GaWC Research Bulletin 1

There is a nested pair of concerns that has provided the stimulus for this paper. First, I doubt whether social scientists as a body of practitioners have fully appreciated the implications of globalization for their research practices.¹ Despite the plethora of recent writings on globalization themes, it is by no means clear that a concomitant proportion of social science research projects address global-level processes. I suspect that this is an important topic where theory has progressed much further than the evidence actually warrants. Second, within this general concern I identify an example of such an 'empirically-challenged' theory. This particular concern is for the 'world city hierarchy', the Achilles heel of research on world and/or global cities.² Despite voluminous researches over the last two decades which have advanced our knowledge of many important new city growth processes and outcomes, the idea that the cities are arranged into a hierarchy has not been credibly advanced. This is by no means an oversight, a world city hierarchy is a common assumption underlying this field of research, but it seems to have drawn the short straw when it comes to rigorous research. In those researches where it is explicitly addressed, it is the contention of this paper that the work is generally flawed either in data, theory or both. For the much world cities research, however, the idea of a hierarchy is left floating in the background, something vaguely obvious, presumably on the grounds that cities are inherently hierarchical in their relations. Historically there is much to be said for this presumption but in a situation where claims are made for categorically new urban processes, it seems to me that the idea of a hierarchy needs fresh empirical grounding and theoretical interpretation.

The paper is divided into two main sections treating problems of evidence and theory in research on world cities respectively. I argue that empirical work is impeded by a double deficiency of published data being both state-centric and dominated by attribute measurements. Quite simply, currently available data do not allow for the specification of a world city hierarchy. We describe a pilot study that identifies an easily available source of information that provides data on relations between cities to explore the idea of hierarchy.

Theory suffers from another double deficit of neglecting both the temporal and spatial contexts in understanding how a world city hierarchy might arise. This is why the title of this paper specifies 'hierarchical tendencies' rather than the static, universal condition implied by an unqualified reference to hierarchy. We propose an expansion of the pilot into a global research programme. The argument is that if there are real tendencies towards a world city hierarchy this can only be adequately assessed at a systems level which means a global study over a series of years. Hence the utility of the easily available data in the pilot study for this large scale research programme. In a short conclusion a plea is made for readers to join in the proposed global study.

THE PROBLEM OF EVIDENCE

The thesis advanced here is that the reason for the common poverty of research on a world city hierarchy is to a large degree a data deficiency problem. Anybody starting research on this topic finds themselves at the intersection of two inherent inadequacies in the way data is produced in the modern world. There is first, a critical bias towards production of data by and for states and second, the resulting data typically measure attributes rather than relations. These problems may impinge on research in all areas of social science but here they combine to obstruct our understanding at the very heart of what a world city hierarchy must mean.

Mosaic and Network

Consider the situation when a researcher wishes to find out about any global level process of change. Where does he or she look for the necessary world wide information? The most obvious source is the volumes of international data published each year by the United Nations and its various sections. 'International' in this context means that the information is ordered by countries - it is more correctly termed 'inter-state'. Such data are very valuable for learning about the members of the UN and generally for comparing states over a wide range of areas of interest. But they are geographically a very special form of data. Organised in terms of sovereign territories, using such data is to privilege political boundaries over other ways of arranging the information. In effect it is forcing the researcher to have states as the object of analysis whether or not this is theoretically justified. In a global-focused study it is likely that the emphasis will be upon processes that transcend state boundaries. Hence 'international data' is precisely what is not required and yet it remains the only published form of world wide data available for the systematic study of social change.

The reasons for this particular geographical monopoly on world wide data is quite obvious. The vast majority of publicly available data is produced by states for state purposes. It is no accident that the word statistics derives from the early modern production of numerical data for states. But if we do not want to be locked into a particular image of the world as a mosaic of countries, information remains very limited. This is one of the great ironies of contemporary globalization. We live in an 'informational age', increasingly a network world rather than a world of political mosaics, and yet data continues to be largely 'stat-istics'. There is a great need for trans-state data but there is no public institution which has an interest in producing such information. Masses of information on trans-state flows do exist, of course, that is why our times have been dubbed the informational age, but it is private and commercially valuable to its corporate owners and thus out of bounds for researchers. It is therefore extremely difficult to move from mosaic to network in any global-level research.³

World cities are the geographical entity that appear as the organising nodes of world wide networks. This research area, therefore, is arch-typical of those requiring trans-state data. However, apart from the 'city-states' of Singapore and, up until 1997, Hong Kong, world cities are not objects of report for world wide data. Of course, data on cities is available in state census reports and this has allowed for comparative city research but there is nothing on relations between cities across state boundaries. For instance, we can find a lot of published information on relations between the UK and France - trade, migration, etc.,. - but almost nothing on relations between Europe's two premier world cities, London and Paris. And yet for any study of hierarchy it is relations between cities that are the vital indicators of structure. This is not to say that such data cannot be collected or constructed for inter-city relations, there are important studies that have done precisely that, but relatively large time and labour resources must be devoted for this purpose so that most research has been restricted to case studies.⁴ On existing evidence, unlike inter-state analyses, inter-city analyses at a global-scale would seem to present insurmountable practical problems of data acquisition.

