolated 'world city'. In: K. Segbers, S. Raiser & K. Volkmann (Eds.). Public problems - private solutions? Globalizing Cities in the South. Aldershot: Ashgate, pp. 17- 34. Rosenberg, M. (2010, July 13). Secretive Mexico copper baron thrives despite crises. Thompson Reuters News. [on-line]. Available: . [last accessed: 2010 December 16]. Rostow, W.W. (1960). The stages of economic growth: a non-communist manifesto. Cambridge: Cambridge University Press. Rumney, R. (2009, November 27). No big deal. Mail & Guardian Online. Rusal. (2010). Shareholder Structure. [on-line]. Available: . [last accessed: 2010 December 16]. Samers, M. (2002). Immigration and the global city hypothesis: towards an alternative research agenda. International Journal of Urban and Regional Research, 26(2), 389-402. Sassen, S. (Ed.). (2002). Global networks, linked cities. New York: Routledge. Sassen, S. (2001 [1991]). The global city: New York, London, Tokyo (2nd. ed.). Princeton: Princeton University Press. Sassen, S. (2000). Territory and Territoriality in the Global Economy. International Sociology, 15(2), 372-393. 184 Seidman Makgetla, M. (2004). The post-apartheid economy. Review of African Political Economy, 31(100), 263-281. Seshoka, L. (2009, July 16). An Open Letter to the Chamber of Mines. [on- line]. Available: option=com_content&task=view&id=425&Itemid=79 ; . [last accessed: 2011 August 26]. Shankleman, J. (no date). Going global: Chinese oil and mining companies and the governance of resource wealth. Washington, DC: Woodrow Wilson International Center for Scholars. Sheppard, E. (2002). The spaces and times of globalization: place, scale, networks, and positionality. Economic Geography, 78(3), 307-330. Sheppard, E.S., Porter, P.W., Faust, D.R. & Nagar, R. (2009). A world of difference: encountering and contesting development. New York: Guilford Press. Short, J.R., Breitbach, C. Buckman, S. & Essex, J. (2000). From world cities to gateway cities: extending the boundaries of globalization theory. City - analysis of urban trends, culture, theory, policy, action, 4(3), 317-340. Short, J.R., Kim, Y., Kuus, M. & Wells, H. (1996). The dirty little secret of world city research: data problems in comparative analysis. International Journal of Urban and Regional Research, 20(4), 697-717. Sklair, L. (2001). The transnational capitalist class. Oxford: Blackwell. Sklair, L & Robbins, P.T. (2002) Global Capitalism and Major Corporations from the Third World. Third World Quarterly, 23(1), 81-100. Smith, D.A., and Timberlake, M. (1995a). Cities in global matrices: toward mapping the world-system’s city-system. In: Knox, P.L. and Taylor, P.J. (Eds.). World cities in a world-system (pp. 79–97). Cambridge, UK: Cambridge University Press. Smith, D.A., and Timberlake, M. (1995b). Conceptualising and mapping the structure of the world system’s city system. Urban Studies, 32(2), 287– 302. Smith, M.P. (2001). Transnational urbanism: locating globalization. Malden: Blackwell. 185 Smith, M.P. (1998). The Global City-whose social construct is it anyway?: a comment on White. Urban Affairs Review, 33(4), 482-488. Smith, N. (1984). Uneven development: nature, capital and the production of space. Oxford: Blackwell. Smith, R.G. (2007). Poststructuralism, power and the global city. In P. Taylor, B. Derudder, P. Saey and F. Witlox. (Eds.), Cities in globalization: practices, policies and theories (pp. 258-272). London: Routledge. Smith, R.G. (2003a). World city actor-networks. Progress in Human Geography, 27(1), 25-44. Smith, R.G. (2003b). World city topologies. Progress in Human Geography 27(5), 561–582. Smith, R.G. & Doel, M. (2011). Questioning the Theoretical Basis of Current Global-City Research: Structures, Networks and Actor-Networks. International Journal of Urban and Regional Research, 35(1), 24-39. Southall, R. (2004). The ANC & black capitalism in South Africa. Review of African Political Economy, 31(100), 313 - 328. South African Human Rights Commission. (2008). Mining-related observations and recommendations: Anglo Platinum, affected communities and other stakeholders, in and around the PPL Mine, Limpopo. Johannesburg: South African Human Rights Commission. Sowetan. (2010, August 19). Cosatu hits at 'elitist' BEE deals. The Sowetan. Swyngedouw, E. (2004). Globalisation or ‘glocalisation’? Networks, territories and rescaling. Cambridge Review of International Affairs, 17(1), 25-48. Taylor, P.J. (2005a). New political geographies: global society and global governance through world city networks. Political Geography, 24, 703- 730. Taylor, P.J. (2005b). Leading world cities: empirical evaluations of urban nodes in multiple networks. Urban Studies, 42(9), 1593-1608. Taylor, P.J. (2004a). World city network: a global urban analysis. London: Routledge. Taylor, P.J. (2004b). The new geography of global civil society: NGOs in the world city network. Globalizations, 1(2), 265-277. 