Taylor & Francis, Ltd. Economic Geography: This Content Downloaded From 169.226.92.8 On Sun, 10 May 2020 22:31:46 UTC

Sticky Places in Slippery Space: A Typology of Industrial Districts
Author(s): Ann Markusen

Source: Economic Geography, Vol. 72, No. 3 (Jul., 1996), pp. 293-313
Published by: Taylor & Francis, Ltd.
Stable URL: https://www.jstor.org/stable/144402
Accessed: 10-05-2020 22:31 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to
Economic Geography
This content downloaded from 169.226.92.8 on Sun, 10 May 2020 22:31:46 UTC
All use subject to https://about.jstor.org/terms
Sticky Places in Slippery Space:
A Typology of Industrial Districts*
Ann Markusen
Project on Regional and Industrial Economics, Rutgers University, New
Brunswick, NJ 08901-1983
Abstract: As advances in transportation and information obliterate distance, cities

and regions face a tougher time anchoring income-generating activities. In probing
the conditions under which some manage to remain "sticky" places in "slippery"
space, this paper rejects the "new industrial district," in either its Marshallian or
more recent Italianate form, as the dominant paradigmatic solution. I identify three
additional types of industrial districts, with quite disparate firm configurations,
internal versus external orientations, and governance structures: a hub-and-spoke
industrial district, revolving around one or more dominant, externally oriented
firms; a satellite platform, an assemblage of unconnected branch plants embedded
in external organization links; and the state-anchored district, focused on one or
more public-sector institutions. The strengths and weaknesses of each are reviewed.
The hub-and-spoke and satellite platform variants are argued to be more prominent
in the United States than the other two. The findings suggest that the study of
industrial districts requires a broader institutional approach and must encompass
embeddedness across district boundaries. The research results suggest that a purely
locally targeted development strategy will fail to achieve its goals.
Key words: industrial districts, regional growth.
The Puzzle of Stickiness in an capital and labor. Movement is, of course,

Increasingly Slippery World costly and disruptive to both. David
Harvey's (1982) work on capital's need for
In a world of dramatically improved "spatial fix" and Storper and Walker's
communications systems and corporations (1989) work on labor and reproduction
that are increasingly mobile internation- suggest generic reasons why hypermobil-
ally, it is puzzling why certain places are
ity cannot completely obliterate produc-
able to sustain their attractiveness to both
tion ensembles in space. But neither
account explains why certain places man-
age to anchor productive activity while
* This research was supported by the Na- others do not.
tional Science Foundation, Program in Geogra- The problem is most acute in advanced
phy and Regional Science, and by Rutgers Uni-
capitalist countries, where wage levels
versity, though responsibility for the results
remains the author's. The author would like to and standards of living are substantially
thank Clelio Campolina Diniz, Masatomi Fun- higher than in newly incorporated labor-
aba, Amy Glasmeier, Elyse Golob, Mia Gray, rich and increasingly technically compe-
Bennett Harrison, Candace Howes, Andy Isser- tent countries (Howes and Markusen
man, Mary Ellen Kelley, Yong Sook Lee, Marlen 1993). Production space in these countries
Llanes, Michael Oden, Sam Ock Park, Eric has become increasingly "slippery," as the
Parker, Mohamad Razavi, Masayuki Sasaki, and
ease to capital of moving plants grows and
Frank Wilkinson for their insights from joint re-
search and conversation on this subject, and Bill as new competing lines are set up in
Beyers, Dan Knudsen, and three anonymous re- lower-cost regions elsewhere. Often the
viewers for comments on a previous version. Spe- only alternative for the region of exit or
cial thanks to Barbara Brunialti for research sup- any other aspirant appears to be matching
port. local production conditions to those in the
293
294 ECONOMIC GEOGRAPHY
competitor place, lowering wages and the state and/or multinational corpora-
reproduction costs to the lower common tions under certain circumstances to
denominator. Much of the stress on shape and anchor industrial districts,
improving local "business climates" providing
in a the glue that makes it difficult
country like the United States in thefor
pastsmaller firms to leave, encouraging
two decades is driven by the belief them
that to stay and expand, and attracting
localities have no other options. newcomers into the region. These models
Alarmed by the welfare implications of
exhibit greater propensities for network-
ing across district lines, rather than
such a strategy, economists, geographers,
and economic development planners within,
have and a much greater tendency to
sought for more than a decade for be exogenously driven and thus focused
alternative models of development oninexternal policy issues than do NIDs.
From
which existing activities are sustained or a welfare point of view, the four
transformed in ways that maintain types
rela- perform quite differently with re-
tively high wage levels, social wages,gard
and to income distribution, permissive-
ness toward labor organization, short-to-
quality of life. They have done so largely
with inductive methods, searching for medium-term
the cyclicality, and longer-term
exceptions to the rule and examining vulnerability
the to secular change.
structure and operation of such "sticky
places." One extensively researched for-
mulation is that of the "flexibly special-
Identifying and Analyzing
ized" or "new industrial district" (NID),
Sticky Places
based on the phenomenon of successful
expansion of mature industries in the The three alternative models of sticky
Emilio-Romagna region of Italy (Best
places developed in this paper were
1990; Goodman and Bamford 1989; Piore constructed through a process of induc-
and Sabel 1984; Scott 1988a, 1988b; tive inquiry similar to that used in
Storper 1989). NIDs owe their stickiness researching NIDs. In the NID literature,
to the role of small, innovative firms, intensive research on particular cases,
embedded within a regionally cooperative sometimes comparing across several, has
system of industrial governance which been used to identify causal forces and
enables them to adapt and flourish despite structural configurations. Piore and Sabel
globalizing tendencies. (1984, 1989) studied the Third Italy
In this paper, I argue that there are at intensively in developing their notions of
least three other types of industrial flexible specialization and industrial dis-
districts, or "sticky places," that have tricts. In the United States, Christopher-
demonstrated resiliency in the postwar son and Storper's work on the film
period in advanced industrialized coun- industry in Los Angeles (1986), Scott
tries. Stickiness connotes both ability to (1986) and Scott and Paul's work on
attract as well as to keep, like fly tape, and Orange County (1990), and Saxenian's
thus it applies to both new and estab- work on Silicon Valley (1990, 1991a,
lished regions. Based on an inductive 1991b, 1994) enabled these authors to
analysis of the more successful metropoli- derive propositions about how secular
tan regions in the United States, I show changes in technology and markets enable
that structures and dynamic paths quite and reward new forms of regional indus-
different from those captured in the NID trial organization. Vigorous debate on the
formulation have enabled both relatively accuracy and applicability of the NID
mature and up-and-coming regions to formulation ensued, enlivening the eco-
weather heightened capital mobility. Con- nomic geography literature for the better
trary to the emphasis on small firms in the part of a decade (e.g., Amin and Robins
NID formulation, these alternative mod- 1990; Amin and Thrift 1992; Ettlinger
els demonstrate the continued power of 1992; Florida and Kenney 1990; Gertler
STICKY PLACES 295
1988; Glasmeier 1988; Harrison 1992; Our method involved a two-stage pro-
Lovering 1990, 1991; Malecki 1987; cess. First, we surveyed metropolitan
Markusen 1991; Pollert 1988; Schoen- growth since 1970 for each of the four coun-
berger 1988). tries, identifying the universe of those who
The research summarized here had its posted growth rates significantly higher
origins in a larger research project to than the national average (tables showing
determine the extent to which the NID each of these regional sets may be found in
model could explain the durability andMarkusen (1995)). We then chose a subset
flourishing of regional economies in the of each of these for further case study re-
United States, Japan, Korea, and Brazil assearch, relying on both disaggregated data
adequately as it appeared to do so in the on industrial structure and expert opinion
Third Italy. Empirical testing of the NID on industrial organization. For each coun-
model has been surprisingly thin. Fewtry, we selected at least one case with ap-
attempts have been made to determine parent conformity to the NID formulation
whether existing agglomerations are "flex- and three to five others whose industrial
ibly specialized"-an exception is Feld- structure and organization appeared to be
man's (1993) remarkable study of U.S.quite different. We used techniques pio-
industrial agglomeration-or to deter- neered in social science case study re-
mine whether major industries are well search (Yin 1984) and leaned heavily on
characterized by this post-Fordist formu-interviews with business firms, trade asso-
lation (see Luria (1990) for an excellent ciations, trade unions, and regional econ-
investigation of the auto industry in thisomy watchers, incorporating and adding to
regard). No author has rigorously set outthe literature on enterprise studies and cor-
the features of new industrial districts in porate interviews as a method for studying
ways that permit easy assessment of their regions (McNee 1960; Krumme 1969;
incidence and growth across space and Schoenberger 1985, 1991; Healey and Raw-
time. The limits of the flexibly specialized linson 1993; Markusen 1994).
new industrial district as an emergent Conceptually, we inquired into the pres-
paradigmatic form (a claim made by Scott ence or absence of features specified in the
(1988a, 1988b)) are best established by NID formulation: firm size distribution, up-
demonstrating that other industrial dis- and downstream industrial linkages, de-
trict profiles are both theoretically plausi- gree of vertical disintegration, networks
ble and empirically demonstrable. among district firms, districtwide gover-
In each country studied in our larger nance structures, innovative capabilities,
project, it was clear that certain mature as the organization of production. In addi-
well as newer agglomerations exhibited an tion, we explored a number of features not
ability to weather the leveling effect of generally incorporated into NID studies
accelerated world market integration and (Park and Markusen 1994). First, we ex-
the global search for profitability, at- amined the role of the state at both the
tributes that make space "slippery." But national and regional/local level as rule
most of these enclaves did not match the maker, as producer and consumer of goods
features of the flexibly specialized indus- and services, and as underwriter of inno-
trial district of the NID literature. Just as vation, with consequences for the distri-
deindustrializing regions are quite re- bution and anchoring of employment within
markably distinguishable from each other, and across regions (Christopherson 1993,
as Massey and Meegan have deftly shown 1994; Linge and Rich 1991; Markusen et
(Massey and Meegan 1982; Massey 1984), al. 1991; Markusen and Park 1993; Saxe-
regions hosting rapid growth and/or es- nian 1995). Second, we scrutinized the role
caping industrial decline exhibit distinctly of large firms, especially those with inter-
different structures. Through inductive nal and external market power, in indus-
research, we were able to identify three trial agglomerations (Amin and Robins
alternative patterns. 1990; Dicken 1992; Gereffi and Korze-
niewicz 1994; Harrison 1994; Sayer 1989). ti

