Global production networks and theanalysis of economic development

Jeffrey Henderson Peter Dicken Martin HessNeil Coe and Henry Wai-Chung Yeung

University of Manchester and National University of Singapore

ABSTRACT

This article outlines a framework for the analysis of economic integrationand its relation to the asymmetries of economic and social developmentConsciously breaking with state-centric forms of social science it arguesfor a research agenda that is more adequate to the exigencies and conse-quences of globalization than has traditionally been the case in lsquodevel-opment studiesrsquo Drawing on earlier attempts to analyse the cross-borderactivities of rms their spatial congurations and developmental conse-quences the article moves beyond these by proposing the framework ofthe lsquoglobal production networkrsquo (GPN) It explores the conceptual elementsinvolved in this framework in some detail and then turns to sketch a stylized example of a GPN The article concludes with a brief indication of the benets that could be delivered by research informed by GPN analysis

KEYWORDS

Globalization economic development business networks institutionsembeddedness

The analysis of economic development has been bedevilled by a seriesof analytic disjunctions that have resulted in work either at macro ormeso levels of abstraction or where empirical investigations have probedmicro level processes the larger analytic picture has often been absentmerely implicit or at best weakly developed While there are notableexceptions to this general rule (for instance Armstrong and McGee 1985)behind it lies half a century and more of scholarship in developmenteconomics (irrespective of its paradigmatic stripe) and in the politicaleconomy and sociology of development1 What is more from the begin-nings of lsquodependencyrsquo approaches to development in the 1940s throughto debates over the respective roles of states and markets in the East

Review of International Political Economy 93 August 2002 436ndash464

Review of International Political EconomyISSN 0969-2290 printISSN 1466-4526 online copy 2002 Taylor amp Francis Ltd

httpwwwtandfcoukDOI 10108009692290210150842

Asian lsquomiraclersquo and its recent demise the central agent in developmenthas often been perceived as the state whether the assessment of its rolehas been positive or negative2 Although the developmental signicanceof labour gender and other social movements as well as internationalagencies such as the IMF and World Bank have gured in radicalanalyses the analytic space given to development actors other than thesehas been limitedNowhere is this relative absence more obvious than with regard to

the rm Although there is a long history of work on foreign investmentand development (summarized for instance in Jenkins 1987 and Dicken1998) this has tended to deal largely with the role of transnational cor-porations (TNCs) and has relied primarily on secondary data for itsempirical bases Little of it has probed the organizational dynamics ofTNC subsidiaries as they emerge evolve and impact on particulareconomies and even less of it has dealt with domestic rms be theyassociated or not with foreign companies3There is of course a considerable amount of research on rms that

has been conducted by sociologists of work and organization and byspecialists in management studies However this has been largely con-ned to companies in developed economies and the former state-socialistsocieties of Central and Eastern Europe and where it has been conductedby management specialists it has remained outside the social sciencemainstream and thus has largely failed to inuence (or be inuencedby) more general discourses Where work of this nature has been con-ducted in the developing world it has been done largely by feministresearchers and has tended to engage more with gender-related issuesthan with the broader questions of industrial organization and economicdevelopment (see for instance Heyzer 1986 Mitter and Rowbotham1995)A further ndash and given contemporary circumstances perhaps fatal ndash

analytic disjuncture is that research on economic development (as withthe vast majority of social science) has been state-centric in its assump-tions and analyses4 While world-systems theory has provided an analyticframework that promises to move beyond these limitations it is a frame-work that has yet to act as a signicant guide to empirical work oncontemporary problems of development In this context the national statecontinues to be the conventional unit of analysis for the majority of studiesof the world economy However exclusive attention to this level of aggre-gation is becoming les useful in light of the changes occurring in theorganization of economic activities which increasingly tend to slicethrough while still being unevenly contained within state boundariesIndeed Castells has argued that the world is being transformed from

a lsquospace of placesrsquo into a lsquospace of owsrsquo (Castells 2000a 2000b) Moreaccurately perhaps the world is now constituted by both a space of places

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

437

and a space of ows and thus a key issue has become the nature of thedialectical relation between these spaces and the consequences of thatrelationIn order to understand the dynamics of development in a given place

then we must comprehend how places are being transformed by owsof capital labour knowledge power etc and how at the same timeplaces (or more specically their institutional and social fabrics) are trans-forming those ows as they locate in place-specic domains Globaliza-tion (for that is the shorthand for our concerns) has undercut the valid-ity of traditional state-centred forms of social science and with that the agendas that hitherto have guided the vast majority of research oneconomic and social development Investigations adequate to the studyof globalization and its consequences demand of social scientists theelaboration of analytic frameworks and research programmes that simul-taneously foreground the dynamics of uneven development transna-tionally nationally and sub-nationally Such investigations require us tofocus on the ows and the places and their dialectical connections as thesearise and are realized in the developed and developing worlds alikeAdditionally if the object of our endeavours is the possibilities for econ-omic development and prosperity then we should recognize that in orderto speak authoritatively on these issues we need to study what rmsdo where they do it why they do it why they are allowed to do it andhow they organize the doing of it across different geographic scalesIn this article we outline an analytic framework which we believe

helps us to understand some of these processes more effectively Theframework we propose is that of the lsquoglobal production networkrsquo (GPN)While the GPN is not advanced as a totalizing framework capable ofgrasping the myriad complexities of economic globalization we believethat it is capable of delivering a better analytic purchase on the changinginternational distribution of production and consumption ndash and theviability of different development strategies to which they relate ndash thanhas previously been possibleWe begin with some brief critical reections on the most relevant pre-

cursors to our work We then outline the conceptual elements of the GPNand in so doing highlight the reasons for its analytic superiority overcompeting frameworks Penultimately we present a stylised example ofa GPN and conclude with a brief comment on the benets that GPNresearch could deliver

CRITICAL REFLECTIONS ON RELATED APPROACHES

Over the past 20 years or so a plethora of studies has emerged usingsome variant or another on the concept of chains or networks5 The resultis a considerable degree of confusion in the use and meaning of the

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

438

terminologies employed (cf Sturgeon 2001) Although the approachesoften overlap with one another they derive from different intellectualdomains and therefore carry with them different kind of intellec-tual lsquobaggagersquo One difference between these approaches is between thosethat stem from the business-managerial literature and those that haveevolved within an economic-developmental framework A second dif-ference is between those that employ a lsquochainrsquo metaphor and those thatadopt a lsquonetworkrsquo perspective (although the distinction is not alwaysclear-cut)

Chain concepts

The value chain or value-adding chain is an old-established concept in indus-trial economics and in the business studies literature It has been usedmost prominently by Michael Porter (1985 1990) and has achieved verywide currency in the management community Like all uses of the chainmetaphor its value lies in its emphasis on the sequential and inter-connected structures of economic activities with each link or element inthe chain adding value to the process (value being dened in terms of the pay-off to the business rm) For our purposes Porterrsquos concep-tualization has a limited utility because it is bounded by the rm or inter-rm network and pays no attention to issues of corporate power theinstitutional contexts of ndash and inuences upon ndash rm-based activities orto the territorial arrangements (and their profound economic and socialasymmetries) in which the chains are embedded As a consequence ithas little relevance for the study of economic developmentOf greater importance is the concept of the liegravere which is dened as

a system of agents producing and distributing goods and services forthe satisfaction of a nal demand Developed in the 1970s by Frencheconomists in order to achieve a more structured understanding of econ-omic processes within production and distribution systems (Lenz 199721) the concept stems from a predominantly empirical tradition themain objectives of which are to map commodity ows and to identifythe agents and activities within the liegravere (Raikes et al 2000 404ndash5) Bydoing so hierarchical relationships between the agents can be identiedallowing for a detailed analysis of the dynamics of economic integrationand disintegrationIt is difcult to identify a distinct theoretical core for the liegravere approach

Indeed there is a plurality of theories underlying recent liegravere analysesparticularly those of regulation and convention theory6 Although theliegravere approach focuses on agents within the system as well as on depen-dency and the distribution of power it concentrates mainly on two typesof agent ndash large rms and (national) state institutions ndash and how theirscope of activity is limited by technological constraints Hence the