186 Taylor, P.J. (2004c). Reply to “A Critical Comment on the Taylor Approach for Measuring World City Interlock Linkages” by C. Nordlund. Geographical Analysis, 36(3), 297-298. Taylor, P.J. (2001). Specification of the world city network. Geographical Analysis, 33(2), 181–94. Taylor, P.J. (2000). The 1999 Annual Political Geography Lecture: World cities and territorial states under conditions of contemporary globalization. Political Geography, 19(1), 5–32. Taylor, P.J. (1997). Hierarchical tendencies amongst world cities: a global research proposal. Cities, 14(6), 323-332. Taylor, P.J. (1996). The way the modern world works: world hegemony to world impasse. Chichester: John Wiley. Taylor, P.J., Catalano, G. & Walker, D.R.F. (2002). Measurement of the world city network. Urban Studies, 39(13), 2367-2376. Taylor, P.J., Hoyler, M., Walker, D.R.F. & Szegner, M.J. (2001). A new mapping of the world for the new millenium. The Geographical Journal, 167(3), 213-222. Taylor, P.J. & Walker, D. R. F. (2001). World Cities: A First Multivariate Analysis of their Service Complexes. Urban Studies, 38(1), 23– 47. Taylor, P.J., Walker, D.R.F., Catalano, G. & Hoyler, M. (2002). Diversity and power in the world city network. Cities, 19(4), 231-241. Teschke, B. (2003). The myth of 1648: class, geopolitics, and the making of modern international relations. London: Verso. Todes, A. (2000). Reintegrating the apartheid city? Urban policy and urban restructuring in Durban. In S. Watson, & G. Bridges (Eds.), A companion to the city. London: Blackwell,  pp. 617–630. United Nations Development Programme. (2009). Human Development Report 2009 Overcoming barriers: Human mobility and development. New York: Palgrave Macmillan. United States Geological Survey. (2010a, January). 2008 Minerals Yearbook Platinum-Group Metals [advance release]. [on-line]. Available: 2008-plati.pdf . [last accessed: 2011, January 18]. 187 United States Geological Survey. (2010b). Mineral Commodity Summaries 1997. [on-line]. Available: pdf . [last accessed: 2011, February 7]. United States Geological Survey. (2009). Mineral commodity summaries 2009. Washington, DC: U.S. Geological Survey. United States Geological Survey. (2008, October). 2007 Minerals Yearbook Platinum-Group Metals [advance release]. [on-line]. Available: 2007-plati.pdf . [last accessed: 2011, January 18]. Urry, J. (2003). Global complexity. Cambridge, UK: Polity. Vale. (2010a, April 29). Annual Report: Overcoming Challenges, Form 20-F Annual report pursuant to section 13 or 15(d) of the Securities Exchange Act of 1934. Rio de Janeiro: Vale. Vale. (2010b). Vale lists on Hong Kong Stock Exchange. [on-line]. Available: na-bolsa-de-hong-kong.aspx . [last accessed: 2010, December 16]. van der Merwe, I.J. (2004). The Global Cities of Sub-Saharan Africa: Fact or Fiction? Urban Forum, 15(1), 36-47. Waldmeir, P. (1997). Anatomy of a miracle : the end of apartheid and the birth of the new South Africa. New York : W.W. Norton & Co. Wallerstein, I.M. (2004). World-systems analysis: an introduction. Durham: Duke University Press. Wallerstein, I. (2000). Globalization or the Age of Transition?: A Long-Term View of the Trajectory of the World-System. International Sociology 2000, 15(2), 249-265. Walters, S.S. & Prinsloo, J.W. (2002, September). The impact of offshore listings on the South African economy. Quarterly Bulletin of the South AfricanSouth African Reserve Bank, 225, pp. 62-63. Warren, B. (1980). Imperialism: pioneer of capitalism. London: New Left Books. World Bank Group. (2010). Country and Lending Groups. [on-line]. Available: classifications/country-and-lending-groups . [last accessed: 2010, December 16]. 