Ranks and Relations

Consider a collection of any objects. We can make two types of measurement on them: (i) their individual characteristics may be recorded and (ii) connections between them may be determined. These produce two very different sorts of data, attribute and relational respectively, which are amenable to very different forms of analysis. Attribute data allow for comparative analysis of objects, for instance ranking objects from highest to lowest by a given characteristic. Relational data allows for network analysis of objects, for instance by creating a flow matrix between objects in terms of given connections. In the case of the former, the collection of objects can be simply an aggregation, a mere collection of independent individuals. For the latter, the objects are assumed to form a system, an organised set of inter-dependent individuals. These differences are crucial in any consideration of hierarchies.

Attribute data can be manipulated in many ways but they can never show the structure of a hierarchy. Ranking produces only an ordered list. It tells us nothing about the relations between the objects on the list. Since the concept of hierarchy is intrinsically relational, it can only be defined through data ordered as a network; a statement that objects in a ranked list constitute a hierarchy is conjecture and nothing more. For any study of hierarchy, therefore, attribute data are of limited utility beyond showing a functional basis for the hierarchy. And yet the literature on world and/or global cities abounds with tables of ranked cities often masquerading as a hierarchy. One example will suffice. Saskia Sassen's seminal Global City is a valuable comparative study of New York, London and Tokyo viewed as the apex of a hierarchy of cities but she uses only attribute data - lists of ranked cities plus tables showing the 'big three' versus the rest - and thus provides no evidence about patterns of relations between cities, that is to say, the actual structure of the hierarchy her three cities posedly head is missing.⁵

We are not, of course, the first to point to this dearth of studies of relations between world cities. Smith and Timberlake have recently explored the relational data needs of world cities research and have constructed a typology of inter-city linkages based upon the form (human, material, information) and function (economic, political, cultural and social) of flows.⁶ Twelve types of flow are thus identified but the authors have to admit that they constitute little more than a 'wish list' for world city research. As well as the reasons outlined above, such data are not available because, in any case, most data collection agencies focus upon attribute data, presumably due to its general ease of collection and because most (state) demands seem to favour this form of information. The only possible conclusion we can draw from all of this is that for our 'wish list', or something like it, to come true, as global researchers we are alone, we must collect information specifically for our own needs.

A Pilot Study of Relations Among Selected US Cities⁷

Saying we need to collect our own data is easier said than done. Within the literature we can identify two major approaches for defining relations between world cities. Both have much potential but we will argue here that they both have severe limitations when confronted by the research task of providing sensitive world wide information. In contrast to the state-centric nature of most world-wide data, information on transport and communications are typically node oriented, for instance recording flows by city airports and city telephone codes.⁸ The problem with analysis along these lines is that the data incorporates a large proportion of flows that are not part of the new city building processes which are our particular concern. In contrast, in recent studies of services by economic geographers, trade directories and company surveys have been used to provide very detailed information on operational links between cities.⁹ The emphasis has been on particular services rather than the whole range of activities that constitute world city processes so that they offer only partial glimpses of possible urban hierarchies. In addition, translating this approach to a world wide collection of data would seem to be beyond most feasible research resources. In essence, therefore, we need something between these two approaches: data as easily accessible as airline information but without providing results that are too general, and data sensitive to real world city processes but without being prohibitive for time and cost reasons.

We ask the question - what is the basic daily information world in which the business decision makers of a world city operate? One answer to this question leads us to a very accessible information source, the business pages of the city's leading newspaper. Below we employ a standard content analysis of this newspaper section focusing on geographical information defining the linkages of the city.¹⁰ This is the business editor's interpretation of the city's everyday information needs in relation to other cities and places. It is relational data from which we should be able to measure the intensity and direction of linkages between cities. The purpose of this pilot study is twofold. First, we evaluate the general utility of the method for delineating inter-city connections: does it work? Second, we evaluate the credibility of the method to produce results that are both consistent with what we already know and can go beyond what we know. The latter requires the study to be in a region where the urban hierarchy is well-known; we have chosen the USA.