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

439

spectrum of agents in production networks their role in shaping thesenetworks and thus inuencing development at different scales is onlypartially dealt withBy far the most useful of the chain conceptualizations of econ-

omic activities is Gary Gerefrsquos global commodity chain (GCC) The char-acteristics of the GCC framework have been extensively outlined both in Gerefrsquos own writings (see for example Geref and Korzeniewicz1994 Geref 1995 1999a) and in appraisals by others (see for exampleDicken et al 2001 Czaban and Henderson 1998 Whitley 1996) so thereis no need for recuperation here It is important however to understandthe intellectual lineage of Gerefrsquos GCC concept and the extent to whichit may meet our needsGerefrsquos work is set within the (broadly dened) lsquodependencyrsquo tradi-

tion of analysis In focusing on the dynamics of the global organizationof production however it has a particular afnity with the work in thelate 1970s and 1980s on the emergence of a lsquonew international divisionof labourrsquo and its economic and socio-spatial consequences (cf Froumlbel etal 1980 Henderson and Castells 1987 Henderson 1989) As with thework of Froumlbel and his colleagues Gerefrsquos contribution was an explicitattempt to operationalise some of the world-systems categories for theempirical study of cross-border rm-based transactions and their rela-tion to development (Geref 1995) Unlike their work however it broke with the static (and now empirically redundant) spatial categoriesof the coresemiperipheryperiphery typology and as such was betterable to grasp the reality of the lsquonewrsquo forms of industrial organisationthat had become the objects of scholarly attention during the 1980s and1990sFor Geref and his collaborators global commodity chains consist of

sets of interorganizational networks clustered around one com-modity or product linking households enterprises and states toone another within the world-economy These networks are situa-tionally specic socially constructed and locally integrated under-scoring the social embeddedness of economic organization

(Geref et al 1994 2)

With the exception of trade unions and other NGOs this denitionincorporates most of the elements relevant to the organization of rm andinter-rm networks and their relation to the possibilities for economic and social development However only a few of these elements have beenfollowed through empirically or analytically by Geref his collaboratorsor others who have worked in this vein7 In particular the focus has been overwhelmingly on the governance dimension of GCCs and on a bi-modal distinction between producer-driven and buyer-driven GCCs atthat This distinction however is a crude one and it leads to problems

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

440

First although the rationale for this distinction lies in differentialbarriers to entry into the various product markets (Dicken et al 2001)it is clear that the distinction is intended to refer to sectorally and orga-nizationally specic empirical realities It is not then an ideal-typicalconstructionSecond much of the work from within the GCC tradition has been

concerned with currently existing chains Hardly any of it seeks to re-construct the history of the nature and implications of the chains Thisis an important omission because the social relations embodied in chainsat one point in time impose a path-dependency and constrain the futuretrajectories of chain development For example the institutional contextsand social arrangements of the state-socialist period linger on and cir-cumscribe in important ways the potential for economic and politicaldevelopment in the lsquotransitionalrsquo economies of Eastern Europe (cf Stark1992 Hausner et al 1995 Czaban and Henderson 1998)Third there have been few attempts to understand the signicance of

rm ownership (domestic or foreign and in the latter case by nation-ality) for economic and social development in particular societies Eventhough this lsquosilencersquo may be a product of the GCC schemersquos primaryconcern with buyer-driven chains there is clearly a need to recognizethat the lsquonationalityrsquo of rm ownership may be a key element in economicand social progress8The fourth problematic issue for the GCC framework is the fact that

commodity chains link not only rms in different locations but also thespecic social and institutional contexts at the national (sometimes sub-national) level out of which all rms arise and in which all ndash thoughto varying extents ndash remain embedded The implication of the GCCframework seems to be that rms are principally reexes of the waygiven commodity chains are organized and of the structural require-ments this imposes on their operation in any given location In thisscheme of things rms appear to have little autonomy to develop rela-tively independent strategies (though this seems crucial for the pros-pects for sustained development) Additionally there appears to be littleroom for understanding where national and local differences in labourmarket organization working conditions etc come from In our viewthese issues cannot be effectively theorized unless it is understood thatinter-rm networks link societies which exhibit signicant social andinstitutional variation embody different welfare regimes and have dif-ferent capacities for state economic management in short representdifferent forms of capitalism (cf Boyer and Drache 1996 Whitley 1999Coates 2000)As an emerging theory of development however the GCC perspec-

tive has much to recommend it Not least it has helped to spawn impor-tant empirical work on footwear garments electronics horticulture

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

441

tourism and auto-components for instance and has provided the ana-lytical rationale for what could become new policy initiatives from theInternational Labour Ofce (ILO)9 It carries forward the task of tran-scending the limitations of state-centred forms of analysis and in so doing highlights the restrictions on rm ndash and thus economic and socialndash development that arise from the structure of corporate power embed-ded in the intra and inter-rm networks which circle the globe By help-ing to show that the capacities to generate value are asymmetricallydistributed because of the structure of GCCs the perspective points tothe existence of new forms of lsquodependent developmentrsquo as well as to possible ways of transcending those constraints

Network concepts

A chain maps the vertical sequence of events leading to the deliveryconsumption and maintenance of goods and services ndash recognisingthat various value chains often share common economic actors andare dynamic in that they are reused and recongured on an ongoingbasis ndash while a network highlights the nature and extent of theinter-rm relationships that bind sets of rms into larger economicgroupings

(Sturgeon 2001 10)

A major weakness of the lsquochainrsquo approach is its conceptualization ofproduction and distribution processes as being essentially vertical andlinear In fact such processes are better conceptualized as being highlycomplex network structures in which there are intricate links ndash horizontaldiagonal as well as vertical ndash forming multi-dimensional multi-layeredlattices of economic activity For that reason an explicitly relationalnetwork-focused approach promises to offer a better understanding ofproduction systemsOne such approach is actor-network theory (ANT) which emphasizes the

relationality of both objects and agency in heterogeneous networks (lsquorela-tional materialityrsquo) pointing out that entities in networks are shaped byand can only be understood through their relations and connectivity toother entities (Law 1999 4) For the study of global production networksthis means that space and distance have to be seen not in absoluteEuclidean terms but as lsquospatial eldsrsquo and relational scopes of inuencepower and connectivity (Harvey 1969 Murdoch 1998) Among otherthings this has important implications for the conceptualization of thelsquoglobalrsquo and of lsquoglobalizationrsquo10Another important aspect of ANT is its rejection of articial dualisms

such as the traditional global-local and the structure-agency dichotomiesFinally ANT conceptualizes networks as hybrid collectivities of human

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

442

and non-human agents and thus allows the consideration of the impor-tant technological elements that underlie and inuence economic activi-ties However while ANT offers an interesting methodology that hasbeen adopted already for the study of globalization and production net-works (see for instance Whatmore and Thorne 1997) its contributionto the analysis of economic development is constrained by the fact thatit lacks an appreciation of the structural preconditions and power rela-tions that inevitably shape production networks (Dicken et al 2001 107)One contribution with a direct afnity to our work is Dieter Ernstrsquos

version of the global production network Developed contemporaneouslybut independently of our work11 Ernst conceives of a GPN as a particularkind of organizational innovation namely one that

combine(s) concentrated dispersion of the value chain across rmand national boundaries with a parallel process of integration ofhierarchical layers of network participants

(Ernst and Kim 2001 1)