188 Yeoh, B.S.A. (1999). Global/globalizing cities. Progress in Human Geography, 23(4), 607-616. Yeung, H. W.-c. (2002). The Limits to Globalization Theory: A Geographic Perspective on Global Economic Change. Economic Geography, 78(3), 285-305. Yeung, H. (1995). Qualitative Personal Interviews in International Business Research: Some Lessons from a Study of Hong Kong Transnational Corporations. International Business Review, 4(3), 313-339. Zajac, E.J. & Westphal, J.D. (1996). Director Reputation, CEO-Board Power, and the Dynamics of Board Interlocks. Administrative Science Quarterly, 41(3), 507-529. 189 Appendix 1 Estimation of shares traded at specific exchanges Market capitalisation is defined as follows: The  total  value  of  all  outstanding  shares  of  a  publicly-traded  company.  The market capitalization is calculated by multiplying the shares outstanding by the price per share. Market capitalization is one of the basic measures of a publicly- traded  company;  it  is  a  way  of  determining  the  rough  value  of  a  company. Generally  speaking,  a  higher  market  capitalization  indicates  a  more  valuable company. Many exchanges and indices are weighted for market capitalization. It is informally known as market cap. Source: Farlex, Inc., 2010. The percentages of shares in a company purchased through a specific stock exchange for cross listed companies is difficult to determine as these data are changing constantly and are not published. The allocation of control to a specific exchange in Table 3.2 is therefore based on estimates and the best available data from the companies or public sources. If a company has a sponsored ADR programme, the number or proportion of shares traded through the American exchange is usually published and this proportion is attributed to the American exchange. For other cross listings such data is not usually published. In such cases, the respective allocations of shares purchased through a specific exchange is estimated using data on the registered addresses of shareholders, if these data are reported. If a breakdown by country for registered addresses of shareholders is not provided in the reports, the locations of the major shareholders are considered to arrive at an estimate. For example if 60 % of major shareholders are located in the country of the primary listing and 40 % in the country of the secondary listing, it is assumed that overall 60 % of shares are traded at the exchange with the primary listing and 40 % at the other one. 190 Data generation by company Vale: The company is held with 28.7 % through ADRs, which are traded at the NYSE as well as the Euronext Paris. Division of trading volume between the two exchanges was estimated. The remaining shares are traded at the BM&FBOVESPA as well as the LATIBEX. Respective trading volumes for the two exchanges were estimated. BHP Billiton: The following information is quoted from BHP Billiton's Annual Report 2010 entitled "Our strategy delivers: Annual Report 2010": BHP Billiton is a Dual Listed Company comprising BHP Billiton Limited and BHP Billiton Plc. The two entities continue to exist as separate companies but operate as a combined Group known as BHP Billiton. (inside back cover). BHP Billiton Limited has a primary listing on the Australian Securities Exchange (ASX) in Australia and BHP Billiton Plc has a premium listing on the UK Listing Authority’s  Official  List  and its ordinary shares are admitted to trading on the London Stock Exchange (LSE). BHP Billiton Plc also has a secondary listing on the Johannesburg Stock Exchange (JSE). In addition, BHP Billiton Limited and BHP Billiton Plc are listed on the New York Stock  Exchange  (NYSE).  Trading  on  the  NYSE  is  via  American  Depositary Shares (ADSs), each representing two ordinary shares evidenced by American Depositary  Receipts  (ADRs).  