Third, we examined the embeddedness of
firms both within their districts and in non
local networks extending across national
international space (Granovetter 1985; Stor-
per and Harrison 1991; Linge 1991; NIDs. Like Storper and Harrison (1991),
Markusen 1994). Fourth, since sources of we opt for an expansive connotation of
profitability vary over the course of an inindustrial district, which does not confine
dustry's maturation and are linked to chang- it to the most common usage, called here
ing forms of competition, organizational the Marshallian (or Italianate variant)
structures, occupational characteristics, and district. Elsewhere, we offer the following
locational tendencies (Markusen 1985), we definition: an industrial district is a sizable
investigated the longer-term developmen- and spatially delimited area of trade-
tal dynamic of major industries and their oriented economic activity which has a
constituents present in the district, to de- distinctive economic specialization, be it
termine their resiliency and/or vulnerabil- resource-related, manufacturing, or ser-
ity to longer-term atrophy. Fifth, we as- vices (Park and Markusen 1994).
sessed the long-term dynamic potential of In what follows, I present four distinc-
each region, including the likely trajectory tive industrial spatial types: (1) the Mar-
and future competitiveness of its existing shallian NID, with its recent Italianate
industrial ensemble and the ability of the variety; (2) the hub-and-spoke district,
latter to release locally anchored re- where regional structure revolves around
sources, human and physical, into new, un- one or several major corporations in one
related specialized sectors. Finally, we or a few industries; (3) the satellite
searched for connections between district industrial platform, comprised chiefly of
structure and operation and a number of branch plants of absent multinational
social welfare metrics, including employ- corporations-this type of district may
ment growth rates over time, cyclical sta- either be comprised of high-tech branch
bility, associated income and wealth dis- plants or consist chiefly of low-wage,
tribution, trade union presence, and low-tax, publicly subsidized establish-
political diversity. ments; and (4) the state-centered district,
A bit more may be said about this final a more eclectic category, where
component of the research. Evaluation of government tenant anchors the re
the welfare implications of each type of sticky economy (a capital city, key m
place is a complex task and rarely under- research facility, public corporatio
taken. Scholars of NID literature have gen- hypothesized features of each are
erally written in a normatively favorable if rized in Table 1. Schematic visua
implicit way about the virtues of NIDs as of each of the first three, showin
providers of good jobs and long-term sta- firm size and interfirm connectio
bility and dynamism; this is especially pal- inside and outside the district, are
pable in the treatments of Piore and Sabel in Figure 1. Here, firm relat
(1984), Best (1990), and Saxenian (1994). A within the region are depicted insi
sticky place is "better," in our normative circle versus those outside of it--s
view, if it (1) ensures average or better-than- to the left, customers to the r
average growth for a region as a whole over real-world district may be an am
time; (2) insulates a region from the job loss one or more types, and over time
and firm failures of short-to-intermediate- may mutate from one type to a
term business or political spending cycles; This conceptualization complemen
(3) provides relatively good jobs, amelio- geographic industrialization sch
rates tendencies toward income duality, and Storper and Walker (1989,
prevents undue concentration of wealth and While theirs is process-centered
ownership; (4) fosters worker representa- offered here is region-centered, w
STICKY PLACES 297
the differences among sticky places and

their ability to persist are presence (or
absence) of distinctive and lopsided power
relationships, sometimes within the dis-
trict and sometimes between district
entities and those residing elsewhere.
Examples of each type can only be
mentioned in passing here, but are the
subject of complementary papers (e.g.,
Gray, Golob, and Markusen 1996; Golob
et al. 1995; Markusen and Park 1993;
Markusen and Sasaki 1994; Markusen
1994; Park and Markusen 1994).
The focus on rapidly growing industrial
spaces helps us to develop an impression-
istic sense of the relative contribution of
each type to overall regional restructuring.
In the United States, for instance, the fastest-
growing industrial cities (as opposed to res-
identiary cities, where retirement commu-
nities account for the bulk of growth) include
the 15 listed in Table 2, all of which added
manufacturing employment at rates of 50
percent or more over the period 1970 to
1990, compared with a zero rate of growth
nationally. These may be contrasted with
the performance of the four major older in-
dustrial centers of New York, Boston, Chi-
cago, and Los Angeles, at the bottom of the
table. Very few of these fast-growing re-
gions, I shall argue, can be characterized as
NIDs, but many- of them reproduce the con-
ditions present in the other three models of
"sticky places."
Marshallian and Italianate

Industrial Districts
Figure 1. Firm size, connections, and local
versus nonlocal embeddedness. An extensive and recent literature on in-
dustrial districts focuses on the Marshallian
industrial district and its more cooperative,
focus on firm size, interconnections, and embedded Italianate progeny. Since the
internal versus external orientations. characteristics hypothesized for these dis-
Each spatial type is presented with a tricts are relatively well known, I summa-
set of hypothesized traits, and the resil- rize them briefly here, with particular em-
ience and/or vulnerability of each to phasis on those which may be contrasted to
events in the changing global economy the district types presented below.
are noted. Districts which are sticky in In his original formulation of the
one era may fail to cohere in the longer industrial district, Marshall envisioned a
run-the glue may dry up, become brittle region where the business structure is
and lose its adhesive quality. Central to comprised of small, locally owned firms
Table 1
Hypothesized Features of New Industrial District Types
Marshallian industrial districts

* Business structure dominated by small, locally owned firms
* Scale economies relatively low
* Substantial intradistrict trade among buyers and suppliers
* Key investment decisions made locally
* Long-term contracts and commitments between local buyers and suppliers
* Low degrees of cooperation or linkage with firms external to the district
* Labor market internal to the district, highly flexible
* Workers committed to district, rather than to firms
* High rates of labor in-migration, lower levels of out-migration
* Evolution of unique local cultural identity, bonds
* Specialized sources of finance, technical expertise, business services available in district outside of firms
* Existence of "patient capital" within district
* Turmoil, but good long-term prospects for growth and employment
Italianate variant (in addition to the above)

* High incidence of exchanges of personnel between customers and suppliers
* High degree of cooperation among competitor firms to share risk, stabilize market, share innovation
* Disproportionate shares of workers engaged in design, innovation
* Strong trade associations that provide shared infrastructure-management, training, marketing, technical or
financial help, i.e., mechanisms for risk sharing and stabilization
* Strong local government role in regulating and promoting core industries
Hub-and-spoke districts
* Business structure dominated by one or several large, vertically integrated firms surrounded by suppliers
* Core firms embedded nonlocally, with substantial links to suppliers and competitors outside of the district
* Scale economies relatively high
* Low rates of turnover of local business except in third tier
* Substantial intradistrict trade among dominant firms and suppliers
* Key investment decisions made locally, but spread out globally
* Long-term contracts and commitments between dominant firms and suppliers
* High degrees of cooperation, linkages with external firms both locally and externally
* Moderate incidence of exchanges of personnel between customers and suppliers
* Low degree of cooperation among large competitor firms to share risk, stabilize market, share innovation
* Labor market internal to the district, less flexible
* Disproportionate shares of blue-collar workers
* Workers committed to large firms first, then to district, then to small firms
* High rates of labor in-migration, but less out-migration
* Specialized sources of finance, technical expertise, business services dominated by large firms
* Little "patient capital" within district outside of large firms
* Absence of trade associations that provide shared infrastructure -management, training, marketing, technical
or financial help, i.e., mechanisms for risk sharing and stabilization
* Strong local government role in regulating and promoting core industries in local and provincial and national
government
* High degree of public involvement in providing infrastructure
* Long-term prospects for growth dependent upon prospects for the industry and strategies of dominant firms
Satellite industrial platforms