As an exploratory project our coverage of US world cities and potential world cities is not comprehensive. We include in our sample those cities in John Friedmann's¹¹ latest world city listing - New York, Los Angeles, San Francisco, and Miami - plus two other cities: Chicago that usually ranks high among US cities and for whom Janet Abu-Lughod¹² claims world city status, and Boston, which Hall¹³ calls 'sub-global', representing the classic US regional city, the 'capital' of New England. For each of these six cities we chose the leading city newspaper for business news: New York Times, Los Angeles Times, San Francisco Chronicle, Miami Herald, Chicago Tribune and Boston Globe. We chose 1990 as our year of study and sampled the front page of the business section. Hence we focused on the stories the editor thought to be particularly important. 4,246 place references were recorded from these front pages. The precise methodology that was followed is described in appendix 1.

Does this methodology produce a delineation of the global scope of our selected cities? The 4246 place references were initially organised into a table showing place counts by world regions for each city (Table 1). Let us remind ourselves what these data represent. They are the places that occur in stories and news items that the business editor of the various newspapers has deemed newsworthy enough to appear on his or her front page. The resulting information worlds of US world cities are far from isotropic and loosely reflect uneven economic development. For instance, Sub-Saharan Africa is revealed as a huge information lacuna: the world's poorest region is a 'nowhere place' for these US world cities.¹⁴ (Ironically, the only reference to Sub-Saharan Africa outside of South Africa was to the address of a Swiss bank account.) Table 1 suggests we can answer yes to the question asked above.

Does this methodology differentiate between the geographical scopes of these cities within the world-economy? Table 2 provides a simple index of global scope by specifying the percentage of all references that are non-US. These figures are very revealing. They confirm that New York is the US city with the greatest international orientation - the only one in our sample with more international than home references.¹⁵ Of the other cities, they all have about a third of their counts as international except for Boston with somewhat less than a quarter references beyond the US. This seems to confirms Boston's role largely as a US regional city. In Table 3 these international references are broken into three geographical orientations: the American, Atlantic and Pacific references of the sample cities. The first point to make is the relative unimportance of the Americas. The major exception is Miami whose world city role at the interface between the USA and Latin America, especially the Caribbean is confirmed in this data.¹⁶ But for all our sample cities it is the Atlantic orientation that is the greatest although in the case of Los Angeles its Atlantic and Pacific counts almost balance. Boston with its weak international orientation is, perhaps as expected, the most Atlantic in orientation but New York comes a close second providing a clear contrast with Los Angeles which leads easily in Pacific orientation. Tables 2 and 3 enable us to answer yes to the question we posed above.

What the above findings show is that this methodology produces results that are consistent with our knowledge of US cities in the world but their credibility will be enhanced if we can further show that they can generate less predictable results. I think this occurs when we turn to using the data to explore city hierarchical relations within the US. In Table 4 we focus upon place references to major US cities by adding Atlanta, Dallas, Detroit, Houston, Philadelphia and Washington, DC. to our original sample. None are usually identified as world cities but they are viewed as at least the equivalent of Boston in the US urban hierarchy. The total column counts for the twelve cities are not comparable because of the self-referencing boost given to the sample cities. Hence an 'external' average count measure is added which is computed for the sample cities by excluding self-referencing. As well as clearly confirming the predominance of New York within the US city system, Table 4 provides two surprises. First, there is the relative sparceness of the table: 19 zeros and 7 ones plus other single digit counts despite a total of over 4,000 place references recorded. This suggests that most major US cities are not really newsworthy places for business editors. Second, although Los Angeles ranks a clear second in our sampled cities she trails Washington, DC in average references. It would seem that business editors are more prone to recognise the importance of Washington than world city theorists!¹⁷ These findings are illustrated in Figure 1 which portrays all inter-city referencing above a threshold of 10% of non self-references. Among the sampled cities references to New York abound with little reciprocity and few other referencing except to Washington. (There can be no reciprocity in this case, of course, since Washington was not in our original sample.)

This diagram illustrates what we might refer to as New York's 'conditional primacy'. In the world cities literature Washington, DC is largely ignored¹⁸ despite the fact that, in other parts of the world, world cities are often capital cities, although this function usually is ignored in their designation as world city.¹⁹ Our evidence suggests world city theorists have been too hasty in their separation of the economic and the political in identifying world cities. In terms of information salience, what happens in Washington, as well as often dominating other parts of city newspapers, is second only to New York for business editors. There are two interpretations we can draw from this. Either we can begin to explore the notion of Washington as a world city in its own right or we can think of Washington as the political annex to New York as the USA premier world city. It follows from the latter that unlike London and Tokyo, or indeed Paris and Madrid, New York on its own is only a partial world city since it does not incorporate political control functions that are crucial in what is an international political economy world. Perhaps it is time to see New York-Washington as America's premier world city. But whatever our interpretation of this result, it confirms the credibility of our methodology to pose ideas beyond the existing state of our knowledge.