Citibank  N.A.  is  the  Depositary  for  both  ADR programs. BHP Billiton Limited’s ADSs have been listed for trading on the NYSE (ticker BHP) since 28 May 1987 and BHP Billiton Plc’s since 25 June 2003 (ticker BBL). (p. 285) Ownership information for the group is divided into BHP Billiton Limited and BHP Billiton Plc. The total global share of on-fuel mineral production of 4.8 percent was therefore attributed proportionately to their market capitalisation to the companies. The report states that 0.2 % of shares are held as ADRs (American Depository Rights) for BHP Billiton Limited and 0.7 % for BHP Billiton Plc. ADRs are traded at the NYSE and control was proportionately attributed to New York in the resulting list of cities. In the case of BHP Billiton Limited, which has only a single listing at the ASX, the remaining share of 99.8 % of stocks was attributed to be traded in Sydney. This assumption was also based on the information that 97.81 % and 1.18 % of shareholders 191 have their addresses in Australia and Nw Zealand respectively. In the case of BHP Billiton Plc, which has a dual listing at the LSE and the JSE, control was attributed proportionately to London and Johannesburg, based on the distribution of registered addresses of shareholders. 79.7 % of shareholders are have registered addresses in the UK and 19.61 % in South Africa, with the remaining 0.69 % in other countries. These remaining 0.69 % were assumed to be traded roughly 80 % at the LSE and 20 % at the JSE, i.e. 0.552 and 0.138 % of total trade respectively. Anglo American: Anglo American's Annual Report does not suggest a break down of shares traded at the stock exchanges at which the company is listed. The 'Anglo American AGM 2010: Notice of Annual General Meeting and Shareholder Information' provides the following information: The Company’s ordinary shares are listed on the London Stock Exchange (the primary listing), the JSE Limited, the SWX Swiss Exchange, the Botswana Stock Exchange and the Namibian Stock Exchange. Certain  US  banks  have  issued  American  Depository  Receipts  (“ADRs”)  in respect of Anglo American’s ordinary shares. These unsponsored ADRs trade on the OTCBB Pink Sheets under the ticker AAUKY. Persons in the US beneficially hold shares representing approximately 15% of the Company’s issued share capital. (Anglo American, 2010). This means that 15 % of shares traded is attributed to New York through the ADRs. Of these no single shareholder appears to have a significant shareholding of more than 3 %. Of the major shareholdings more than 45 % can clearly be attributed to Johannesburg or Pretoria, which are presumably traded at the JSE and in the case PLC nominees (33.54) certainly traded at the JSE. Approximately 11 % of major shareholdings can be attributed to London and presumably these shares are traded at the LSE. This means that a maximum of 40 % of shares can be traded at Anglo American's primary listing the LSE. As most South African shareholders are probably represented by PLC nominees, 50 % of trade was attributed to the JSE, i.e. assuming that about 5 percent are not significant shareholders and represent themselves at the AGM. 192 There is no suggestions about volumes traded at the SWX Swiss Exchange, the Botswana Stock Exchange or the Namibian Stock Exchange, but volumes are presumably very small. To reflect the ability to buy or sell Anglo American shares at these exchanges 0.1 % of trade were attributed to each. Rio Tinto: Rio Tinto is a dual listed company, consisting of Rio Tinto plc with headquarters in London and Rio Tinto limited with headquarters in Melbourne. Rio Tinto Plc has a primary listing at the LSE with a secondary listing at the Euronext Paris. It also has a registered ADR programme. The annual report states that 5.94 % are traded through the ADR programme at the NYSE. There is no information about the breakdown of trade of the remaining 94.06 % of shares, but given the major shareholder locations, only 10 % have been allocated to Paris. All shares of Rio Tinto limited are traded at the ASX in Sydney. FreeportMcMoran: Freeport McMoran is 100 % traded at the NYSE and all indications are that it is an almost exclusively or at least largely American owned company. (see also Appendix 2 for more details). Codelco: Codelco is an unlisted company owned 100 % by the Chilean State (Codelco, 2010b) Xstrata: Xstrata is traded at the LSE and the Swiss Stock Exchange. There is no indication in the report as to the proportion of shares traded at either exchange. Given the location of major shareholders and considering that Xstrata is registered in Switzerland, it was assumed that 45 % of shares were traded in Zurich and 