The fundamental rationale for rms to establish GPNs of this natureis supposedly to access exible specialized suppliers in lower-cost loca-tions The GPN is seen to supersede the transnational corporation as themost effective form of industrial organization a shift that has emergedin response to three constituent processes of globalization namely theascendancy of liberalization policies the rapid up-take of informationand communication technologies and the onset of lsquoglobalrsquo competitionThe empirical evidence used to illustrate this alleged wholesale shift

in industrial organization is anecdotal and almost exclusively drawnfrom the electronics and information technology industries Conse-quently rather than having developed an explanatory category of generalrelevance Ernst has tended to highlight only one particular form ofindustrial organization and one at that which seems to be drawn froma sectorally narrow range Ernstrsquos work is particularly helpful howeverin that he highlights a number of key problems that have hinderedprevious research in this areaFirst he criticizes the tendency to focus narrowly on the role of key

lsquoagshiprsquo rms within GPNs at the expense of attention to networksuppliers that are more than one stage removed from the agshipSecond he notes that in mapping the dispersion of production unitsresearch has often overlooked the wide range of service functions (fromdesign to marketing and beyond) that are crucial to the viability of GPNsThird Ernst notes a pre-occupation with formal RampD and technologytransfers which may preclude an appreciation of the importance of diffu-sion of less codied forms of knowledge Indeed much of Ernstrsquos researchunder the GPN banner has been concerned with the potential for differentforms of knowledge (which he variously terms lsquoembrainedrsquo lsquoembeddedrsquo

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

443

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

and a space of ows and thus a key issue has become the nature of thedialectical relation between these spaces and the consequences of thatrelationIn order to understand the dynamics of development in a given place

then we must comprehend how places are being transformed by owsof capital labour knowledge power etc and how at the same timeplaces (or more specically their institutional and social fabrics) are trans-forming those ows as they locate in place-specic domains Globaliza-tion (for that is the shorthand for our concerns) has undercut the valid-ity of traditional state-centred forms of social science and with that the agendas that hitherto have guided the vast majority of research oneconomic and social development Investigations adequate to the studyof globalization and its consequences demand of social scientists theelaboration of analytic frameworks and research programmes that simul-taneously foreground the dynamics of uneven development transna-tionally nationally and sub-nationally Such investigations require us tofocus on the ows and the places and their dialectical connections as thesearise and are realized in the developed and developing worlds alikeAdditionally if the object of our endeavours is the possibilities for econ-omic development and prosperity then we should recognize that in orderto speak authoritatively on these issues we need to study what rmsdo where they do it why they do it why they are allowed to do it andhow they organize the doing of it across different geographic scalesIn this article we outline an analytic framework which we believe

helps us to understand some of these processes more effectively Theframework we propose is that of the lsquoglobal production networkrsquo (GPN)While the GPN is not advanced as a totalizing framework capable ofgrasping the myriad complexities of economic globalization we believethat it is capable of delivering a better analytic purchase on the changinginternational distribution of production and consumption ndash and theviability of different development strategies to which they relate ndash thanhas previously been possibleWe begin with some brief critical reections on the most relevant pre-

cursors to our work We then outline the conceptual elements of the GPNand in so doing highlight the reasons for its analytic superiority overcompeting frameworks Penultimately we present a stylised example ofa GPN and conclude with a brief comment on the benets that GPNresearch could deliver

CRITICAL REFLECTIONS ON RELATED APPROACHES

Over the past 20 years or so a plethora of studies has emerged usingsome variant or another on the concept of chains or networks5 The resultis a considerable degree of confusion in the use and meaning of the

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

438

terminologies employed (cf Sturgeon 2001) Although the approachesoften overlap with one another they derive from different intellectualdomains and therefore carry with them different kind of intellec-tual lsquobaggagersquo One difference between these approaches is between thosethat stem from the business-managerial literature and those that haveevolved within an economic-developmental framework A second dif-ference is between those that employ a lsquochainrsquo metaphor and those thatadopt a lsquonetworkrsquo perspective (although the distinction is not alwaysclear-cut)

Chain concepts

The value chain or value-adding chain is an old-established concept in indus-trial economics and in the business studies literature It has been usedmost prominently by Michael Porter (1985 1990) and has achieved verywide currency in the management community Like all uses of the chainmetaphor its value lies in its emphasis on the sequential and inter-connected structures of economic activities with each link or element inthe chain adding value to the process (value being dened in terms of the pay-off to the business rm) For our purposes Porterrsquos concep-tualization has a limited utility because it is bounded by the rm or inter-rm network and pays no attention to issues of corporate power theinstitutional contexts of ndash and inuences upon ndash rm-based activities orto the territorial arrangements (and their profound economic and socialasymmetries) in which the chains are embedded As a consequence ithas little relevance for the study of economic developmentOf greater importance is the concept of the liegravere which is dened as

a system of agents producing and distributing goods and services forthe satisfaction of a nal demand Developed in the 1970s by Frencheconomists in order to achieve a more structured understanding of econ-omic processes within production and distribution systems (Lenz 199721) the concept stems from a predominantly empirical tradition themain objectives of which are to map commodity ows and to identifythe agents and activities within the liegravere (Raikes et al 2000 404ndash5) Bydoing so hierarchical relationships between the agents can be identiedallowing for a detailed analysis of the dynamics of economic integrationand disintegrationIt is difcult to identify a distinct theoretical core for the liegravere approach

Indeed there is a plurality of theories underlying recent liegravere analysesparticularly those of regulation and convention theory6 Although theliegravere approach focuses on agents within the system as well as on depen-dency and the distribution of power it concentrates mainly on two typesof agent ndash large rms and (national) state institutions ndash and how theirscope of activity is limited by technological constraints Hence the

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

439

spectrum of agents in production networks their role in shaping thesenetworks and thus inuencing development at different scales is onlypartially dealt withBy far the most useful of the chain conceptualizations of econ-

omic activities is Gary Gerefrsquos global commodity chain (GCC) The char-acteristics of the GCC framework have been extensively outlined both in Gerefrsquos own writings (see for example Geref and Korzeniewicz1994 Geref 1995 1999a) and in appraisals by others (see for exampleDicken et al 2001 Czaban and Henderson 1998 Whitley 1996) so thereis no need for recuperation here It is important however to understandthe intellectual lineage of Gerefrsquos GCC concept and the extent to whichit may meet our needsGerefrsquos work is set within the (broadly dened) lsquodependencyrsquo tradi-

tion of analysis In focusing on the dynamics of the global organizationof production however it has a particular afnity with the work in thelate 1970s and 1980s on the emergence of a lsquonew international divisionof labourrsquo and its economic and socio-spatial consequences (cf Froumlbel etal 1980 Henderson and Castells 1987 Henderson 1989) As with thework of Froumlbel and his colleagues Gerefrsquos contribution was an explicitattempt to operationalise some of the world-systems categories for theempirical study of cross-border rm-based transactions and their rela-tion to development (Geref 1995) Unlike their work however it broke with the static (and now empirically redundant) spatial categoriesof the coresemiperipheryperiphery typology and as such was betterable to grasp the reality of the lsquonewrsquo forms of industrial organisationthat had become the objects of scholarly attention during the 1980s and1990sFor Geref and his collaborators global commodity chains consist of

sets of interorganizational networks clustered around one com-modity or product linking households enterprises and states toone another within the world-economy These networks are situa-tionally specic socially constructed and locally integrated under-scoring the social embeddedness of economic organization

(Geref et al 1994 2)

With the exception of trade unions and other NGOs this denitionincorporates most of the elements relevant to the organization of rm andinter-rm networks and their relation to the possibilities for economic and social development However only a few of these elements have beenfollowed through empirically or analytically by Geref his collaboratorsor others who have worked in this vein7 In particular the focus has been overwhelmingly on the governance dimension of GCCs and on a bi-modal distinction between producer-driven and buyer-driven GCCs atthat This distinction however is a crude one and it leads to problems