I will proceed on the assumption that this section has indeed illustrated that the proposed methodology for measuring inter-city relations produces credible results and, because it is based upon easily accessible sources, has the potential for large-scale use beyond the pilot study.

THE PROBLEM OF THEORY

Any measurement exercise is ultimately only as good as the theoretical context in which it is used. In the world cities literature the nature of the theory suggesting a hierarchy is not well developed. Cohen's pioneer work relied largely on city ranking by corporate headquarters and this corporate emphasis was followed by Friedmann in his initial formulation of the world city hypothesis.²⁰ Friedmann's 'second thesis' refers to a 'complex spatial hierarchy' but, he adds, lack of data leads only to a 'possible' ordering based upon the 'presumed nature of their (i.e. cities') integration with the world economy'.²¹ In the event he identifies eleven primary and nineteen secondary world cities. Later he refers to 'a hierarchy of spatial articulations' in which the complexity is represented by the types of spaces being integrated by different classes of cities from global to regional.²² In a similar manner Hall refers to 'global' and 'sub-global' cities with the latter as a 'second rung' of national and regional cities.²³ This is all very reminiscent of the 'range of a good' argument for services in traditional central place theory but with the caveat of additional complexity. Of course, it is the complexity that is interesting, that needs investigation.

One reason the hierarchy is complex is because of the volatility in ordering inherent in a situation of intense city competition. Freidmann makes this argument in order to suggest that defining a hierarchy among world cities is 'a futile undertaking' and therefore he urges us to stay with a 'rough notion' of hierarchy with no further specification.²⁴ But there is an alternative response to his presumed volatility: the hierarchy should not be measured as a static structure but as a dynamic ordering of cities. Perhaps Freidmann was put off this possibility by the immense data requirements that are implied. But for this paper, the data problem is at least partially solved by our argument of the last section. Hence we will proceed without having to make do with a rough notion of hierarchy and assume that we will be able to measure a changing global hierarchy.

The fundamental reason for the complexity of the relations amongst world cities is to be found in the historical and geographical variation of the world-system in which they operate. In this section I begin the task of tackling this complexity by investigating the timing and spacing of world cities. Much of the material I draw upon is implicit in the world cities literature but it has not been used to inform the question of hierarchy.

Moyenne Durée and Longue Durée

We are to study the world city hierarchy as dynamic, but which dynamic? Fernand Braudel has taught us that to properly understand our changing world we must be cognisant of three 'social time' frames: events of the short term (court durée), medium term trends (moyenne durée) and long term structures (longue durée).²⁵ The first social time is represented by Friedmann's demoting of Johannesburg because of political events in South Africa²⁶ but the world cities literature is overwhelmingly moyenne durée in its temporal orientation. From Mayer's and Friedmann's original linking of world cities to the 'new international division of labour' to Hall's recent discussion of globalization as the context for his global cities, the literature focuses upon trends of the past two or three decades.²⁷ Although there are references to the modern world-system as the context for the rise of these cities, the longue durée is conspicuous by its absence from the discussion.²⁸ But today's complexity in the relations between world cities is to a large degree the result of the unfolding of world-system before 1970, that is to say of patterns and processes that long pre-date a medium term analysis.

From a longue durée perspective²⁹, the world cities that are our concern here are interpreted as the particular form that dominant urban nodes in the system have taken in just the last few decades of the half-millennial existence of the modern world-system. What form did their predecessors take? Braudel has identified a sequence of Italian 'world-cities' as financial centres of an incipient European world-economy ending with Genoa in the sixteenth century to be continued by the leading cities of the hegemonic states that subsequently rose to sustain the modern world-economy.³⁰ These 'hegemonic cities' were Amsterdam in the seventeenth century which finally gave way to London in the Revolutionary and Napoleonic Wars with the latter being replaced by New York in World War I.³¹ This focus on finance continues through to contemporary world cities but it is not the only way in which urban nodes have come to pre-eminence in the modern world-system. Coincident with the growth of the capitalist world-economy was the development of the inter-state system and the resulting political centralisation of the new nation-states created great capital cities. Hence, in the decline phase of Britain's hegemonic cycle beginning in the last third of the nineteenth century, London came to be challenged as the world's leading city, not by financial rivals, but by other capital cities. In this 'nationalisation' phase of the world-system, the importance of these cities depended on the great power status of their state. Thus, after 1870 London's great rival as a world city was Berlin, the capital of the new German Empire and other capital cities were also ranked according their state's power: Paris, Vienna, St Petersburg, Rome. These were, of course, all imperial capitals with British hegemony having been replaced by the 'age of imperialism', which in current terminology might be referred to as 'political globalization' - all inhabited parts of the world were incorporated into the inter-state system.