55 % in London. Norilsk Nickel: 193 All Norilsk Nickel shares are traded in Moscow. However, 30.5 % are held as ADRs, which are not registered with an American exchange, but traded as OTC (over- the-counter) in the United States (OTC Market, New York) as well as the electronic International Order Book section of the LSE and the Freiverkehr section of the Berlin stock exchange. The attribution of the 30.5 % of shares traded as ADRs to a place is a very rough estimate and attributed 20.5 % to New York, 7.5 % to London and 2.5 % to Berlin. Barrick: Barrick's annual report shows almost identical trading volumes for the NYSE and the TSX (1153 and 1154 million shares respectively). The allocation of traded shares was therefore estimated at 50 % for each of the two exchanges. Grupo Mexico: Grupo Mexico is traded at the Bolsa Mexicana de Valores (BMV) (Mexican Stock Exchange) only. It consists of three divisions, of which the mining division represents approximately 85 %. The mining division is divided into two companies, Southern Copper Corporation (“SCC”) in Mexico and Peru, and Asarco LLC (“Asarco”) in the United States. Grupo Mexico holds 80 % of SCC, which is listed at the New York and Lima stock exchanges. Asarco is a fully owned subsidary of Grupo Mexico. This means a small proportion of the mining share attributed to Grupo Mexico is in fact held by others (the remaining 20 % of SCC). However, this share is very small. 194 Appendix 2 This appendix includes notes on the data generation for Table 3.3. Locations for all known owners of the ten largest non-fuel mineral producers by production value were assembled in this table. If the owner is a nominee and the nominee's administrative location cannot be reasonably safely attributed, the bank's/institution's head office location was usually used as place of control. Data generation by company Vale: Vale's major shareholder is Valepar, holding 32.8 % of all shares (preferred and ordinary). Valepar is owned 49 % by Litel Participações S.A. Locational control has been allocated proportionately to the owners of Valepar and Litel Participações S.A. BHP Billiton: PLC Nominees (Proprietary) Limited is a South African organisation representing shareholders that hold equity in a company through the STRATE system (an electronic trading platform used by the JSE). Locational control is attributed to Johannesburg, whereas this organisation represents a number of shareholders, most or all of who are likely located in South Africa, but not necessarily in Johannesburg. For shareholders suspected of being in London, the following site was often used for verification: query=Vidacos+Nominees+Limited+%8BCLRLUX2&live=&x=0&y=0 [last accessed: 2010 November 16.] Anglo American: Anglo American does not publish shareholder information, including significant shareholdings, in its annual report. This data is published and extracted from the 'Anglo 195 American AGM 2010: Notice of Annual General Meeting and Shareholder Information', which was publicly available on its web site at the time of writing. See note above (BHP Billiton) re. PLC Nominees. Two significant shareholders in Anglo Amrican, Epoch Two Investment Holdings Limited and Tarl Investment Holdings Limited, "are two of the independent companies which have purchased shares as part of Anglo’s share buy back programme. Epoch 2 and Tarl have waived their right to vote all the shares they hold or will hold in Anglo American plc." (Anglo American, 2010). This means that shares held by Epoch Two Investment Holdings Limited and Tarl Investment Holdings Limited are equivalent to shares held in treasury (own shares). Company announcements, declare that these shares need to be included when calculating total voting rights.26 A Johannesburg Stock Exchange Notifications of major interests in shares in Anglo American from 31 May 2007 (17:29 GMT) lists a Johannesburg address for Tarl Investment Holdings and a Johannesburg Stock Exchange notification on 'Anglo American Plc - Interim Results for the six months ended' from 4 August 2006 (15:21 GMT) states that "Epoch is a South African registered entity owned by a charitable trust whose trustees are independent of the Group. Although the Group has no voting rights in Epoch and cannot appoint or remove trustees, it does meet the accounting definition of a subsidiary in accordance with IAS 27 Consolidated and Separate Financial Statements and as a result is consolidated." The place of control for both entities was attributed to Johannesburg. All other significant shareholders could be attributed to a specific address. Rio Tinto: The Rio Tinto Group consists of two companies, Rio Tinto Plc with a primary listing at the LSE and Rio Tinto limited with a primary listing at the ASX. The two entities act as one group. 