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

440

First although the rationale for this distinction lies in differentialbarriers to entry into the various product markets (Dicken et al 2001)it is clear that the distinction is intended to refer to sectorally and orga-nizationally specic empirical realities It is not then an ideal-typicalconstructionSecond much of the work from within the GCC tradition has been

concerned with currently existing chains Hardly any of it seeks to re-construct the history of the nature and implications of the chains Thisis an important omission because the social relations embodied in chainsat one point in time impose a path-dependency and constrain the futuretrajectories of chain development For example the institutional contextsand social arrangements of the state-socialist period linger on and cir-cumscribe in important ways the potential for economic and politicaldevelopment in the lsquotransitionalrsquo economies of Eastern Europe (cf Stark1992 Hausner et al 1995 Czaban and Henderson 1998)Third there have been few attempts to understand the signicance of

rm ownership (domestic or foreign and in the latter case by nation-ality) for economic and social development in particular societies Eventhough this lsquosilencersquo may be a product of the GCC schemersquos primaryconcern with buyer-driven chains there is clearly a need to recognizethat the lsquonationalityrsquo of rm ownership may be a key element in economicand social progress8The fourth problematic issue for the GCC framework is the fact that

commodity chains link not only rms in different locations but also thespecic social and institutional contexts at the national (sometimes sub-national) level out of which all rms arise and in which all ndash thoughto varying extents ndash remain embedded The implication of the GCCframework seems to be that rms are principally reexes of the waygiven commodity chains are organized and of the structural require-ments this imposes on their operation in any given location In thisscheme of things rms appear to have little autonomy to develop rela-tively independent strategies (though this seems crucial for the pros-pects for sustained development) Additionally there appears to be littleroom for understanding where national and local differences in labourmarket organization working conditions etc come from In our viewthese issues cannot be effectively theorized unless it is understood thatinter-rm networks link societies which exhibit signicant social andinstitutional variation embody different welfare regimes and have dif-ferent capacities for state economic management in short representdifferent forms of capitalism (cf Boyer and Drache 1996 Whitley 1999Coates 2000)As an emerging theory of development however the GCC perspec-

tive has much to recommend it Not least it has helped to spawn impor-tant empirical work on footwear garments electronics horticulture

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

441

tourism and auto-components for instance and has provided the ana-lytical rationale for what could become new policy initiatives from theInternational Labour Ofce (ILO)9 It carries forward the task of tran-scending the limitations of state-centred forms of analysis and in so doing highlights the restrictions on rm ndash and thus economic and socialndash development that arise from the structure of corporate power embed-ded in the intra and inter-rm networks which circle the globe By help-ing to show that the capacities to generate value are asymmetricallydistributed because of the structure of GCCs the perspective points tothe existence of new forms of lsquodependent developmentrsquo as well as to possible ways of transcending those constraints

Network concepts

A chain maps the vertical sequence of events leading to the deliveryconsumption and maintenance of goods and services ndash recognisingthat various value chains often share common economic actors andare dynamic in that they are reused and recongured on an ongoingbasis ndash while a network highlights the nature and extent of theinter-rm relationships that bind sets of rms into larger economicgroupings

(Sturgeon 2001 10)

A major weakness of the lsquochainrsquo approach is its conceptualization ofproduction and distribution processes as being essentially vertical andlinear In fact such processes are better conceptualized as being highlycomplex network structures in which there are intricate links ndash horizontaldiagonal as well as vertical ndash forming multi-dimensional multi-layeredlattices of economic activity For that reason an explicitly relationalnetwork-focused approach promises to offer a better understanding ofproduction systemsOne such approach is actor-network theory (ANT) which emphasizes the

relationality of both objects and agency in heterogeneous networks (lsquorela-tional materialityrsquo) pointing out that entities in networks are shaped byand can only be understood through their relations and connectivity toother entities (Law 1999 4) For the study of global production networksthis means that space and distance have to be seen not in absoluteEuclidean terms but as lsquospatial eldsrsquo and relational scopes of inuencepower and connectivity (Harvey 1969 Murdoch 1998) Among otherthings this has important implications for the conceptualization of thelsquoglobalrsquo and of lsquoglobalizationrsquo10Another important aspect of ANT is its rejection of articial dualisms

such as the traditional global-local and the structure-agency dichotomiesFinally ANT conceptualizes networks as hybrid collectivities of human

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

442

and non-human agents and thus allows the consideration of the impor-tant technological elements that underlie and inuence economic activi-ties However while ANT offers an interesting methodology that hasbeen adopted already for the study of globalization and production net-works (see for instance Whatmore and Thorne 1997) its contributionto the analysis of economic development is constrained by the fact thatit lacks an appreciation of the structural preconditions and power rela-tions that inevitably shape production networks (Dicken et al 2001 107)One contribution with a direct afnity to our work is Dieter Ernstrsquos

version of the global production network Developed contemporaneouslybut independently of our work11 Ernst conceives of a GPN as a particularkind of organizational innovation namely one that

combine(s) concentrated dispersion of the value chain across rmand national boundaries with a parallel process of integration ofhierarchical layers of network participants

(Ernst and Kim 2001 1)

The fundamental rationale for rms to establish GPNs of this natureis supposedly to access exible specialized suppliers in lower-cost loca-tions The GPN is seen to supersede the transnational corporation as themost effective form of industrial organization a shift that has emergedin response to three constituent processes of globalization namely theascendancy of liberalization policies the rapid up-take of informationand communication technologies and the onset of lsquoglobalrsquo competitionThe empirical evidence used to illustrate this alleged wholesale shift

in industrial organization is anecdotal and almost exclusively drawnfrom the electronics and information technology industries Conse-quently rather than having developed an explanatory category of generalrelevance Ernst has tended to highlight only one particular form ofindustrial organization and one at that which seems to be drawn froma sectorally narrow range Ernstrsquos work is particularly helpful howeverin that he highlights a number of key problems that have hinderedprevious research in this areaFirst he criticizes the tendency to focus narrowly on the role of key

lsquoagshiprsquo rms within GPNs at the expense of attention to networksuppliers that are more than one stage removed from the agshipSecond he notes that in mapping the dispersion of production unitsresearch has often overlooked the wide range of service functions (fromdesign to marketing and beyond) that are crucial to the viability of GPNsThird Ernst notes a pre-occupation with formal RampD and technologytransfers which may preclude an appreciation of the importance of diffu-sion of less codied forms of knowledge Indeed much of Ernstrsquos researchunder the GPN banner has been concerned with the potential for differentforms of knowledge (which he variously terms lsquoembrainedrsquo lsquoembeddedrsquo

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

443

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

terminologies employed (cf Sturgeon 2001) Although the approachesoften overlap with one another they derive from different intellectualdomains and therefore carry with them different kind of intellec-tual lsquobaggagersquo One difference between these approaches is between thosethat stem from the business-managerial literature and those that haveevolved within an economic-developmental framework A second dif-ference is between those that employ a lsquochainrsquo metaphor and those thatadopt a lsquonetworkrsquo perspective (although the distinction is not alwaysclear-cut)

Chain concepts

The value chain or value-adding chain is an old-established concept in indus-trial economics and in the business studies literature It has been usedmost prominently by Michael Porter (1985 1990) and has achieved verywide currency in the management community Like all uses of the chainmetaphor its value lies in its emphasis on the sequential and inter-connected structures of economic activities with each link or element inthe chain adding value to the process (value being dened in terms of the pay-off to the business rm) For our purposes Porterrsquos concep-tualization has a limited utility because it is bounded by the rm or inter-rm network and pays no attention to issues of corporate power theinstitutional contexts of ndash and inuences upon ndash rm-based activities orto the territorial arrangements (and their profound economic and socialasymmetries) in which the chains are embedded As a consequence ithas little relevance for the study of economic developmentOf greater importance is the concept of the liegravere which is dened as

a system of agents producing and distributing goods and services forthe satisfaction of a nal demand Developed in the 1970s by Frencheconomists in order to achieve a more structured understanding of econ-omic processes within production and distribution systems (Lenz 199721) the concept stems from a predominantly empirical tradition themain objectives of which are to map commodity ows and to identifythe agents and activities within the liegravere (Raikes et al 2000 404ndash5) Bydoing so hierarchical relationships between the agents can be identiedallowing for a detailed analysis of the dynamics of economic integrationand disintegrationIt is difcult to identify a distinct theoretical core for the liegravere approach

Indeed there is a plurality of theories underlying recent liegravere analysesparticularly those of regulation and convention theory6 Although theliegravere approach focuses on agents within the system as well as on depen-dency and the distribution of power it concentrates mainly on two typesof agent ndash large rms and (national) state institutions ndash and how theirscope of activity is limited by technological constraints Hence the