* Business structure dominated by large, externally owned and headquartered firms
* Scale economies moderate to high
* Low to moderate rates of turnover of platform tenants
* Minimal intradistrict trade among buyers and suppliers
* Key investment decisions made externally
* Absence of long-term commitments to suppliers locally
* High degrees of cooperation, linkages with external firms, especially with parent company
* High incidence of exchanges of personnel between customers and suppliers externally but not locally
* Low degree of cooperation among competitor firms to share risk, stabilize market, share innovation
* Labor market external to the district, internal to vertically integrated firm
* Workers committed to firm rather than district (continued)
STICKY PLACES 299
Table 1
(continued)
* High rates of labor in-migration and out-migration at managerial, professional, technical levels; little at blue-
and pink-collar levels
* Little evolution of unique local cultural identity, bonds
* Main sources of finance, technical expertise, business services provided externally, through firm or external
purchase
* No "patient capital" within district
* No trade associations that provide shared infrastructure-management, training, marketing, technical, or
financial help, i.e., mechanisms for risk sharing and stabilization
* Strong local government role in providing infrastructure, tax breaks, and other generic business inducements
* Growth jeopardized by intermediate-term portability of plants and activities elsewhere to similarly con-
structed platforms
State-anchored industrial districts

* Business structure dominated by one or several large, government institutions such as military bases, state
or national capitals, large public universities, surrounded by suppliers and customers (including those reg-
ulated)
* Scale economies relatively high in public-sector activities
* Low rates of turnover of local business
* Substantial intradistrict trade among dominant institutions and suppliers, but not among others
* Key investment decisions made at various levels of government, some internal, some external
* Short-term contracts and commitments between dominant institutions and suppliers, customers
* High degrees of cooperation, linkages with external firms for externally headquartered supplier organizations
* Moderate incidence of exchanges of personnel between customers and suppliers
* Low degree of cooperation among local private-sector firms to share risk, stabilize market, share innovation
* Labor market internal if state capital, national if university or military facility or other federal offices for
professional/technical and managerial workers
* Disproportionate shares of clerical and professional workers
* Workers committed to large institutions first, then to district, then to small firms
* High rates of labor in-migration, but less out-migration unless government is withdrawing or closing down
* No specialized sources of finance, technical expertise, business services
* No "patient capital" within district
* Weak trade associations to share information about public-sector client
* Weak local government role in regulating and promoting core activities
* High degree of public involvement in providing infrastructure
* Long-term prospects for growth dependent on prospects for government facilities at core
that make investment and production external markets on the right, in the form
decisions locally. Scale economies are of exchange rather than cooperative rela-
relatively low, forestalling the rise of large tionships external to the region.
firms. Within the district, substantial What makes the industrial district so
trade is transacted between buyers and special and vibrant, in Marshall's account,
sellers, often entailing long-term contracts is the nature and quality of the local labor
or commitments. Although Marshall did market, which is internal to the district
not explicitly say so, linkages and/or and highly flexible. Individuals move from
cooperation with firms outside the district firm to firm, and owners as well as
are assumed to be minimal. The Marshal- workers live in the same community,
lian industrial district is depicted in the where they benefit from the fact that "the
top portion of Figure 1, with many smallsecrets of industry are in the air."
firms buying and selling from each otherWorkers are committed to the district
for eventual export from the region. Therather than to the firm. Labor out-
arrows show necessary purchases of rawmigration is minimal, while in-migration
materials and business services from occurs as growth permits. The district is
outside the region on the left and sales seen
to as a relatively stable community,
Table 2
Selected U.S. Metropolitan Employment Growth Rates, 1970-1990
Manufacturing Service
Employment Employment Employment, Manufacturing Em
1990 Change (%), 1990 Change (%), 1990 Change (%),
(in thousands) 1970-90 (in thousands) 1970-90 (in thousands) 1970-90
Colorado Springs, Colo. 228 104 24 261 60 214

Austin, Tex. 471 178 50 249 131 253
Reno, Nev. 145 155 9 202 70 184
Tucson, Ariz. 316 123 28 199 101 219
Huntsville, Ala. 163 76 34 177 42 82
Orlando, Fla. 569 246 56 162 236 465
Albuquerque, N. Mex. 305 125 22 131 98 184
Melboure/Titusville, Fla. 202 112 31 122 66 119
San Jose. Calif. 1,015 128 273 119 301 199
San Diego, Calif. 1,397 120 141 109 390 254
Anaheim-Santa Ana, Calif. 1,552 192 261 111 464 352
Raleigh-Durham, N.C. 513 123 66 94 145 175
Seattle, Wash. 1,339 114 227 73 362 206
Madison, Wis. 262 73 26 53 62 147
Elkhart-Goshan, Ind. 116 64 52 50 20 123
Los Angeles-Long Beach, Calif. 5,200 56 893 9 1,707 129
Boston-Lawrence-Salem, Mass. 1,672 30 340 -12 894 108
Chicago, Ill. 3,673 23 569 -33 1,128 101
New York, N.Y. 4,765 2 428 -51 1,704 50
United States 110,321 56 19,742 0 37,573 126
Source: U.S. Department of Commerce, Bureau of

were computed by Andrew Isserman and Oleg Smi
and compiled by Ann Markusen and Mia Gray.
which enables that any

the of these actors should
evolution of be stron
local cultural identity and shared
consciously cooperating with each otherindu
trial expertise. in order for the district to exist and
The Marshallian district also encom-operate as such. But in a more recent
passes a relatively specialized set offormulation, emerging from research on
services tailored to the unique products/ Italian industrial districts and extended to
industries of the district. These servicesother venues in Europe and the United
include technical expertise in certain States, researchers have argued that con-
product lines, machinery and marketing,certed efforts to cooperate among district
and maintenance and repair services.members and to build governance struc-
They include local financial institutions tures to improve districtwide competitive-
offering so-called "patient capital," willing ness can improve prospects-that is,
to take longer-term risks because they increase the stickiness of the district.
have both inside information and trust in Features characterizing Italianate dis-
the entrepreneurs of local firms. tricts are articulated in intensive case
All of these features are subsumable studies on the Italian case (Piore and
under the notion of agglomeration, whichSabel 1984; Bellandi 1989; Bull, Pitt, and
suggests that the stickiness of a place
Szarka 1991; Goodman 1989; Sforzi 1989).
resides not in the individual locational
These have been reworked and adapted to
calculus of firms or workers, but in the
American cases--Orange County (Scott
external economies available to each firm 1986; Scott and Paul 1990) and Silicon
from its spatial conjunction with other Valley (Saxenian 1994)-though not with-
firms and suppliers of services. In Mar- out debate (Malecki 1987; Florida and
shall's formulation, it was not necessary Kenney 1990; Saxenian 1991a). The unify-
STICKY PLACES 301
ing notion is that firms (often with the costs, innovation (so the theory goes) will
help of regional governments and trade ensure the revitalization of these "seed-
associations) consciously "network" to beds of innovation." But other hypotheses
solve problems of cycles and overcapacity have been advanced. Agglomerative spe-
and to respond to new demands for cialization and success in one industry,
flexibility (Amin and Thrift 1992). In the especially when associated with some
American version, rigidities in older in- degree of market power and/or domi-
dustrial cities tend to encourage these nance over regional factor markets, can
agglomerations to root anew in relatively actually impede the development of other
virgin locations (Markusen 1991; Scott sectors, whose presence might diversify
1988b; Storper and Walker 1989). Few the economy and counteract maturation
cases have been identified outside of or instability in the original sector. Pitts-
Europe or the United States, but good burgh in the late nineteenth century and
candidates for study are subdistricts Detroit
such in the early decades of the
as the southern sector of Tokyo and twentieth century resembled Italianate
districts and Silicon Valley, but the
Kangwan, a south side district in Seoul.
evolution of oligopoly and the crowding
Unlike the passivity of Marshall's firms,
Italianate districts exhibit frequent andout
in-of other sectors left both quite
tensive exchanges of personnel between vulnerable to the inevitable maturation
customers and suppliers and cooperation and decentralization of those industries
among competitor firms to share risk,(Chinitz
sta- 1960; Markusen 1985).
bilize markets, and share innovation. Dis- On the equity front, the high-tech
proportionate shares of workers areSilicon en- Valleys and Orange Counties
gaged in design and innovative activities.
depart strikingly from the Italian indus-
trial districts. Italian industrial districts
Activist trade associations provide shared
infrastructure - management, training, are often the creatures of resilient cul-
tures, organized politically on the basis of
marketing, technical, or financial help--as
well as providing forums to hammerlong-standing
out communities, unions, and
the Italian communist party. Fundamen-
collective strategy. Local and regional gov-
ernments may be central in regulatingtal andto their governance structures are
promoting core industries. Trust amongstrong
dis- leadership roles for unions and
guarantees
trict members is central to their ability to that most enterprises will be
stabilized and nurtured, even during
cooperate and act collectively (Harrison
1992; Saxenian 1994), although critics ar-
downturns. This has helped to stabilize
incomes
gue that the power of large corporations to and assure relatively good in-
shape Italian industrial districts has come
beendistributions within the districts. In
the California cases, in contrast, district
understated (see the discussion in Harri-
son 1994, Chap. 4). cooperation, where it exists, is purely
In assessing the growth, stability,between
eq- entrepreneurs and firms, who
uity, and politics of Italianate industrial
operate in a non-union environment and
districts, the Italian variety mustwherebe there is little preexisting commu-
distinguished from the Silicon Valleynity
and
to ameliorate vicious competition and
failure in periods of instability. Income
Orange County cases, and each from their
Marshallian predecessors. In terms of
distribution tends to be highly dualized in
such regions (Saxenian 1983; Harrison
growth and stability, as long as agglomer-
ation economies remain and are not 1994). Furthermore, politics within such
districts tends toward the conservative,
replicated in other locales, both Marshal-
lian and Italianate industrial districts laissez-faire end of the spectrum-Orange
retain good long-term prospects for
County is famous as the home of the John
growth and development. Although moreBirch society and Silicon Valley as a
standardized functions may be hived hotbed
off of free trade and anti-union
business
and driven elsewhere by inflated regional activism.
Despite the often extravagant claims of diversified developmental process where