Today we are living through the decline phase of US hegemony. The great economic growth of the world-economy during American 'high hegemony' (c. 1945-71) has culminated in contemporary economic globalization so that, unlike London a century ago, New York is being challenged by other world cities as economic nodes in the modern world-system. The growth of the overall world-economy has resulted in an expansion of its core which exists now as three sections, northern America, western Europe and the western Pacific rim. It is the leading financial centres within each of these sections, New York, London and Tokyo, that constitute Sassen's three 'global cities' which are assumed to be at the apex of a world hierarchy of cities.³² It is this world hierarchy hypothesis, seen as the geographical dimension of economic globalization, which is unique to contemporary world cities. Braudel's world-cities assumed a single dominant node, hegemonic cities likewise, and imperial world cities each had their own separate imperial urban hierarchies, but today economic globalization has created the conditions for the development of a single world-wide urban hierarchy.

Creating conditions for something is not the same as creating its actual existence. The potential exists for the development of a world city hierarchy but there is no reason to assume that such could be created in just a couple of decades, eradicating the legacy of previous arrangements of major cities. A fully developed world city hierarchy implies an integrated global economic space that has overcome the obstacles represented by the political spaces of states. Such a change, the demise of the states, is epochal, a phenomonen of longue durée proportions. Hence, in this sense, it is futile, as Friedmann suggests, to search for a world city hierarchy at this time.³³ Rather we should focus upon searching for tendencies towards a world city hierarchy and measure whether they are strengthening over time. This is the critical conclusion we come to by arguing from outside the usual moyenne durée stance of world city researchers.

Politics and Economics

One of the features of the modern world-system is the way in which the relationship between economics and politics has been negotiated in different ways at different times. That is to say, the relative powers of political and economic elites have fluctuated. And the fortunes of the major cities of the system have been integrally tied up to these fluctuations as the previous discussion illustrated. The world cities literature has emphasised the current tendency for power to accrue to economic elites but in doing so they have neglected the on-going power of the states. Take for example the fact that there are two leading world cities, London and Paris, less than an hour's travel apart. This is not what an urban hierarchy derived from economic globalization would look like. These two cities are global cities not just because of past imperial legacies but also because of contemporary political processes where their home states encourage their continuing pre-eminent status in the world. Alongside globalizing tendencies that promote a world city hierarchy there are continuing 'nationalising' tendencies that are countering the development of a world city hierarchy. In other words, there is currently severe tension between continuing nation-state reproduction and economic globalization with both impinging upon each other: as well as globalization constraining states, states continue to be obstacles on the road to a uniform world economic space.³⁴ The development of a world city hierarchy out of economic globalization is, therefore, neither a simple evolution nor, indeed, even an inevitable outcome of contemporary economic and political changes in the modern world-system.

The tripartite division of the contemporary core of the world-economy is actually very interesting with respect to economic and political processes pulling in different directions. These three economic zones contrast with each other in terms of their political structures. They can be ordered in terms of political integration from the most unified, northern America, through partial unification, western Europe, to the least unified, western Pacific rim.³⁵ This will have profound implications for the development of world city hierarchic tendencies. With Tokyo, for example, any hierarchic relations would seem to be very problematic in a section encompassing a veritable pot pourri of rivals - other capital cities (e.g. Seoul), city states (e.g. Singapore) and old trading cities (e.g. Shanghai) - resulting from both a highly variable imperial experience and rapid contemporary economic changes. London , on the other hand, exists in a section of nation-states trying to come together as the European Union. Nevertheless 'historical inertia', as Hall³⁶ calls it, adds a deal of stability to contemporary dynamism to create a very complex pattern of cities including rival capitals (e.g. Paris), traditional finance centres (e.g. Zurich) and growing regional political (Brussels) and economic (Frankfurt) centres. In contrast, New York exists within a core section almost wholly in one state and in which there are no historical or political rivals. In summary, we might say that the core of the world-economy has evolved like a social laboratory designed to test the effects of variable political conditions on the development of the world city hierarchy.

In conclusion, we have found that the contemporary pattern of world cities is eminently suited for investigating the relationship between economic developments and political obstacles in the putative development of a world city hierarchy. We take advantage of this 'laboratory' situation in devising a programme for investigating the variable nature of tendencies towards a world city hierarchy.

Research Programme for a Global Study

Peter Hall begins his recent plea for 'a general urban theory' with reference to 'the global system of cities'.³⁷ A serious, comprehensive study of this system is surely long overdue. It is our contention here that a combination of easily available and credible relational data within the theoretical context of geographically variable hierarchical tendencies provides what has hitherto been missing: the empirical and conceptual tools to carry out the task.

