Globalization and the World City System: Region, Role, and Position since 1981 by Arthur S. Alderson Department of Sociology Indiana University *This research.

Globalization and the World City System: Region, Role, and Position since 1981

by

Arthur S. AldersonDepartment of Sociology

Indiana University

*This research was supported by a grant from the World Society Foundation. Direct all correspondence to Arthur S. Alderson, Department of Sociology, Indiana University, Ballantine Hall 744, Bloomington, IN 47405. Email: [email protected].

World City Hypotheses

I. Stephen Hymer. 1972. "The Multinational Corporation and the Law of Uneven Development."

II. John Friedmann. 1986. "The World City Hypothesis."III. Saskia Sassen. 1991. The Global City: New York, London, and Tokyo

• All three see globalization fundamentally altering the geography of inequality, but they differ strongly on exactly how it is changing.

• Hymer argues that globalization will reproduce and magnify existing patterns of inequality and dependency.

• Friedmann and Sassen propose that a new geography of centrality and marginality is emerging, one that cuts across established North/South and East/West divides in the global system.

World City Hypotheses

• Hymer, Friedmann, and Sassen share the view the globalization has been a fundamentally implosive, rather than explosive process. While economic activity has been spatially dispersed, its organization is described as growing increasingly centralized, centralized in the “world” or “global cities.”

• All three thus share the view that the fate of cities has become increasingly tied to their position in international flows of investment and trade.

• Friedmann and Sassen differ in suggesting that such flows are increasingly orthogonal to those of an earlier geography of inequality constructed around the nation-state system.

Globalization and the World City System: Region, Role, and Position since 1981

• How extensively has the global urban hierarchy been altered?

• Have power and prestige in the world city system grown more concentrated?

• Does the new world city system cut across pre-existing patterns of global inequality?

Data

• Our research focuses on a key relation linking cities into a world system of cities: that between multinational enterprises and their subsidiaries.

• Headquarter and branch locations of the world’s 500 largest multinationals in 1981 and 2000.

• Relational (network) data. Cities are linked by tie between multinational headquarters and subsidiaries.

• Data coded as directional (i.e., distinguishing between senders and receivers) and valued (i.e., allowing multiple ties between two cities).

• Data set links more than 3700 cities at two time points.

• In 2000, the total revenue of the Global 500 totaled $12.6 trillion, more than twice the combined gross domestic product of the world’s 156 poorest societies.

Figure 1. 2- and 3-step paths from Indianapolis to Durham

Figure 2. 2- 3- and 4-step paths from Indianapolis to the Durham

Point Centrality

• Power of cities is assessed as outdegree, closeness, and betweenness.

• Prestige of cities is assessed as indegree.

Point Centrality

The Star The Circle

A

B

C

D

E

F

GA

B

C

DE

F

G

outdegree = the number of ties sent from city i

closeness = the inverse average distance between city i and all others

betweenness = the probability that the paths linking cities j and k contain city i

indegree = the number of ties that city i receives

Blockmodeling

• Regular equivalence blockmodeling partitions cities into sets composed of cities have that the same relation to members of other equivalence sets (e.g., Susan is daughter of Jane’s sister and Tess is daughter of Lisa’s sister → “niece” and “aunt”).

• We focus in particular on two positions that emerge in the blockmodeling: the primary and isolate positions.

Primary blocks = the “core” of the world city system, highly cliquish sets of cities whose members are involved in high levels of relations with outsiders (i.e., >exp in-group preference + >exp outdegree and indegree)

Isolate blocks = the “periphery” of the world city system, internally disconnected sets of cities whose only relations are with members of more active sets (i.e., <exp in-group preference + <exp outdegree and indegree)

Q1: How extensively has the global urban hierarchy been altered across the era of globalization?

Method: Calculate the Spearman rank-order correlation between each measure

of power and prestige in 1981 and 2000

Results: outdegree 1981, 2000 r = 0.672 closeness 1981, 2000 r = 0.568 betweenness 1981, 2000 r = 0.665 indegree 1981, 2000 r = 0.616

Conclusions: Spearman correlations in this range mean that there has been

substantial change between 1981 and 2000 in the rank order of cities in the global urban hierarchy. Plainly stated, the “deck” has been “reshuffled” in the past two decades, consistent with the idea that globalization is generating a new urban hierarchy.

Q2: Have power and prestige in the world city system grown more concentrated?