some of its protagonists, the "new indus- new firms form few connections to hub
trial district" approach has much to offer firms other than benefiting from the
and has deservedly captured the imagina- urbanization and agglomeration econo-
tion of scholars and local economic devel- mies they have created.
opment activists alike. But many of the Hub-and-spoke districts are thus domi-
faster-growing regions of the world turn nated by one or several large, vertically
out not to be primarily characterized by integrated firms, in one or more sectors,
these same features. Furthermore, other surrounded by smaller and less powerful
structural forms may be associated with suppliers. Hub-and-spoke districts may
superior welfare and political cultures. It exhibit either a strongly linked form,
is to these other types of sticky places we where smaller firms are quite dependent
now turn. upon the large anchor firm or institution
for either markets or supplies, or a
weaker, more nucleated form, in which
small firms enjoy the agglomerative exter-
Hub-and-Spoke nalities of the larger organization's pres-
Industrial Districts ence without necessarily buying or selling
to them. In some versions, the large
Another quite different type of indus- player(s) may be oligopolists in a single
trial district is present in regions where a industry, as with the Big Three auto
number of key firms and/or facilities act as corporations in Detroit or Toyota in
anchors or hubs to the regional economy, Toyota City. Unrelated or loosely linked
with suppliers and related activities hubs in several industries may also coexist
spread out around them like spokes of a in a region. In Seattle, for instance, the
wheel. Examples are Seattle and central economy is organized around Weyer-
New Jersey, United States; Toyota City, hauser as the dominant resource-sector
Japan; Ulsan and Pohang, South Korea; company, Boeing as the dominant indus-
San Jose dos Campos in Brazil. A simple trial employer (commercial aircraft and
version of this form is depicted in the military/spacecraft), Microsoft as the lead-
middle frame of Figure 1, where a single ing services firm, the Hutchinson Cancer
large firm (e.g., Boeing in Seattle or Center as the progenitor of a series of
Toyota in Toyota City) buys from both biotechnology firms, and the Port of
local and external suppliers and sells Seattle as the transportation hub. Core
chiefly to external customers, who may be firms or institutions are embedded nonlo-
large (e.g., the airlines, the military in the cally, with substantial links to suppliers,
case of Boeing) or masses of individual competitors, and customers outside the
consumers (Toyota). Intensive case stud- district. Internal scale and scope econo-
ies of hub-and-spoke districts include mies are relatively high, and turnover of
Seattle (Gray, Golob, and Markusen firms and personnel is relatively low
1996), central New Jersey (Fineberg et al. except in third-tier suppliers or in major
1993), San Jose dos Campos and Campi- downturns in hub industries. Key invest-
nas, Brazil (Diniz and Razavi 1994). ment decisions are made locally, but their
The dynamism in hub-and-spoke econ- consequences are spread out globally.
omies is associated with the position of Hub-and-spoke districts may exhibit
these anchor organizations in their na- intradistrict cooperation, but it will gener-
tional and international markets. Other ally be on the terms of the hub firm.
local firms tend to have subordinate Substantial intradistrict trade will take
relationships to them. If over time place
the among suppliers and hub firms,
often embodied in long-term contracts
anchors evoke a critical mass of agglomer-
ated skilled labor and business services and commitments. Cooperation may en-
around them, they may set off a more tail efforts to upgrade supplier quality,
STICKY PLACES 303
timeliness, and inventory control, and it trade tend to be tied up as retained