I propose the following research programme:

1. Objects of study: all primary/secondary (or global/sub-global) cities which have been identified as part of a world city hierarchy. Friedmann actually refers to lower rungs of the hierarchy but at this stage we will concentrate on the top two rungs. My provisional list is as follows:

northern American core - New York, Washington DC, Los Angeles, San Francisco, Chicago, Miami, Boston, Seattle, Houston, Toronto, Montreal, Vancouver;

western European core - London, Paris, Frankfurt, Milan, Brussels, Amsterdam, Madrid, Dusseldorf, Munich, Lyon, Barcelona, Vienna, Rotterdam, Copenhagen, Rome, Zurich;

western Pacific core - Tokyo, Singapore, Osaka, Seoul, Hong Kong, Taipei, Kuala Lumpur, Sydney, Melbourne;

non-core - Mexico City, Caracas, Sao Paulo, Buenos Aires, Rio de Janeiro, Johannesburg, Bombay, Bangkok, Manila;

ex-command economies - Moscow, Prague, Berlin,³⁸ Beijing, Shanghai.

2. Timing of study: we need to go back to before the presumed rise of contemporary world cities in order to provide a base from which to measure a globalization of the urban system. I suggest using five year intervals from 1955 to 1995. The early years represent US high hegemony to be followed by the whole period identified in the literature as the moyenne durée of globalization. For each city for each year data such as that collected above in the pilot study would be constructed. For the cities of the ex-command economies, of course, only recent years will be surveyed.

3. Analysis: for each year we can repeat the sort of analysis we carried out for the pilot 1990 study but the overall emphasis will be on change and difference. That is to say we monitor relations between cities over time with particular reference to differences between sections of the core and their respective peripheries. Of course, for the east Asian cities the actual rise to core status will be recorded. We will not be defining a world city hierarchy but will be looking for indications of tendencies both promoting and obstructing its possible creation.

4. Organisation: this requires a substantial planning operation involving data collection and subsequent interpretation for cities in 29 countries. A steering group consisting of persons from each of the geographical zones identified above should both oversee the work by putting 'flesh' on the bones of this proposal, and ensure consistent data collection and analysis from a team of researchers located in all countries involved.

5. Output: the data will be lodged in the Global Observatory, a Loughborough University initiative on the world wide web which specialises in trans-state data.³⁹ This will make it publicly available for all urban researchers. Before the latter happens we will convene a conference of all participants in which the essence of the project results will be reviewed and book planned. In addition a division of labour for other analysis and writings to come out of the project will be agreed.

For my purposes here there is no need to develop this proposal any further. If we are to go beyond vague discussion of a 'world city hierarchy', or indeed globalization itself, there is an urgent need for studies along the lines presented above. My objective now is to stimulate thinking on such global study and try and attract people to join me in this endeavour.

CONCLUSION AS INVITATION

We return to the point we began with: I do not think that social scientists have fully appreciated the implications of globalization for their research practices. Some sizeable segment of research effort is required to look at the world-system as a whole. The sizes of research tasks implied by this are daunting but they cannot be avoided if we are to come to grips with contemporary social change. The Global Observatory at Loughborough University has been set up to help facilitate such work. This initiative has now completed its establishment phase with a general plea to social science researchers to become involved in its trans-state enterprise. This heart of this paper is a special plea to readers to join in this particular research project to investigate the global system of cities.

Clearly in the arguments above I think I have proposed a methodology that meets our data needs and have also provided the necessary theoretical context for the study. Of course, this is not the only way of proceeding with investigation of the topic and debate on the matter is welcomed. But I do think there is great urgency to begin a theoretically-informed global research study so that we have 'hard' empirical evidence for what have become commonplace assumptions within the globalization thesis. It is in this spirit that this paper has been written.

ACKNOWLEDGEMENTS

The work on the pilot study was done jointly with Travis R Longcore and Carmen McWilliams and I am indebted to them for both their careful and diligent data collection and their contributions to planning the US cities project. The current paper was read and commented upon by my Loughborough colleague, Jon Beaverstock, who was thus instrumental in clarifying some key points for me.