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

439

spectrum of agents in production networks their role in shaping thesenetworks and thus inuencing development at different scales is onlypartially dealt withBy far the most useful of the chain conceptualizations of econ-

omic activities is Gary Gerefrsquos global commodity chain (GCC) The char-acteristics of the GCC framework have been extensively outlined both in Gerefrsquos own writings (see for example Geref and Korzeniewicz1994 Geref 1995 1999a) and in appraisals by others (see for exampleDicken et al 2001 Czaban and Henderson 1998 Whitley 1996) so thereis no need for recuperation here It is important however to understandthe intellectual lineage of Gerefrsquos GCC concept and the extent to whichit may meet our needsGerefrsquos work is set within the (broadly dened) lsquodependencyrsquo tradi-

tion of analysis In focusing on the dynamics of the global organizationof production however it has a particular afnity with the work in thelate 1970s and 1980s on the emergence of a lsquonew international divisionof labourrsquo and its economic and socio-spatial consequences (cf Froumlbel etal 1980 Henderson and Castells 1987 Henderson 1989) As with thework of Froumlbel and his colleagues Gerefrsquos contribution was an explicitattempt to operationalise some of the world-systems categories for theempirical study of cross-border rm-based transactions and their rela-tion to development (Geref 1995) Unlike their work however it broke with the static (and now empirically redundant) spatial categoriesof the coresemiperipheryperiphery typology and as such was betterable to grasp the reality of the lsquonewrsquo forms of industrial organisationthat had become the objects of scholarly attention during the 1980s and1990sFor Geref and his collaborators global commodity chains consist of

sets of interorganizational networks clustered around one com-modity or product linking households enterprises and states toone another within the world-economy These networks are situa-tionally specic socially constructed and locally integrated under-scoring the social embeddedness of economic organization

(Geref et al 1994 2)

With the exception of trade unions and other NGOs this denitionincorporates most of the elements relevant to the organization of rm andinter-rm networks and their relation to the possibilities for economic and social development However only a few of these elements have beenfollowed through empirically or analytically by Geref his collaboratorsor others who have worked in this vein7 In particular the focus has been overwhelmingly on the governance dimension of GCCs and on a bi-modal distinction between producer-driven and buyer-driven GCCs atthat This distinction however is a crude one and it leads to problems

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

440

First although the rationale for this distinction lies in differentialbarriers to entry into the various product markets (Dicken et al 2001)it is clear that the distinction is intended to refer to sectorally and orga-nizationally specic empirical realities It is not then an ideal-typicalconstructionSecond much of the work from within the GCC tradition has been

concerned with currently existing chains Hardly any of it seeks to re-construct the history of the nature and implications of the chains Thisis an important omission because the social relations embodied in chainsat one point in time impose a path-dependency and constrain the futuretrajectories of chain development For example the institutional contextsand social arrangements of the state-socialist period linger on and cir-cumscribe in important ways the potential for economic and politicaldevelopment in the lsquotransitionalrsquo economies of Eastern Europe (cf Stark1992 Hausner et al 1995 Czaban and Henderson 1998)Third there have been few attempts to understand the signicance of

rm ownership (domestic or foreign and in the latter case by nation-ality) for economic and social development in particular societies Eventhough this lsquosilencersquo may be a product of the GCC schemersquos primaryconcern with buyer-driven chains there is clearly a need to recognizethat the lsquonationalityrsquo of rm ownership may be a key element in economicand social progress8The fourth problematic issue for the GCC framework is the fact that

commodity chains link not only rms in different locations but also thespecic social and institutional contexts at the national (sometimes sub-national) level out of which all rms arise and in which all ndash thoughto varying extents ndash remain embedded The implication of the GCCframework seems to be that rms are principally reexes of the waygiven commodity chains are organized and of the structural require-ments this imposes on their operation in any given location In thisscheme of things rms appear to have little autonomy to develop rela-tively independent strategies (though this seems crucial for the pros-pects for sustained development) Additionally there appears to be littleroom for understanding where national and local differences in labourmarket organization working conditions etc come from In our viewthese issues cannot be effectively theorized unless it is understood thatinter-rm networks link societies which exhibit signicant social andinstitutional variation embody different welfare regimes and have dif-ferent capacities for state economic management in short representdifferent forms of capitalism (cf Boyer and Drache 1996 Whitley 1999Coates 2000)As an emerging theory of development however the GCC perspec-

tive has much to recommend it Not least it has helped to spawn impor-tant empirical work on footwear garments electronics horticulture

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

441

tourism and auto-components for instance and has provided the ana-lytical rationale for what could become new policy initiatives from theInternational Labour Ofce (ILO)9 It carries forward the task of tran-scending the limitations of state-centred forms of analysis and in so doing highlights the restrictions on rm ndash and thus economic and socialndash development that arise from the structure of corporate power embed-ded in the intra and inter-rm networks which circle the globe By help-ing to show that the capacities to generate value are asymmetricallydistributed because of the structure of GCCs the perspective points tothe existence of new forms of lsquodependent developmentrsquo as well as to possible ways of transcending those constraints

Network concepts

A chain maps the vertical sequence of events leading to the deliveryconsumption and maintenance of goods and services ndash recognisingthat various value chains often share common economic actors andare dynamic in that they are reused and recongured on an ongoingbasis ndash while a network highlights the nature and extent of theinter-rm relationships that bind sets of rms into larger economicgroupings

(Sturgeon 2001 10)

A major weakness of the lsquochainrsquo approach is its conceptualization ofproduction and distribution processes as being essentially vertical andlinear In fact such processes are better conceptualized as being highlycomplex network structures in which there are intricate links ndash horizontaldiagonal as well as vertical ndash forming multi-dimensional multi-layeredlattices of economic activity For that reason an explicitly relationalnetwork-focused approach promises to offer a better understanding ofproduction systemsOne such approach is actor-network theory (ANT) which emphasizes the

relationality of both objects and agency in heterogeneous networks (lsquorela-tional materialityrsquo) pointing out that entities in networks are shaped byand can only be understood through their relations and connectivity toother entities (Law 1999 4) For the study of global production networksthis means that space and distance have to be seen not in absoluteEuclidean terms but as lsquospatial eldsrsquo and relational scopes of inuencepower and connectivity (Harvey 1969 Murdoch 1998) Among otherthings this has important implications for the conceptualization of thelsquoglobalrsquo and of lsquoglobalizationrsquo10Another important aspect of ANT is its rejection of articial dualisms

such as the traditional global-local and the structure-agency dichotomiesFinally ANT conceptualizes networks as hybrid collectivities of human

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

442

and non-human agents and thus allows the consideration of the impor-tant technological elements that underlie and inuence economic activi-ties However while ANT offers an interesting methodology that hasbeen adopted already for the study of globalization and production net-works (see for instance Whatmore and Thorne 1997) its contributionto the analysis of economic development is constrained by the fact thatit lacks an appreciation of the structural preconditions and power rela-tions that inevitably shape production networks (Dicken et al 2001 107)One contribution with a direct afnity to our work is Dieter Ernstrsquos

version of the global production network Developed contemporaneouslybut independently of our work11 Ernst conceives of a GPN as a particularkind of organizational innovation namely one that

combine(s) concentrated dispersion of the value chain across rmand national boundaries with a parallel process of integration ofhierarchical layers of network participants

(Ernst and Kim 2001 1)