may extend outside district boundaries to earnings in the major hub firms, who are
suppliers farther afield. Exchanges of happy to redeploy it wherever across the
personnel may take place, though not to globe their strategic plans call for. The
the extent found in Italianate industrial few trade associations that exist are
districts. Markedly lacking is the coopera- relatively weak, often because top hub
tion among competitor firms to share risk, managers absent themselves from their
stabilize the market, and share innova- deliberations and activities. Hub firms
tion. Strategic alliances on the part of the will concern themselves with state and
larger firms are more apt to be forged local governmental activities that impinge
with partners outside the region. upon their land use, tax, and regulatory
The labor market in hub-and-spoke situations and will try politically to ensure
districts is internal to both large hub firms that area politicians represent the inter-
and to the district, though it is less flexible ests of their firm and industries at the
than in the Italianate model. Workers' national and international levels. They
loyalties are to core firms first, then to themay also be actively involved in issues
district, and only after that to small firms.that affect their work force and ability to
If jobs open up in hub firms, workers will do business-especially in improving area
often abandon smaller employers to geteducational institutions and the provision
of infrastructure.
onto the hub firms' payroll. This factor
makes it tougher for smaller firms in some In the long run, hub-and-spoke districts
segments of the industry to survive. Hubare quite dependent on their major
firms attract new labor into the conurba- industries and firms within them for their
tion, however, which helps to counterbal- stickiness. Growth and stability can be
ance the power imbalance in the labor jeopardized by intermediate-term porta-
market. bility of plants and activities away from
Hub-and-spoke districts do evolve the region, or by the long-term decline of
unique local cultures related to hub the industry, or by poor management of
activities. Detroit is known as Motor City, the principal firms. But stickiness also
and sports teams of many cities have been depends on the degree to which mature
named after dominant sectors-the Oil- sectors can release local resources into
ers, the Steelers, the Brewers, the Pis- new, unrelated sectors. A sobering histor-
tons, the Millers (the old Minneapolis ical example of the vulnerability of hub-
team). They develop considerable exper- and-spoke districts is Detroit, where a
tise in the labor pool in specialized turn-of-the-century Marshallian district
industrial capabilities, and they engender (perhaps with some Italianate features)
specialized business service sectors tai-transformed itself into a hub-and-spoke
lored to their needs. Although these district around the auto oligopoly by the
business services are focused on the large 1930s. Here, to vastly oversimplify, De-
hub firms, some can become less depen-troit's vitality was severely taxed by the
dent by extending their markets to other oligopolistic rigidity of the locally head-
competitor firms in far-flung locales. Anquartered auto industry, combined with
extensive discussion of how a small firm concerted investment on the part of the
experiences its position in a hub-and- Japanese state and auto corporation in
spoke economy is included in Markusen building a rival agglomeration around
(1994). Toyota near Nagoya, Japan. Furthermore,
Districts of this sort lack some of the tight oligoponistic control over the De-
more celebrated governance structures of troit area's resources prevented the diver-
the Italianate industrial districts. They sification of its economy (Chinitz 1960). A
often lack "patient capital," local venture counter example is Seattle, where several
capital specially tailored to start-ups in unique features of Boeing as the undis-
their industry. The largest returns to puted anchor to the regional economy
(and the undisputed lead firm in the competitors and external suppliers or cus-
world aerospace industry) have contrib- tomers (Glasmeier 1988).
uted to (or at least not prevented) the Satellite platforms may be found in
region's diversification into other sec- almost all countries, regardless of devel-
tors-port-related activities, software, opment. An outstanding high-end exam-
biotechnology-positioning it well to ple in the United States is the internation-
withstand retrenchment and global de- ally much-admired Research Triangle
centralization in the aircraft industry Park, a collection of unrelated research
(Gray, Golob, and Markusen 1996). centers of major multinational corpora-
Hub-and-spoke industrial districts may tions (Luger and Goldstein 1990), while a
be characterized by relatively good in- comparable low-end U.S. case is Elkhart,
come distributions. If so, this is due to Indiana, where a number of auto-related
both structural and institutional causes. branch plants have been attracted by
Market power, often present in hub-and-
relatively low-wage labor. In South Korea,
spoke cases, results in relatively highKumi constitutes a low-end textile and
electronics platform, while Ansan oper-
returns to capital, a necessary though not
sufficient condition for sharing of suchates as an odd collection of disparate
returns with the work force in the form ofindustrial
polluters grouped together
higher wages. The presence of large (Park and Markusen 1994). In Japan, some
anchor firms, nonprofit and public institu-of the better-performing technopoles,
tions may also reflect natural economies ofsuch as Oita and Kumamoto, fall into this
scale, which are associated with largecategory (Markusen and Sasaki 1994). In
capital outlays and therefore high levels ofBrazil, a remarkable case is the state-
labor productivity, available for distribu-sponsored expansion of Manaus as an
import/export zone (Diniz and Borges
tion in wages. Securing this labor share is
most often dependent on the presence of Santos 1995).
unions or the threat of their emergence. In satellite platforms, business struc-
More vigorous political competition be-ture is dominated by large, externally
tween probusiness and prolabor constitu-situated firms that make key investment
encies is apt to hold sway in such districts.decisions. Scale economies within each
facility are moderate to high, and rates of
Satellite Platforms turnover of platform tenants are low to
moderate. Minimal intradistrict trade or
Yet a third variant of rapidly growing even conversation takes place among
platform tenants. Orders and commit-
industrial districts may be termed the sat-
ellite platform-a congregation of branch ments to local suppliers are conspicuously
facilities of externally based multiplant absent. Since platforms generally host
firms. Often these are assembled at a dis- heterogeneous firms in terms of product if
tance from major conurbations by national not industry and are remotely controlled,
governments or entrepreneurial provin- they do not operate as cooperative ven-
cial governments as a way of stimulating tures among resident plants to share risk,
regional development in outlying areas and stabilize the market, or engage in innova-
simultaneously lowering the cost of busi- tive partnerships. In this they differ from
ness for competitively squeezed firms bris- hub-and-spoke district, where the large
tling under relatively high urban wages, multilocational firm or institution is lo-
rents, and taxation. Tenants of satellite plat- cally based. This type of sticky place is
forms may range from routine assembly presented in the lower portion of Figure
functions to relatively sophisticated re- 1; its most conspicuous feature is the
search, but they must be able to more or absence of any connections or networks
less "stand alone," detachable spatially from within the region and the predominance
either up- or downstream operations within of links to the parent corporation and
the same firm or from agglomerations of other branch plants elsewhere.
STICKY PLACES 305
It is not as if branch operations, local government efforts and services

however, are not embedded in relation- offered by Chambers of Commerce and
ships external to the facility. They cooper- other associations of local fixed capital.
ate and communicate daily with the Satellite platforms' future growth is
parent company. Personnel exchanges are jeopardized by the intermediate-term
common between branch operations and portability of plants and activities else-
the headquarters firm, but not locally with where to similarly constructed platforms.
other branch facilities. To buttress this Those concentrating on higher-end activi-
nonplace embeddedness, the labor mar-ties, where stability and amenities in the
ket within which each facility operates, atresidential sphere are essential to drawing
least in the high-end version and for and keeping skilled personnel, will be less
management and some technical talent in vulnerable in this regard, while purely
the low-end version, cuts across districtlow-cost districts will be more so, espe-
boundaries; it is internal to the verticallycially if fixed capital investment is low.
Since individual plants and facilities are
integrated firm, rather than to the district.
This means that there will be high rates of disparate and outward looking, satellite
platforms do not engender the develop-
labor migration in and out of the district at
the managerial, professional, and techni- ment of unique local cultural bonds or
cal levels. Often skilled professionals who new identities, even though they may
originated from the region will be dispro-destroy preexisting ones. Thus they may
portionally represented. Only blue- andbe less sticky, especially if less skilled,
pink-collar labor will be hired locally,
than other types of district. Hosting
which may, however, not be inconsequen-
communities face the challenge of trying
tial. to parlay resources assembled by such
Over time, districts built around plat- facilities into other diversifying and home-
forms may begin to host growth of grown sectors. They do remain sticky,
suppliers, oriented toward platform ten- however, to the extent that large capital
ants, and they may enjoy some increase in investments are made in the process of
local entrepreneurship because the plat- occupying them.
form enhances the pool of skilled person- The record on income distribution in sat-
nel resident in the region. But in cases ellite platforms is mixed. In all countries
studied to date, the incidence of such studied, the entry of such platforms into
activity is small, and the aggregate growth previously depressed regions does contrib-
ute to higher overall per capita incomes
of the region is still very much tied to the
number of tenants that can be attracted (and perhaps a depression of those in re-
and to the ability to retain them (Howes gions of exit). Within the region, income
1993). distributional consequences depend on the
A number of features of the satellite nature of the industry and activity. Good
platform constrain its development into blue-collar
a jobs in a depressed agricultural
better-articulated regional economy. First region will improve the income distribu-
of all, the main sources of finance, tion. In technical branch plant platforms,
technical expertise, and business services the creation of a significant number of cler-
are external to the region, furnished ical and technician jobs may help to ame-
through corporate headquarters. Satellite liorate the skewness introduced by opera-
districts have little "patient capital" to tions that are top-heavy with managers and
draw upon, and because substantive activ- professionals. This seems to have occurred
ities are diverse, they lack industry- in the case of Research Triangle Park (Luger
specific trade associations that would and Goldstein 1990). However, satellite
provide shared infrastructure and help platforms by their very nature artificially
with management, training, and market- cordon off employment in some operations
ing problems. These will only be partially of a corporation from those in other re-
compensated for by strong national or gions, spreading income inequality out spa-
tially. Somewhat better jobs for rural Ja- performance to the presence, new loca-
pan or small-town Alabama placed on a tion, or expansion of state facilities. Mili-
satellite platform obscure the concentra- tary bases, military academies, and weap-
tion of top-paid corporate jobs elsewhere ons labs, for instance, explain the
and the deterioration in the income distri- phenomenal postwar growth of U.S. cities
bution in a Detroit or inner-city Tokyo, like Santa Fe, Albuquerque, San Diego,
especially for blue-collar workers. and Colorado Springs, while defense plants
The implications for the complexion of contributed dramatically to the growth of
politics in satellite platform regions are also Los Angeles, Silicon Valley, and Seattle
mixed. In some cases-Japan, for in- (Markusen et al. 1991). State universities
stance-the creation of such platforms un- and/or state capitals explain the promi-
der the technopolis strategy has co-opted nence of cities like Madison, Ann Arbor,
militant, often antibusiness prefectural Sacramento, Austin, and Boulder among
movements for environmental cleanup and fastest growing U.S. cities. Denver owes
an improved quality of life, redirecting their much of its postwar growth to its hosting of
energies and local resources into specula- the second largest concentration of federal
tive economic development activities. In government offices in the nation. In Japan
other cases, new satellite platforms have and South Korea, the government re-
helped break the stranglehold of tradition- search complexes at Tsukuba and Taejon,
ally dominant "good old boy" parties by respectively, have fueled growth in their
introducing educated people and new im- environs. In Brazil, Campinas owes much
migrants into the region and contributing to its top-ranked university, while San Jose
to more contested local politics. dos Campos's growth is based on the gov-
ernment-owned, military-oriented aero-
State-anchored Districts space complex (Diniz and Bazavi 1994).
In general, scale economies are rela-
A fourth form of sticky place is what we tively high in such complexes. Because
call the state-anchored industrial district,state-owned or state-dependent facilities
where a public or nonprofit entity, be it a are so large, supplier sectors do grow up
military base, a defense plant, a weaponsaround them, dependent on the level of
lab, a university, a prison complex, or apublic expenditure. Short-term contracts
concentration of government offices, is aand commitments do exist between state
key anchor tenant in the district. Here,"customers" and their suppliers, subject
the local business structure is dominated to political change. In the case of state
by the presence of such facilities, whose capitals and universities, high degrees of
locational calculus and economic relation- cooperation may exist between the cus-
ships are determined in the political tomer and suppliers, and activity will be
realm, rather than by private-sector firms. relatively immune from the threat of
This type of district is much more difficult exodus. This is less true for national
to theorize, because contingencies partic- facilities, especially in times of fiscal
ular to the type of activity involved color stringency or redundancy of function
its operation and characteristics. It is apt (e.g., the current closing of military bases
to look much like the hub-and-spoke in the United States). In nationally funded
district in Figure 1, although a facility can facilities, decisions are made external to
operate with few connections to the the district and may be more indifferent
regional economy, resembling the satel- to regional development impacts.
lite platform case. Nevertheless, some When government contracting is in-
commonalities may be noted. volved, especially in areas like defense,
Before doing so, however, I shall simply the arcane and elaborate nature of the
cite examples of such districts. Many of contracting process may encourage the
the fastest growing industrial districts in development of long-term supply rela-
the United States and elsewhere owe their tionships, based on a fairly strong degree
STICKY PLACES 307
of trust and cooperation. However, these tion to the local economy. Often, the
ties need not be localized; they may span mammoth size of the facility-New Mexi-
thousands of miles between Los Angeles co's Los Alamos Laboratories, for in-
or Silicon Valley and Washington, D.C., stance, with an annual budget of $1.4
for instance, or most of the length of a billion, mostly for personnel, or New
country like Korea, as between Changwon London, Connecticut's Electric Boat sub-
and Seoul (Golob et al. 1995; Markusen marine manufacturing facility, with its
and Park 1993). 20,000 workers-overwhelms any contri-
Labor markets will be tailored to the bution, real or potential, that may be
particular state activity hosted. For state
made through second effects. This means
capitals, the labor market will tend to bethat local business and political energies
relatively local or regional. Personnel may tend to be focused on solidifying the
cycle between state customers and localfacility's commitment and its level of
suppliers. For universities and nationalfunding. This must be pursued through
facilities, labor markets will operate exter-politics at the relevant level and thus
nally for the higher-skilled occupations. Inrequires a relatively unique governance
the case of military bases, blue-collar and
structure.
unskilled positions will also be filled from Politics in state-anchored industrial