REFERENCES

Abu-Lughod J L (1995) 'Comparing Chicago, New York and Los Angeles: testing some world cities hypotheses' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP)

Braudel F (1980) On History (London: Weidenfeld and Nicolson)

Braudel F (1984) The Perspective of the World (London: Collins)

Castells M (1996) Network Society (Oxford: Blackwell)

Cheshire P and Gordon I (1995) 'European integration: the logic of territorial competition and Europe's urban system' in Brotchie J et al (eds) Cities in Competition (Melbourne: Longman)

Cohen R J (1981) 'The new international division of labour, multinational corporations and urban hierarchy' in Dear M and Scott A J (Eds) Urbanisation and urban planning in capitalist society (London: Methuen)

Cook I G, Doel M A and Li R (1996) Fragmented Asia: Regional Integration and National Disintegration in Pacific Asia (Aldershot: Avebury)

Daniels P W, Dinteren Van J H J and Monnoyor M C (1992) 'Consultancy services and the urban hierarchy in Western Europe' Environment and Planning A, 24, 1731-1748

Friedmann J (1995) 'Where we stand: a decade of world city research' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP)

Grosfoguel R (1995) 'Global logics in the Caribbean city system: the case of Miami' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP)

Hall P (1966) The World Cities (London: Weidenfeld and Nicolson)

Hall P (1995) 'Towards a general urban theory' in Brotchie J et al (eds) Cities in Competition (Melbourne: Longman)

Hall P and Preston P (1988) The Carrier Wave (London: Unwin and Hyman)

Hamnett C (1995) 'Controlling space; global cities' in Allen J and Hamnett C (eds) A Shrinking World? (Oxford: Oxford University Press)

Keeling D J (1995) 'Transportation and the world city paradigm' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP) 115-131

Korff R (1987) 'The world city hypothesis: a critique' Development and Change, 18, 483-95

Lee, R and Pelizzon, S (1991) 'Hegemonic cities in the modern world-system' in Kasaba, R (ed) Cities in the World-System (New York: Greenwood)

Leyshon A (1995) 'Annihilating space?: the speed-up of communications' in Allen J and Hamnett C (eds) A Shrinking World? (Oxford: Oxford University Press)

Lynch J and Meyer D R (1992) 'Dynamics of the US system of cities, 1950 to 1980' Urban Affairs Quarterly 28, 38-68

Meyer D (1986) 'World system of cities: relations between international financial metropolises and South American cities' Social Forces, 64, 3, 553-81

Nijman J (1996) 'Breaking the rules. Miami in the urban hierarchy' Urban Geography 17, 1, 5-22

ó hUallacháin B (1994) 'Foreign banking in the American urban system of financial organization' Economic Geography 70, 3, 206-228

Pred A (1980) Urban Growth and City Systems in the United States, 1840-1860 (London: Hutchinson)

Sassen S (1991) The Global Cities (Princeton: Princeton University Press) CAPTION

Smith D A and Timberlake M (1995) '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 (Cambridge: CUP) 79-97

Taylor P J (1995) 'World cities and territorial states: the rise and fall of their mutuality' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP)

Taylor P J (1996) 'Embedded statism and the social sciences: opening up to new spaces', Environment and Planning A, 28, 1917-28

Thrift N (1987) 'The fixers? the urban geography of international commercial capital' in Henderson J and Castells E (eds) Global Restructuring and Territorial Development (Beverly Hills, CA: Sage) 203-233

NOTES

1. I have elsewhere developed this as a general argument for trans-disciplinary research: Taylor P J (1996) 'Embedded statism and the social sciences: opening up to new spaces', Environment and Planning A, 28,

2. The question of when world cities become global cities is discussed in Hamnett C (1995) 'Controlling space; global cities' in Allen J and Hamnett C (eds) A Shrinking World? (Oxford: Oxford University Press). Here I use the terms interchangeably for contemporary dominant cities on the following grounds: world cities are dominant cities within world-systems and, since the modern world-system is now global in scope, today world cities may be reasonably designated global.

3. It is noteworthy that the data used by Manuel Castells in his new book on the 'network society' is overwhelmingly state-centric in nature: of 43 tables, 42 compare states and one compares city population totals; of 25 maps and graphs, 22 portray states, one shows city populations and two plot world-wide trends.

4. I am thinking here of Nigel Thrift's work on international financial centres which focuses on London, for example: Thrift N (1987) 'The fixers? the urban geography of international commercial capital' in Henderson J and Castells E (eds) Global Restructuring and Territorial Development (Beverly Hills, CA: Sage) 203-233

5. Sassen S (1991) The Global Cities (Princeton: Princeton University Press)

6. Smith D A and Timberlake M (1995) '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 (Cambridge: CUP) 79-97

7. This section is based on work carried out by Travis R Longcore and Carmen McWilliams as graduate students during a course I taught at the University of Delaware. (See acknowledgements.)