The fundamental rationale for rms to establish GPNs of this natureis supposedly to access exible specialized suppliers in lower-cost loca-tions The GPN is seen to supersede the transnational corporation as themost effective form of industrial organization a shift that has emergedin response to three constituent processes of globalization namely theascendancy of liberalization policies the rapid up-take of informationand communication technologies and the onset of lsquoglobalrsquo competitionThe empirical evidence used to illustrate this alleged wholesale shift

in industrial organization is anecdotal and almost exclusively drawnfrom the electronics and information technology industries Conse-quently rather than having developed an explanatory category of generalrelevance Ernst has tended to highlight only one particular form ofindustrial organization and one at that which seems to be drawn froma sectorally narrow range Ernstrsquos work is particularly helpful howeverin that he highlights a number of key problems that have hinderedprevious research in this areaFirst he criticizes the tendency to focus narrowly on the role of key

lsquoagshiprsquo rms within GPNs at the expense of attention to networksuppliers that are more than one stage removed from the agshipSecond he notes that in mapping the dispersion of production unitsresearch has often overlooked the wide range of service functions (fromdesign to marketing and beyond) that are crucial to the viability of GPNsThird Ernst notes a pre-occupation with formal RampD and technologytransfers which may preclude an appreciation of the importance of diffu-sion of less codied forms of knowledge Indeed much of Ernstrsquos researchunder the GPN banner has been concerned with the potential for differentforms of knowledge (which he variously terms lsquoembrainedrsquo lsquoembeddedrsquo

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

443

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

spectrum of agents in production networks their role in shaping thesenetworks and thus inuencing development at different scales is onlypartially dealt withBy far the most useful of the chain conceptualizations of econ-

omic activities is Gary Gerefrsquos global commodity chain (GCC) The char-acteristics of the GCC framework have been extensively outlined both in Gerefrsquos own writings (see for example Geref and Korzeniewicz1994 Geref 1995 1999a) and in appraisals by others (see for exampleDicken et al 2001 Czaban and Henderson 1998 Whitley 1996) so thereis no need for recuperation here It is important however to understandthe intellectual lineage of Gerefrsquos GCC concept and the extent to whichit may meet our needsGerefrsquos work is set within the (broadly dened) lsquodependencyrsquo tradi-

tion of analysis In focusing on the dynamics of the global organizationof production however it has a particular afnity with the work in thelate 1970s and 1980s on the emergence of a lsquonew international divisionof labourrsquo and its economic and socio-spatial consequences (cf Froumlbel etal 1980 Henderson and Castells 1987 Henderson 1989) As with thework of Froumlbel and his colleagues Gerefrsquos contribution was an explicitattempt to operationalise some of the world-systems categories for theempirical study of cross-border rm-based transactions and their rela-tion to development (Geref 1995) Unlike their work however it broke with the static (and now empirically redundant) spatial categoriesof the coresemiperipheryperiphery typology and as such was betterable to grasp the reality of the lsquonewrsquo forms of industrial organisationthat had become the objects of scholarly attention during the 1980s and1990sFor Geref and his collaborators global commodity chains consist of

sets of interorganizational networks clustered around one com-modity or product linking households enterprises and states toone another within the world-economy These networks are situa-tionally specic socially constructed and locally integrated under-scoring the social embeddedness of economic organization

(Geref et al 1994 2)

With the exception of trade unions and other NGOs this denitionincorporates most of the elements relevant to the organization of rm andinter-rm networks and their relation to the possibilities for economic and social development However only a few of these elements have beenfollowed through empirically or analytically by Geref his collaboratorsor others who have worked in this vein7 In particular the focus has been overwhelmingly on the governance dimension of GCCs and on a bi-modal distinction between producer-driven and buyer-driven GCCs atthat This distinction however is a crude one and it leads to problems

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

440

First although the rationale for this distinction lies in differentialbarriers to entry into the various product markets (Dicken et al 2001)it is clear that the distinction is intended to refer to sectorally and orga-nizationally specic empirical realities It is not then an ideal-typicalconstructionSecond much of the work from within the GCC tradition has been

concerned with currently existing chains Hardly any of it seeks to re-construct the history of the nature and implications of the chains Thisis an important omission because the social relations embodied in chainsat one point in time impose a path-dependency and constrain the futuretrajectories of chain development For example the institutional contextsand social arrangements of the state-socialist period linger on and cir-cumscribe in important ways the potential for economic and politicaldevelopment in the lsquotransitionalrsquo economies of Eastern Europe (cf Stark1992 Hausner et al 1995 Czaban and Henderson 1998)Third there have been few attempts to understand the signicance of

rm ownership (domestic or foreign and in the latter case by nation-ality) for economic and social development in particular societies Eventhough this lsquosilencersquo may be a product of the GCC schemersquos primaryconcern with buyer-driven chains there is clearly a need to recognizethat the lsquonationalityrsquo of rm ownership may be a key element in economicand social progress8The fourth problematic issue for the GCC framework is the fact that

commodity chains link not only rms in different locations but also thespecic social and institutional contexts at the national (sometimes sub-national) level out of which all rms arise and in which all ndash thoughto varying extents ndash remain embedded The implication of the GCCframework seems to be that rms are principally reexes of the waygiven commodity chains are organized and of the structural require-ments this imposes on their operation in any given location In thisscheme of things rms appear to have little autonomy to develop rela-tively independent strategies (though this seems crucial for the pros-pects for sustained development) Additionally there appears to be littleroom for understanding where national and local differences in labourmarket organization working conditions etc come from In our viewthese issues cannot be effectively theorized unless it is understood thatinter-rm networks link societies which exhibit signicant social andinstitutional variation embody different welfare regimes and have dif-ferent capacities for state economic management in short representdifferent forms of capitalism (cf Boyer and Drache 1996 Whitley 1999Coates 2000)As an emerging theory of development however the GCC perspec-

tive has much to recommend it Not least it has helped to spawn impor-tant empirical work on footwear garments electronics horticulture

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

441

tourism and auto-components for instance and has provided the ana-lytical rationale for what could become new policy initiatives from theInternational Labour Ofce (ILO)9 It carries forward the task of tran-scending the limitations of state-centred forms of analysis and in so doing highlights the restrictions on rm ndash and thus economic and socialndash development that arise from the structure of corporate power embed-ded in the intra and inter-rm networks which circle the globe By help-ing to show that the capacities to generate value are asymmetricallydistributed because of the structure of GCCs the perspective points tothe existence of new forms of lsquodependent developmentrsquo as well as to possible ways of transcending those constraints

Network concepts

A chain maps the vertical sequence of events leading to the deliveryconsumption and maintenance of goods and services ndash recognisingthat various value chains often share common economic actors andare dynamic in that they are reused and recongured on an ongoingbasis ndash while a network highlights the nature and extent of theinter-rm relationships that bind sets of rms into larger economicgroupings

(Sturgeon 2001 10)

A major weakness of the lsquochainrsquo approach is its conceptualization ofproduction and distribution processes as being essentially vertical andlinear In fact such processes are better conceptualized as being highlycomplex network structures in which there are intricate links ndash horizontaldiagonal as well as vertical ndash forming multi-dimensional multi-layeredlattices of economic activity For that reason an explicitly relationalnetwork-focused approach promises to offer a better understanding ofproduction systemsOne such approach is actor-network theory (ANT) which emphasizes the

relationality of both objects and agency in heterogeneous networks (lsquorela-tional materialityrsquo) pointing out that entities in networks are shaped byand can only be understood through their relations and connectivity toother entities (Law 1999 4) For the study of global production networksthis means that space and distance have to be seen not in absoluteEuclidean terms but as lsquospatial eldsrsquo and relational scopes of inuencepower and connectivity (Harvey 1969 Murdoch 1998) Among otherthings this has important implications for the conceptualization of thelsquoglobalrsquo and of lsquoglobalizationrsquo10Another important aspect of ANT is its rejection of articial dualisms

such as the traditional global-local and the structure-agency dichotomiesFinally ANT conceptualizes networks as hybrid collectivities of human

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

442

and non-human agents and thus allows the consideration of the impor-tant technological elements that underlie and inuence economic activi-ties However while ANT offers an interesting methodology that hasbeen adopted already for the study of globalization and production net-works (see for instance Whatmore and Thorne 1997) its contributionto the analysis of economic development is constrained by the fact thatit lacks an appreciation of the structural preconditions and power rela-tions that inevitably shape production networks (Dicken et al 2001 107)One contribution with a direct afnity to our work is Dieter Ernstrsquos

version of the global production network Developed contemporaneouslybut independently of our work11 Ernst conceives of a GPN as a particularkind of organizational innovation namely one that

combine(s) concentrated dispersion of the value chain across rmand national boundaries with a parallel process of integration ofhierarchical layers of network participants