Method: Track the change from 1981 to 2000 in the coefficient of variation (COV)

for each measure of power and prestige

Results:COV outdegree: 1981 = 11.100 → 2000 = 13.018COV closeness:1981 = 0.104 → 2000 = 0.095COV betweenness: 1981 = 9.663 → 2000 = 12.672COV indegree: 1981 = 5.917 → 2000 = 5.957

Conclusions: With the exception of closeness, the results are consistent with the expectation that power and prestige will become concentrated in an increasingly small number of “world” or “global” cities as the world economy develops.

Q3: Does the new world city system cut across pre-existing patterns of global inequality?

Method:

Estimate a series of conditional-change (OLS) and logistic regressions to determine the association of world system position (i.e., of a city being located in a core, semiperipheral, or peripheral country) with change in the power, prestige, and blockmodel position of cities 1981-2000.

Q3: Does the new world city system cut across pre-existing patterns of global inequality?

Method:

Estimate a series of conditional-change (OLS) and logistic regressions to determine the association of world system position (i.e., of a city being located in a core, semiperipheral, or peripheral country) with change in the power, prestige, and blockmodel position of cities 1981-2000.

Results for point centrality:

Outdegree: semiperipheral cities -11%, peripheral cities -21%

Closeness: semiperipheral cities -43%, peripheral cities -46%

Betweenness: semiperipheral cities -2%, peripheral cities -3%

Indegree: semiperipheral cities -17%, peripheral cities -23%

Figure 3. Reduced graphs of the world city system 1981 and 2000, nodes sized by proportion non-core cities in block

1981 2000

Q3: Does the new world city system cut across pre-existing patterns of global inequality?

Method:

Estimate a series of conditional-change (OLS) and logistic regressions to determine the association of world system position (i.e., of a city being located in a core, semiperipheral, or peripheral country) with change in the power, prestige, and blockmodel position of cities 1981-2000.

Results for blockmodel position/role (effect on odds of joining a block, 1981-2000):

Primary: semiperipheral cities = 0.053, peripheral cities = 0.007

Isolate: semiperipheral cities = 1.450, peripheral cities = 2.031

Figure 4. Reduced graphs of the world city system 1981 and 2000, nodes sized by proportion non-core cities in block (blue = isolate block)

1981 2000

Q3: Does the new world city system cut across pre-existing patterns of global inequality?

Method:

Estimate logistic regressions to determine the association of region (i.e., of a city being located in Africa, Latin America, or Asia) with change in the blockmodel position of cities 1981-2000.

Results for blockmodel position/role (effect on odds of joining a block, 1981-2000):

Primary: Africa = -------; Lat. Am. = 0.921; Asia = 1.719

Isolate: Africa = 3.978; Lat. Am. = 1.199; Asia = 1.230

Figure 5. Reduced graphs of the world city system 1981 and 2000, nodes sized by proportion African cities in block (blue = isolate block)

1981 2000

Figure 6. Reduced graphs of the world city system 1981 and 2000, nodes sized by proportion Latin American cities in block (blue = isolate block)

1981 2000

Figure 7. Reduced graphs of the world city system 1981 and 2000, nodes sized by proportion Asian cities in block (blue = isolate block)

1981 2000

Q3: Does the new world city system cut across pre-existing patterns of global inequality?

Conclusions:

Rather than cutting across the hierarchy of states in the inter-state system – across the Global North and South – the urban hierarchy maps onto it increasingly well. Cities located in core countries have grown relatively more powerful and prestigious, not less. They are relatively more likely to play the primary role in the world city system and are relatively less likely to be isolates.

Disaggregating the Global South, we again find little evidence for a new geography of inequality that is orthogonal to the “old.” African and Latin American cities are more likely to be “satellites” in 2000 than they were in 1981. The experience of Asian cities is more mixed.

Summary

We find that the world city system is in the midst of substantial restructuring and that it is changing in such a way as to concentrate power and prestige in a small number of cities. However, in contrast to some accounts, we find no evidence for a new geography of inequality that cuts across existing patterns of global inequality. Rather, we find just the opposite – the reproduction of the “old” geography in a starker form.

Limitations and Future Tasks

1) The city system that we describe and analyze should be compared with results of research rooted in different conceptualizations of the world city problematic (e.g., the Loughborough group) and validated with data that allow for more culturally-, socially-, and politically-informed senses “world-cityness.”

• unimodal/bi-modal nets• “levels” of command/control• sector/producer-service focus• cultural flows (e.g., people, telephony, internet)• political (GO/NGO & IGO/INGO nets) • slice and dice

2) The ultimate aim of world city research is to say something meaningful about the changing fortunes of cities and their residents. Combining our measures of power, prestige, and position with, for instance, readily available data at the SMSA (U.S.) or NUTS 5 (E.U.) levels, one could rigorously assess arguments about how the position of cities in the global urban network affects their fate.