8. Keeling D J (1995) 'Transportation and the world city paradigm' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP) 115-131

9. Pioneer uses of such data can be found in the works of David Meyer on how Latin American cities have stronger ties to the USA than between each other, and of Nigel Thrift on 'international financial centres' using foreign banks in London and selected London companies' international networks: Meyer D (1986) 'World system of cities: relations between international financial metropolises and South American cities' Social Forces, 64, 3, 553-81 and Thrift op cit. More recently ó hUallacháin has used foreign banks in US cities to illustrate the global scope of six of them and Lynch and Meyer have used law firm branching patterns to analyse the US urban hierarchy. The largest such organisational study is by Peter Daniels and colleagues on western European cities in which 292 consultancy firms in 30 cities in six countries are surveyed: ó hUallacháin B (1994) 'Foreign banking in the American urban system of financial organization' Economic Geography 70, 3, 206-228; Lynch J and Meyer D R (1992) 'Dynamics of the US system of cities, 1950 to 1980' Urban Affairs Quarterly 28, 38-68; and Daniels P W, Dinteren Van J H J and Monnoyor M C (1992) 'Consultancy services and the urban hierarchy in Western Europe' Environment and Planning A, 24, 1731-1748

10. A similar methodology has been used in Pred A (1980) Urban Growth and City Systems in the United States, 1840-1860 (London: Hutchinson)

11. Friedmann J (1995) 'Where we stand: a decade of world city research' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP)

12. Abu-Lughod J L (1995) 'Comparing Chicago, New York and Los Angeles: testing some world cities hypotheses' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP)

13. Hall P (1995) 'Towards a general urban theory' in Brotchie J et al (eds) Cities in Competition (Melbourne: Longman), p. 22

14. This is a good illustration of the exclusionary tendency in globalization - see Leyshon A (1995) 'Annihilating space?: the speed-up of communications' in Allen J and Hamnett C (eds) A Shrinking World? (Oxford: Oxford University Press), p.42

15. Compare with ó hUallacháin op cit and Lynch and Meyer op cit

16. Grosfoguel R (1995) 'Global logics in the Caribbean city system: the case of Miami' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP); and Nijman J (1996) 'Breaking the rules. Miami in the urban hierarchy' Urban Geography 17, 1, 5-22

17. Washington, DC does not appear in Friedmann's world city hierarchy, for instance, op cit p. p.24

18. The major exception is Lynch and Mayer op cit, p.42 where Washington, DC ranks second among US cities in terms of law firm branching patterns

19. This was not the case in Peter Hall's original identification of world cities but since R B Cohen's linking of what he terms the 'new world cities' to the new international division of labour the political function of world cities has been neglected: Hall P (1966) The World Cities (London: Weidenfeld and Nicolson); and Cohen R J (1981) 'The new international division of labour, multinational corporations and urban hierarchy' in Dear M and Scott A J (Eds) Urbanisation and urban planning in capitalist society (London: Methuen), p.307

20. Cohen op cit; and Friedmann J (1986) 'The world city hypothesis' Development and Change, 17, 69-83

21. Friedmann (1986) op cit, p. 71

22. Friedmann (1995) op cit, p.23

23. Hall (1995) op cit, p. 5

24. Friedmann (1995) op cit, p.23

25. Braudel F (1980) On History (London: Weidenfeld and Nicolson)

26. Friedmann (1986) op cit, p. 39; see also the demise of Berlin: Hall P and Preston P (1988) The Carrier Wave (London: Unwin and Hyman)

27. Meyer op cit; Friedmann (1986) op cit; Hall (1995) op cit

28. Exceptions are Korff R (1987) 'The world city hypothesis: a critique' Development and Change, 18, 483-95; and Taylor P J (1995) 'World cities and territorial states: the rise and fall of their mutuality' in Knox P L and Taylor P J (Eds) World Cities in a World System (Cambridge: CUP)

29. The following argument draws upon Taylor (1995) op cit

30. Braudel, F (1984) The Perspective of the World (London: Collins)

31. Lee, R and Pelizzon, S (1991) 'Hegemonic cities in the modern world-system' in Kasaba, R (ed) Cities in the World-System (New York: Greenwood)

32. Sassen op cit

33. Friedmann (1995) op cit, p.23

34. This is a point emphasized by Hall (1995) op cit, p.21; the whole of the argument of this section can be interpreted as a return to Hall's orginal more rounded view of world city: Hall (1966) op cit

35. A recent book on the 'Pacific rim' is entitled Fragmented Asia which contrasts with institutional development in Europe where tendencies towards political and economic union are directly impinging on the urban system: see Cook I G, Doel M A and Li R (1996) Fragmented Asia: Regional Integration and National Disintegration in Pacific Asia (Aldershot: Avebury) and Cheshire P and Gordon I (1995) 'European integration: the logic of territorial competition and Europe's urban system' in Brotchie J et al (eds) Cities in Competition (Melbourne: Longman)

36. Hall (1995) op cit, p.5

37. Hall (1995) op cit, p.3

38. Berlin, as a divided city, is a special case and will need special treatment but is well worth including both for historical reasons and for its new potential as Germany's capital city.

39. The address is http://www.stile.lut.ac.uk/global.html