(Ernst and Kim 2001 1)

The fundamental rationale for rms to establish GPNs of this natureis supposedly to access exible specialized suppliers in lower-cost loca-tions The GPN is seen to supersede the transnational corporation as themost effective form of industrial organization a shift that has emergedin response to three constituent processes of globalization namely theascendancy of liberalization policies the rapid up-take of informationand communication technologies and the onset of lsquoglobalrsquo competitionThe empirical evidence used to illustrate this alleged wholesale shift

in industrial organization is anecdotal and almost exclusively drawnfrom the electronics and information technology industries Conse-quently rather than having developed an explanatory category of generalrelevance Ernst has tended to highlight only one particular form ofindustrial organization and one at that which seems to be drawn froma sectorally narrow range Ernstrsquos work is particularly helpful howeverin that he highlights a number of key problems that have hinderedprevious research in this areaFirst he criticizes the tendency to focus narrowly on the role of key

lsquoagshiprsquo rms within GPNs at the expense of attention to networksuppliers that are more than one stage removed from the agshipSecond he notes that in mapping the dispersion of production unitsresearch has often overlooked the wide range of service functions (fromdesign to marketing and beyond) that are crucial to the viability of GPNsThird Ernst notes a pre-occupation with formal RampD and technologytransfers which may preclude an appreciation of the importance of diffu-sion of less codied forms of knowledge Indeed much of Ernstrsquos researchunder the GPN banner has been concerned with the potential for differentforms of knowledge (which he variously terms lsquoembrainedrsquo lsquoembeddedrsquo

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

443

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

First although the rationale for this distinction lies in differentialbarriers to entry into the various product markets (Dicken et al 2001)it is clear that the distinction is intended to refer to sectorally and orga-nizationally specic empirical realities It is not then an ideal-typicalconstructionSecond much of the work from within the GCC tradition has been

concerned with currently existing chains Hardly any of it seeks to re-construct the history of the nature and implications of the chains Thisis an important omission because the social relations embodied in chainsat one point in time impose a path-dependency and constrain the futuretrajectories of chain development For example the institutional contextsand social arrangements of the state-socialist period linger on and cir-cumscribe in important ways the potential for economic and politicaldevelopment in the lsquotransitionalrsquo economies of Eastern Europe (cf Stark1992 Hausner et al 1995 Czaban and Henderson 1998)Third there have been few attempts to understand the signicance of

rm ownership (domestic or foreign and in the latter case by nation-ality) for economic and social development in particular societies Eventhough this lsquosilencersquo may be a product of the GCC schemersquos primaryconcern with buyer-driven chains there is clearly a need to recognizethat the lsquonationalityrsquo of rm ownership may be a key element in economicand social progress8The fourth problematic issue for the GCC framework is the fact that

commodity chains link not only rms in different locations but also thespecic social and institutional contexts at the national (sometimes sub-national) level out of which all rms arise and in which all ndash thoughto varying extents ndash remain embedded The implication of the GCCframework seems to be that rms are principally reexes of the waygiven commodity chains are organized and of the structural require-ments this imposes on their operation in any given location In thisscheme of things rms appear to have little autonomy to develop rela-tively independent strategies (though this seems crucial for the pros-pects for sustained development) Additionally there appears to be littleroom for understanding where national and local differences in labourmarket organization working conditions etc come from In our viewthese issues cannot be effectively theorized unless it is understood thatinter-rm networks link societies which exhibit signicant social andinstitutional variation embody different welfare regimes and have dif-ferent capacities for state economic management in short representdifferent forms of capitalism (cf Boyer and Drache 1996 Whitley 1999Coates 2000)As an emerging theory of development however the GCC perspec-

tive has much to recommend it Not least it has helped to spawn impor-tant empirical work on footwear garments electronics horticulture

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

441

tourism and auto-components for instance and has provided the ana-lytical rationale for what could become new policy initiatives from theInternational Labour Ofce (ILO)9 It carries forward the task of tran-scending the limitations of state-centred forms of analysis and in so doing highlights the restrictions on rm ndash and thus economic and socialndash development that arise from the structure of corporate power embed-ded in the intra and inter-rm networks which circle the globe By help-ing to show that the capacities to generate value are asymmetricallydistributed because of the structure of GCCs the perspective points tothe existence of new forms of lsquodependent developmentrsquo as well as to possible ways of transcending those constraints

Network concepts

A chain maps the vertical sequence of events leading to the deliveryconsumption and maintenance of goods and services ndash recognisingthat various value chains often share common economic actors andare dynamic in that they are reused and recongured on an ongoingbasis ndash while a network highlights the nature and extent of theinter-rm relationships that bind sets of rms into larger economicgroupings

(Sturgeon 2001 10)

A major weakness of the lsquochainrsquo approach is its conceptualization ofproduction and distribution processes as being essentially vertical andlinear In fact such processes are better conceptualized as being highlycomplex network structures in which there are intricate links ndash horizontaldiagonal as well as vertical ndash forming multi-dimensional multi-layeredlattices of economic activity For that reason an explicitly relationalnetwork-focused approach promises to offer a better understanding ofproduction systemsOne such approach is actor-network theory (ANT) which emphasizes the

relationality of both objects and agency in heterogeneous networks (lsquorela-tional materialityrsquo) pointing out that entities in networks are shaped byand can only be understood through their relations and connectivity toother entities (Law 1999 4) For the study of global production networksthis means that space and distance have to be seen not in absoluteEuclidean terms but as lsquospatial eldsrsquo and relational scopes of inuencepower and connectivity (Harvey 1969 Murdoch 1998) Among otherthings this has important implications for the conceptualization of thelsquoglobalrsquo and of lsquoglobalizationrsquo10Another important aspect of ANT is its rejection of articial dualisms

such as the traditional global-local and the structure-agency dichotomiesFinally ANT conceptualizes networks as hybrid collectivities of human

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

442

and non-human agents and thus allows the consideration of the impor-tant technological elements that underlie and inuence economic activi-ties However while ANT offers an interesting methodology that hasbeen adopted already for the study of globalization and production net-works (see for instance Whatmore and Thorne 1997) its contributionto the analysis of economic development is constrained by the fact thatit lacks an appreciation of the structural preconditions and power rela-tions that inevitably shape production networks (Dicken et al 2001 107)One contribution with a direct afnity to our work is Dieter Ernstrsquos

version of the global production network Developed contemporaneouslybut independently of our work11 Ernst conceives of a GPN as a particularkind of organizational innovation namely one that

combine(s) concentrated dispersion of the value chain across rmand national boundaries with a parallel process of integration ofhierarchical layers of network participants

(Ernst and Kim 2001 1)

The fundamental rationale for rms to establish GPNs of this natureis supposedly to access exible specialized suppliers in lower-cost loca-tions The GPN is seen to supersede the transnational corporation as themost effective form of industrial organization a shift that has emergedin response to three constituent processes of globalization namely theascendancy of liberalization policies the rapid up-take of informationand communication technologies and the onset of lsquoglobalrsquo competitionThe empirical evidence used to illustrate this alleged wholesale shift

in industrial organization is anecdotal and almost exclusively drawnfrom the electronics and information technology industries Conse-quently rather than having developed an explanatory category of generalrelevance Ernst has tended to highlight only one particular form ofindustrial organization and one at that which seems to be drawn froma sectorally narrow range Ernstrsquos work is particularly helpful howeverin that he highlights a number of key problems that have hinderedprevious research in this areaFirst he criticizes the tendency to focus narrowly on the role of key

lsquoagshiprsquo rms within GPNs at the expense of attention to networksuppliers that are more than one stage removed from the agshipSecond he notes that in mapping the dispersion of production unitsresearch has often overlooked the wide range of service functions (fromdesign to marketing and beyond) that are crucial to the viability of GPNsThird Ernst notes a pre-occupation with formal RampD and technologytransfers which may preclude an appreciation of the importance of diffu-sion of less codied forms of knowledge Indeed much of Ernstrsquos researchunder the GPN banner has been concerned with the potential for differentforms of knowledge (which he variously terms lsquoembrainedrsquo lsquoembeddedrsquo