a labor market national in scope. Workers' districts tend to be complex and tailored
loyalties will be devoted to large state
to the particularities of the form of
institutions and/or state-dependent facili-government involvement. Military-indus-
ties first, then to the district, then totrial districts range from the remarkably
firms. conservative (Colorado Springs) to the
Indigenous firms will play less of a role remarkably liberal (New England). Uni-
in these districts than in Marshallian or versity towns and state capitals tend to be
hub-and-spoke districts. Some may emerge more liberal than other cities of similar
out of specialized technology transfer (uni-size, even within their own states, while
versities) or business service functions (lob-towns hosting military bases and prisons
bying). Firms will not tend to cooperate to tend to line up on the conservative end of
stabilize markets or hedge against risk sincethe spectrum.
they are not preoccupied with stabilizing
demand in the same way that Marshallian
Sticky Mixes
districts with mature industries might. In
general, trade associations will be rela- Although the presence of Marshallian
tively weak, and local government's role inindustrial districts, even the Italianate
regulating and promoting district activitiesversion, can be confirmed in a number of
will be minimal (consider, for instance, theAmerican instances, the claims made for
District of Columbia's almost complete ab- the paradigmatic ascendancy of this form
sence of power). Local fixed capital and of new industrial space (Scott's rubric) do
government may adopt a sycophantic formnot square with the experience of most
ofboosterism, designed to enhance the abil-rapidly growing agglomerations in indus-
ity of the anchor facility to maintain or in-trialized and industrializing countries. In
crease levels of external funding or protectthe United States, for instance, most
it against closure. rapidly growing industrial regions do not
In state-anchored industrial districts,exhibit the characteristics of the Third
long-term growth prospects depend on Italy. Indeed, the lessons of the Italian
two factors: the prospects for the facility atindustrial district experience are being
the core of the region, and the extent toadopted most fruitfully in the Industrial
which the facility encourages growthMidwest as a way of stemming deindustri-
within the region by spawning local alizing and retaining jobs in small and
medium-sized firms, not in explaining
suppliers, spinning off new businesses, or
supplying labor or other factors of produc-new industrial spaces. Even Silicon Val-
ley, as we show elsewhere, is more a mix and how they vary by type of industry and
of industrial district types than a pure case degree of maturity, national and regional
of Italianate industrial district (Golob et rules and cultures, and firm and local
al. 1995). In Japan, South Korea, and economic development strategy.
Brazil, it is difficult to find a single Many localities, especially larger metro-
instance of a flexibly specialized industrial politan areas, exhibit elements of all four
district outside of subareas of the major models. Silicon Valley, for instance, hosts
metropolis. Most rapidly growing metro- an industrial district in electronics (Saxe-
politan areas owe their performance to nian 1994) but also revolves around
hub firms or industries, satellite plat- several important hubs (Lockheed Space
forms, and/or state anchors, or some and Missiles, Hewlett Packard, Stanford
combination thereof. University), as well as hosting large
In the United States, the fast-growing "platform" type branch plants of U.S.,
industrial cities in Table 2 may be Japanese, Korean, and European compa-
allocated to one or more of our industrial nies (e.g., IBM, Oki, NTK Ceramics,
district types. Colorado Springs, Hunts- Hyundai, Samsung). Furthermore, Silicon
ville, Melbourne/Titusville, and San Di- Valley is now and has been the fourth
ego, all military or space-dependent cit- largest recipient of military spending
ies, belong in the fourth, government contracts in the nation, a fact that shapes
facility-anchored growth areas. Madison, it defense electronics and communications
Austin, and Albuquerque also belong in sector (Saxenian 1985; Markusen et al.
this category, the first two because they 1991; Golob et al. 1995).
house both the state university and state An intriguing question is whether re-
capital, and Albuquerque because it hosts gions can maintain their stickiness by
the state capital, state university, and transforming themselves from one type of
various military-related facilities, includ- district to another. Historically, as I have
ing nearby Los Alamos and Sandia labora- pointed out, Detroit made the transition
tories. Reno and Orlando's growth is from a Marshallian district to a hub-and-
primarily entertainment-related, although spoke district. Localities that host satellite
in recent years Reno has benefited from platforms may be able to encourage
warehousing and related operations flee- backward and forward linkages that trans-
ing California's tax structure. Seattle, Los form them into more Marshallian or
Angeles, and the latter's Anaheim/Santa hub-and-spoke type districts; scholars are
Ana neighbor are hub-and-spoke districts debating whether this is occurring around
organized around large defense and com- large Japanese auto transplants in the
mercial corporations, with universities United States. A state-centered district
playing larger or smaller roles. Raleigh- might do the same. A hub-and-spoke
Durham is a prototypical case of a district which loses its anchor tenant may
successful high-tech satellite platform, be able to create a Marshallian district in
while Elkhart-Goshen has flourished from its wake, as some are trying to do in the
low-wage, non-union capacity additions in Los Angeles aerospace industry. Recruit-
aging industries. ment or incubation of a new hub could
The models of sticky places presented transform a Marshallian or state-centered
district into a hub-and-spoke variant,
above are suggestive rather than defini-
tive products of an inductive research which is what Colorado Springs has been
doing with new organizational headquar-
method. Further application of these to an
even broader set of regional economies ters like the U.S. Olympics and the
will be necessary to determine how well right-wing Christian Focus on the Family.
each is constructed and how common its More work could be done on the condi-
incidence is in real space. Comprehensive tions that impede or facilitate these
comparative work across a larger appliedmutations.
set could tell us much about district forms The research reported here was method-
STICKY PLACES 309
ologically confined to places doing better than study why certain places appear to be
average, simply because this ensured that different and/or more successful as a
they met the criterion of superior growth means of developing hypotheses regard-
performance. However, many localities with ing features that may contribute to such
stable or slowly declining growth patterns success. Once identified, these then need
are struggling to be sticky places, and many to be tested against a larger sample, one
are succeeding in stanching their losses by more representative of the universe of
remaking their industrial structures. New En- localities.
gland, for instance, began as early as the 1950s
Furthermore, the study of industrial dis-
to transform itself into a diversified military-
tricts and networks within them has gen-
industrial complex, escaping the deeper dis-
erally been confined to smaller firms in
placement that occurred post-1970 in the In-
particular industries; their links to larger
dustrial Midwest (Markusen et al. 1991).
firms and to other firms and institutions
Although New England has not as a region
posted above-average long-term growth rates,
outside the region have been ignored. As a
result, conclusions have been drawn about
even during the Reagan military buildup, it
deserves study as a sticky place. Midwestern the endogeneity of growth in such districts
cities like Chicago, Milwaukee, and Cleve- that, when viewed on a larger, more com-
land with little comparative advantage in mil- prehensive canvas, are not warranted. Nor
itary-industrial sectors are trying to make is the zero-sum nature of much of this
themselves more sticky by anchoring growth acknowledged--that certain places
and upgrading existing expertise in indus- grow at the expense of other places, that
tries like metals, machining, and automo- high-wage employment in some regions is
biles. linked to low-wage employment in others,
Our study was conducted at the metro- and that only a few places can possibly as-
politan scale, equivalent more or less to a pire to become Silicon Valleys of the fu-
regional labor shed. However, industrial ture.
district features may characterize smaller In reality, sticky places are complex
agglomerations within metropolitan areas. products of multiple forces: corporate
Extension of these models to the subre- strategies, industrial structures, profit
gional scale might require relaxing one or
cycles, state priorities, local and national
more assumptions and altering some politics. Their success cannot be studied
hypotheses. by focusing only on local institutions and
behaviors, because their companies
(through corporate relationships, trade
Research and Policy Implications
associations, trade, government con-
This exercise in distinguishing among tracts), workers (via migration and inter-
types of sticky places illustrates the national unions), and other institutions
diversity in spatial form, industrial com- (universities, government installations)
plexion and maturity, institutional config- are embedded in external relationships--
urations, and welfare outcomes found in both cooperative and competitive-that
contemporary regional economies. It cau- condition their commitment to the local-
tions that the singular enthusiasm for ity and their success there.
flexibly specialized industrial districts, These reflections on research approach
especially the high-tech American variant, are applicable to economic development
is ill-founded on both growth/stability and policy at both regional and national levels
equity grounds. In large part, the problem as well. At the regional level, economic
here lies in the limits of the research developers would be well advised to
assess their existing district structures
strategy used in the NID literature, which
intensively studies particular localities
accurately and design a strategy around
extracted from their embeddedness in a them, rather than committing to a fashion-
larger global economy. It is useful toable strategy of small-firm networking
within the region. Improving cooperative alliances might be as important to locali-