26 This may be a mechanism to artificially keep some shareholdings under 3 %, so that a significant shareholding does not have to be announced. Epoch Two Investment Holdings Limited and Tarl Investment Holdings Limited hold 6.79 % of the company, but have waived their voting rights, which makes it legally different from holding shares in treasury, in which case these 6.79 % would not be included for the calculation of total voting rights. 196 Tinto Holdings Australia Pty Limited, a 100 % subsidary of Rio Tinto Plc, holds 28.19 % of Rio Tinto limited. This means that control of 28.19 % of Rio Tinto limited (listed in Australia) lies directly with the sister company in London. For the purpose of allocating control, the shares held by Rio Tinto Plc in Rio Tinto limited were treated like shares held in treasury and therefore not included in he calculation of market capitalisation. This ensures accurate attribution of control to place, i.e. a higher proportion to the shareholders of Rito Tinto Plc (LSE listed) than Rio Tinto limited (ASX listed). FreeportMcMoran: Freeport McMoran indicates only one beneficial owner holding more than 5 % of the company shares. This owner is BlackRock, Inc. in New York, listed with address on page 24 of the 'Notice of Annual Meeting of Stockholders'. The company is 100 % traded at the NYSE and indications are that it is an almost exclusively or at least largely owned American company. The board of directors, advisory directors and the executive management are exclusively American with many members showing strong linkages to the US south including Texas and particularly strong linkages to Louisiana, where the company had its headquarters prior to the acquisition of Phelps Dodge. Reuters reported that by 30 September 2010 Ken Heebner's Capital Growth Management (Boston) had acquired 3.95 million shares (approximately 0.9 % of the company). Bloomberg reported on 13 August 2010 that Capital Growth Management had sold its shares in Freeport McMoran. Reuters also reported on 16 May 2010 that the Singaporean state investment fund Temasek acquired 382,000 shares in Freeport McMoran (an approximately 0.09 % stake). There is very little other indication of ownership of the company available, but the indications point to largely American ownership. An SEC filing of 26 February 2010 shows a Boston law firm to hold about 3.31 %. Codelco: Codelco is an unlisted company owned 100 % by the Chilean State. Control of the company has been attributed to the Chilean capital Santiago. 197 Xstrata: Locations of major shareholders have been researched on the internet and were attributed accordingly. Norilsk Nickel: Through the companies UC Rusal Investment Management and Rypotus Limited 25.13 % of Norilsk are held by Rusal, which has several major shareholders. Control for this segment was attributed as follows: 18.22 % Moscow, 4.47 % Shelekhov (Irkutsk Region, Russia), 2.44 % Baar (see Rusal, 2010). Mr. Potanin V.O., beneficiary of Bonico Holding Co. limited, Montebella Holdings limited, ICFI (Cyprus) limited and Interros International Investments limited, is assumed to be in Moscow. Barrick: Barrick does not report any substantial shareholding of more than 5 % in it 2009 Annual Report or identifies any other shareholders. However, filings with the SEC in early 2010 indicate ownership by two American investment funds of 6.6 and 6.63 % each. This information is reflected in the allocation of ownership to place. Grupo Mexico: Grupo Mexico does not disclose any major shareholdings in its publications. News stories refer to the Larrea family as controlling the company (Rosenberg, 2010). Control of the company was therefore fully allocated to Mexico City. 198


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items