relationships and building networks that ties as purely local networking ap-
reach outside of the region may prove proaches. Cross-regional networks might
more productive for some localities than be forged to shore up progressive institu-
concentrating on indigenous firms. Fur- tions under attack (labor, environmental
thermore, our work on hub-and-spoke and community development gains) and
and satellite platform structures suggests create better ones at national and interna-
that large firms can be significant contrib- tional levels to curb the worst products of
utors to regional development, albeit capitalist development-poverty, insecu-
posing problems of dominance and vul- rity, income inequality, environmental
nerability, and that recruitment of an degradation. While NID district builders
external firm or plant may be a good struggle to create governance structures
strategy for a region at a particular at the local level, multinational finance
developmental moment. Regions might and industrial leaders have crafted a
also be well advised to target national- World Trade Organization that would be
level policies shaping the competitive highly undemocratic and preempt many
status of their industries and allocating of the existing rights and safeguards that
public infrastructure and procurement workers and communities have fought for
contracts. and won. More sophisticated and pluralis-
At the national level, a strategytictoprofiles of industrial districts and how
they operate, both internally and exter-
ameliorate regional competition and dif-
nally,
ferential growth rates would (1) attempt to must be joined with more intensive
determine how many districts of each study of multinational corporations and
type the national economy mightstate be institutions if a more powerful
geographic contribution to progressive
expected to sustain, (2) develop a strategy
strategy is to emerge.
for stabilizing existing districts and chan-
neling new ones to deficit areas, (3) ban
the use of public funds to subsidizeReferences
competition among regions, and (4) moni-
tor and if necessary alter national policies
Amin, A., and Robins, K. 1990. The re-
with substantial regional implications,
emergence of regional economies? The
such as devolution of powers and respon-
mythical geography of flexible accumulation.
Environment and Planning D: Society and
sibility to subnational levels, new trade
Space 8:7- 34.
regimes (e.g., North American Free Trade
Amin, A., and Thrift, N. 1992. Neo-Marshal-
Agreement, General Agreement on Tariffs
lian nodes in global networks. International
and Trade), macroeconomic policy initia-
Journal of Urban and Regional Research
tives (e.g., deficit reduction and fiscal
16:571- 87.
austerity versus stimulus), financial mar-
Bellandi, M. 1989. The industrial district in
ket structures, Third World development,Marshall. In Small firms and industrial
international labor and human rights,
districts in Italy, ed. E. Goodman and J.
international environmental standards, Bamford, 136- 52. London: Routledge.
immigration restrictions, social safety Best, M. 1990. The new competition: Institu-
tions of industrial restructuring. Cam-
nets, and infrastructure provisions. In the
bridge: Harvard University Press.
United States at present, only the third of
these has any near-term possibility of A.; Pitt, M.; and Szarka, J. 1991. Small
Bull,
firms and industrial districts, structural
being undertaken and then only as a
explanations of small firm viability in three
result of considerable bipartisan clamor in
countries. Entrepreneurship and Regional
Congress. Development 3:83- 99.
The prominence of hub-and-spoke and Chinitz, B. 1960. Contrasts in agglomeration:
satellite platforms among U.S. sticky New York and Pittsburgh. American Eco-
places suggests that economic develop- nomic Association, Papers and Proceedings
ment strategies built on cross-regional 40:279-89.
STICKY PLACES 311
Christopherson, S. 1993. Market rules and of the Institute of British Geographers

territorial outcomes: The case of the United 13:419- 32.
States. International Journal of Urban and Glasmeier, A. 1988. Factors governing the
Regional Research 17:274- 88. development of high tech industry agglom-
.1994. Rules as resources in investment erations: A tale of three cities. Regional
and location decisions. Paper presented at Studies 22:287- 301.
the Centenary Conference for Harold Innis, Golob, E.; Gray, M.; Markusen, A.; and Park,
University of Toronto, September. S. 0. 1995. Valley of the heart's delight:
Christopherson, S., and Storper, M. 1986. The Silicon Valley reconsidered. Working Paper,
city as studio, the world as back lot: The Project on Regional and Industrial Econom-
impact of vertical disintegration on the ics, Rutgers University, presented at the
location of the motion picture industry. Regional Science Association annual meet-
Environment and Planning D: Society and
ings, Niagara Falls, Canada, November,
1994.
Space 4:305-20.
Goodman, E. 1989. Introduction: The political
Dicken, P. 1992. Global shift: The internation-
economy of the small firm in Italy. In Small
alization of economic activity. 2d ed. New firms and industrial districts in Italy, ed. E.
York: Guilford.
Goodman and J. Bamford, 1-3. London:
Diniz, C. C. 1994. Polygonized development Routledge.
in Brazil: Neither decentralization nor con-
Goodman, E., and Bamford, J., eds. 1989.
tinued polarization. International Journal of Small firms and industrial districts in Italy.
Urban and Regional Research 18:293- 314. London: Routledge.
Diniz, C. C., and Borges Santos, F. 1995. Granovetter, M. 1985. Economic action and
Manaus: A satellite platform in the Amazon social structure: The problem of embedded-
region. Working Paper, CEDEPLAR, Uni- ness. American Journal of Sociology 91:481-
versidad Federal de Minas Gerais, Brazil, 510.
May. Gray, M.; Golob, E.; and Markusen, A. 1996.
Diniz, C. C., and Razavi, M. 1994. Emergence Big firms, long arms: A portrait of a "hub
of new industrial districts in Brazil: Sao Jose and spoke" industrial district in the Seattle
dos Campos and Campinas cases. Working region. Regional Studies, forthcoming.
Paper, CEDEPLAR, Universidad Federal Harrison, B. 1992. Industrial districts: Old
de Minas Gerais, Brazil, November. wine in new bottles? Regional Studies
Ettlinger, N. 1992. Modes of corporate organi- 26:469-83.
zation and the geography of development. 1994. Lean and mean: The changing
Papers in Regional Science 71:107- 26. landscape of corporate power in the age of
Feldman, M. 1993. Agglomeration and indus- flexibility. New York: Basic Books.
trial restructuring. CPAD Working Paper Harvey, D. 1982. The limits to capital.
93-02, Graduate Program in Community London: Basil Blackwell.
Planning and Area Development, University Healey, M., and Rawlinson, M. 1993. Inter-
of Rhode Island, June. viewing business owners and managers: A
Fineberg, D.; Gilmore, R. W.; Krantz, J.; review of methods and techniques. Geofo-
Lianes, M.; Miller, R.; Mann, U.; and rum 24:339-55.
Schmitt, B. 1993. The biopharmaceutical Howes, C. 1993. Constructing comparative
industry in New Jersey: Prescriptions for disadvantage: Lessons from the U.S. auto
regional economic development. Report to industry. In Trading industries, trading
the Princeton/Rutgers Research Corridor, regions, ed. H. Noponen, J. Graham, and A.
Department of Urban Planning and Policy Markusen, 45-91. New York: Guilford.
Development, Rutgers University. Howes, C., and Markusen, A. 1993. Trade,
Florida, R., and Kenney, M. 1990. Silicon industry and economic development. In
Valley and Route 128 won't save us. Trading industries, trading regions, ed. H.
California Management Review 33:68- 88. Noponen, J. Graham, and A. Markusen,
Gereffi, G., and Korzeniewicz, M., eds. 1994. 1-44. New York: Guilford.
Commodity chains and global capitalism. Krumme, G. 1969. Toward a geography of
Westport, Conn.: Greenwood. enterprise. Economic Geography 45:30-40.
Gertler, M. 1988. The limits to flexibility: Linge, G. J. R. 1991. Just-in-time: More or less
Comments on the post-Fordist vision of flexible? Economic Geography 67:316-32.
production and its geography. Transactions Linge, G. J. R., and Rich, D. C. 1991. The
state and industrial change. In The state and Industrial Economics, Rutgers University,
the spatial management of industrial New Brunswick, N.J., January.
change, ed. G. J. R. Linge and D. C. Rich, Massey, D. 1984. Spatial divisions of labor:
1-21. London: Routledge. Social structures and the geography of
Lovering, J. 1990. Fordism's unknown succes- production. New York: Methuen.
sor: A comment on Scott's "Theory of Massey, D., and Meegan, R. 1982. The
flexible accumulation and the re-emergence anatomy of job loss: The how, why and
of regional economies." International Jour- where of employment decline. London:
nal of Urban and Regional Research 14:159- Methuen.
74.
Park, S. 0. and Markusen, A. 1994. Generaliz-
. 1991. Theorising post-Fordism: Why ing new industrial districts: A theoretical
contingency matters (A further response to
agenda and an application from a non-
Scott). International Journal of Urban and Western economy. Environment and Plan-
Regional Research 15:298-301.
ning A 27:81-104.
Luger, M., and Goldstein, H. 1990. Technol-
Piore, M., and Sabel, C. 1984. The second
ogy in the garden. Chapel Hill: University of
North Carolina Press. industrial divide: Possibilities for prosperity.
New York: Basic Books.
Luria, D. 1990. Automation, markets and
scale: Can "flexible niching" modernize U.S. Pollert, A. 1988. Dismantling flexibility. Capi-
tal and Class 34:42-75.
manufacturing? International Review of Ap-
plied Economics 4:127-65. Sabel, C. 1989. Flexible specialization and the
McNee, R. 1960. Toward a more humanistic reemergence of regional economies. In
economic geography: The geography of Reversing industries decline, ed. P. Hirst
enterprise. Tijdschrift voor Economische en and J. Zeitlin, 17-70. New York: St. Martin's
Social Geografie 51:201-05. Press.
Malecki, E. J. 1987. Comments on Scott's Saxenian, A. 1983. The urban contradictions of
"High tech industry and territorial develop- Silicon Valley. International Journal of
ment: The rise of the Orange County Urban and Regional Research 17:236-57.
complex, 1955-1984." Urban Geography . 1985. The genesis of Silicon Valley. In
8:77- 81. Silicon landscapes, ed. P. Hall and A.
Markusen, A. 1985. Profit cycles, oligopoly and Markusen, 20-34. Boston: Allen and Unwin.
regional development. Cambridge: MIT . 1990. Regional networks and the
Press. resurgence of Silicon Valley. California
. 1991. The military industrial divide: Management Review 32:89-112.
Cold war transformation of the economy and . 1991a. Silicon Valley and Route 128
the rise of new industrial complexes. Envi- won't save us: Response to Richard Florida
ronment and Planning D: Society and Space and Martin Kenney. California Management
9:391-416. Review 33:136-42.
. 1994. Studying regions by studying . 1991b. The origins and dynamics of
firms. The Professional Geographer 46:477- production networks in Silicon Valley. Re-
90. search Policy 20:423-37.
. 1995. The interaction of regional and . 1994. Regional networks: Industrial
adaptation in Silicon Valley and Route 128.
industrial policies: Evidence from four coun-
tries. Proceedings, World Bank's conference Cambridge: Harvard University Press.
on development economics, 1994. Washing- Sayer, A. 1989. Post-Fordism in question.
ton, D.C.: World Bank. International Journal of Urban and Re-
Markusen, A.; Hall, P.; Campbell, S.; and gional Research 13:666-95.
Schoenberger, E. 1985. Foreign manufactur-
Deitrick, S. 1991. The rise of the Gunbelt.
New York: Oxford University Press. ing investment in the United States: Com-
Markusen, A., and Park, S. 0. 1993. The state petitive strategies and international location.
as industrial locator and district builder: The Economic Geography 61:241-59.
case of Changwon, South Korea. Economic . 1988. From Fordism to flexible
Geography 69:157-81. accumulation: Technology, competitive
Markusen, A., and Sasaki, M. 1994. Satellite strategies, and international location. Envi-
new industrial enclaves: A comparative ronment and Planning D: Society and Space
study of United States and Japanese cases. 6:245-62.
Working Paper, Project on Regional and . 1991. The corporate interview as a
STICKY PLACES 313
research method in economic geography. Storper, M. 1989. The transition to flexible

The Professional Geographer 44:180-89. specialization in industry: External econo-
Scott, A. 1986. High tech industry and territo- mies, the division of labor and the crossing
rial development: The rise of the Orange of industrial divides. Cambridge Journal of
County complex, 1955-1984. Urban Geog- Economics 13:273-305.
raphy 7:3-45. Storper, M., and Harrison, B. 1991. Flexibility,
. 1988a. Flexible production systems hierarchy and regional development: The
and regional development: The rise of new changing structure of industrial production
industrial space in North America and systems and their forms of governance in the
Western Europe. International Journal of
1980s. Research Policy 20:407-22.
Urban and Regional Research 12:171-86.
Storper, M., and Walker, R. 1989. The
. 1988b. New industrial space. London:
Pion. capitalist imperative: Territory, technology
Scott, A., and Paul, A. 1990. Collective order and industrial growth. New York: Basil
and economic coordination in industrial Blackwell.
agglomerations: The technopoles of South-U.S. Department of Commerce, Bureau of

ern California. Environment and PlanningEconomic Analysis. 1973, 1990. Country
C: Government and Policy 8:179-93. business patterns. Washington, D.C.: Gov-
Sforzi, F. 1989. The geography of industrial ernment Printing Office.
Yin, R. 1984. Case study research: Design and
district in Italy. In Small firms and indus-
methods. Applied Social Research Methods
trial districts in Italy, ed. E. Goodman and
J. Bamford, 153-73. London: Routledge. Series 5. Beverly Hills, Calif.: Sage.

Taylor & Francis, Ltd. Economic Geography: This Content Downloaded From 169.226.92.8 On Sun, 10 May 2020 22:31:46 UTC

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Taylor & Francis, Ltd. Economic Geography: This Content Downloaded From 169.226.92.8 On Sun, 10 May 2020 22:31:46 UTC

Uploaded by

Copyright:

Available Formats

Sticky Places in Slippery Space: A Typology of Industrial Districts

Author(s): Ann Markusen

Abstract: As advances in transportation and information obliterate distance, cities

Key words: industrial districts, regional growth.

The Puzzle of Stickiness in an capital and labor. Movement is, of course,

niewicz 1994; Harrison 1994; Sayer 1989). ti

the differences among sticky places and

Marshallian and Italianate

Hypothesized Features of New Industrial District Types

Marshallian industrial districts

Italianate variant (in addition to the above)

Satellite industrial platforms

State-anchored industrial districts

Selected U.S. Metropolitan Employment Growth Rates, 1970-1990

Colorado Springs, Colo. 228 104 24 261 60 214

Source: U.S. Department of Commerce, Bureau of

which enables that any

Despite the often extravagant claims of diversified developmental process where

timeliness, and inventory control, and it trade tend to be tied up as retained

It is not as if branch operations, local government efforts and services

unskilled positions will also be filled from Politics in state-anchored industrial

within the region. Improving cooperative alliances might be as important to locali-

Christopherson, S. 1993. Market rules and of the Institute of British Geographers

research method in economic geography. Storper, M. 1989. The transition to flexible

agglomerations: The technopoles of South-U.S. Department of Commerce, Bureau of

You might also like