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

443

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

tourism and auto-components for instance and has provided the ana-lytical rationale for what could become new policy initiatives from theInternational Labour Ofce (ILO)9 It carries forward the task of tran-scending the limitations of state-centred forms of analysis and in so doing highlights the restrictions on rm ndash and thus economic and socialndash development that arise from the structure of corporate power embed-ded in the intra and inter-rm networks which circle the globe By help-ing to show that the capacities to generate value are asymmetricallydistributed because of the structure of GCCs the perspective points tothe existence of new forms of lsquodependent developmentrsquo as well as to possible ways of transcending those constraints

Network concepts

A chain maps the vertical sequence of events leading to the deliveryconsumption and maintenance of goods and services ndash recognisingthat various value chains often share common economic actors andare dynamic in that they are reused and recongured on an ongoingbasis ndash while a network highlights the nature and extent of theinter-rm relationships that bind sets of rms into larger economicgroupings

(Sturgeon 2001 10)

A major weakness of the lsquochainrsquo approach is its conceptualization ofproduction and distribution processes as being essentially vertical andlinear In fact such processes are better conceptualized as being highlycomplex network structures in which there are intricate links ndash horizontaldiagonal as well as vertical ndash forming multi-dimensional multi-layeredlattices of economic activity For that reason an explicitly relationalnetwork-focused approach promises to offer a better understanding ofproduction systemsOne such approach is actor-network theory (ANT) which emphasizes the

relationality of both objects and agency in heterogeneous networks (lsquorela-tional materialityrsquo) pointing out that entities in networks are shaped byand can only be understood through their relations and connectivity toother entities (Law 1999 4) For the study of global production networksthis means that space and distance have to be seen not in absoluteEuclidean terms but as lsquospatial eldsrsquo and relational scopes of inuencepower and connectivity (Harvey 1969 Murdoch 1998) Among otherthings this has important implications for the conceptualization of thelsquoglobalrsquo and of lsquoglobalizationrsquo10Another important aspect of ANT is its rejection of articial dualisms

such as the traditional global-local and the structure-agency dichotomiesFinally ANT conceptualizes networks as hybrid collectivities of human

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

442

and non-human agents and thus allows the consideration of the impor-tant technological elements that underlie and inuence economic activi-ties However while ANT offers an interesting methodology that hasbeen adopted already for the study of globalization and production net-works (see for instance Whatmore and Thorne 1997) its contributionto the analysis of economic development is constrained by the fact thatit lacks an appreciation of the structural preconditions and power rela-tions that inevitably shape production networks (Dicken et al 2001 107)One contribution with a direct afnity to our work is Dieter Ernstrsquos

version of the global production network Developed contemporaneouslybut independently of our work11 Ernst conceives of a GPN as a particularkind of organizational innovation namely one that

combine(s) concentrated dispersion of the value chain across rmand national boundaries with a parallel process of integration ofhierarchical layers of network participants

(Ernst and Kim 2001 1)

The fundamental rationale for rms to establish GPNs of this natureis supposedly to access exible specialized suppliers in lower-cost loca-tions The GPN is seen to supersede the transnational corporation as themost effective form of industrial organization a shift that has emergedin response to three constituent processes of globalization namely theascendancy of liberalization policies the rapid up-take of informationand communication technologies and the onset of lsquoglobalrsquo competitionThe empirical evidence used to illustrate this alleged wholesale shift

in industrial organization is anecdotal and almost exclusively drawnfrom the electronics and information technology industries Conse-quently rather than having developed an explanatory category of generalrelevance Ernst has tended to highlight only one particular form ofindustrial organization and one at that which seems to be drawn froma sectorally narrow range Ernstrsquos work is particularly helpful howeverin that he highlights a number of key problems that have hinderedprevious research in this areaFirst he criticizes the tendency to focus narrowly on the role of key

lsquoagshiprsquo rms within GPNs at the expense of attention to networksuppliers that are more than one stage removed from the agshipSecond he notes that in mapping the dispersion of production unitsresearch has often overlooked the wide range of service functions (fromdesign to marketing and beyond) that are crucial to the viability of GPNsThird Ernst notes a pre-occupation with formal RampD and technologytransfers which may preclude an appreciation of the importance of diffu-sion of less codied forms of knowledge Indeed much of Ernstrsquos researchunder the GPN banner has been concerned with the potential for differentforms of knowledge (which he variously terms lsquoembrainedrsquo lsquoembeddedrsquo

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

443

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

and non-human agents and thus allows the consideration of the impor-tant technological elements that underlie and inuence economic activi-ties However while ANT offers an interesting methodology that hasbeen adopted already for the study of globalization and production net-works (see for instance Whatmore and Thorne 1997) its contributionto the analysis of economic development is constrained by the fact thatit lacks an appreciation of the structural preconditions and power rela-tions that inevitably shape production networks (Dicken et al 2001 107)One contribution with a direct afnity to our work is Dieter Ernstrsquos

version of the global production network Developed contemporaneouslybut independently of our work11 Ernst conceives of a GPN as a particularkind of organizational innovation namely one that

combine(s) concentrated dispersion of the value chain across rmand national boundaries with a parallel process of integration ofhierarchical layers of network participants

(Ernst and Kim 2001 1)

The fundamental rationale for rms to establish GPNs of this natureis supposedly to access exible specialized suppliers in lower-cost loca-tions The GPN is seen to supersede the transnational corporation as themost effective form of industrial organization a shift that has emergedin response to three constituent processes of globalization namely theascendancy of liberalization policies the rapid up-take of informationand communication technologies and the onset of lsquoglobalrsquo competitionThe empirical evidence used to illustrate this alleged wholesale shift

in industrial organization is anecdotal and almost exclusively drawnfrom the electronics and information technology industries Conse-quently rather than having developed an explanatory category of generalrelevance Ernst has tended to highlight only one particular form ofindustrial organization and one at that which seems to be drawn froma sectorally narrow range Ernstrsquos work is particularly helpful howeverin that he highlights a number of key problems that have hinderedprevious research in this areaFirst he criticizes the tendency to focus narrowly on the role of key

lsquoagshiprsquo rms within GPNs at the expense of attention to networksuppliers that are more than one stage removed from the agshipSecond he notes that in mapping the dispersion of production unitsresearch has often overlooked the wide range of service functions (fromdesign to marketing and beyond) that are crucial to the viability of GPNsThird Ernst notes a pre-occupation with formal RampD and technologytransfers which may preclude an appreciation of the importance of diffu-sion of less codied forms of knowledge Indeed much of Ernstrsquos researchunder the GPN banner has been concerned with the potential for differentforms of knowledge (which he variously terms lsquoembrainedrsquo lsquoembeddedrsquo

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

443

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

lsquoenculturedrsquo) to be diffused from GPNs in developing country locationsand thereby stimulate local industrial upgrading (see for example Ernst2000)

GLOBAL PRODUCTION NETWORKS

The concept of the global production network (GPN) developed in theremainder of this article draws on many aspects of the work outlinedin the preceding section In particular it builds upon the work of Gerefand his collaborators but takes seriously the criticisms that have beenlevelled against it Concomitantly the framework aims to provide a moregenerally applicable conceptualisation of the GPN than that of ErnstBefore we elaborate the nature of the GPN however we need to explainour preferences for the terms lsquoproductionrsquo rather than lsquocommodityrsquo andlsquonetworkrsquo rather than lsquochainrsquo We also need to indicate our understandingof lsquoglobalrsquoIn contemporary usage the term lsquocommodityrsquo generally connotes stan-

dardized products and with that the xity of their production in timeand space While this remains the reality of some forms of productiveactivity and products (some agriculture some heavy industry and min-erals extraction for instance) it clearly does not capture adequately thepost-fordist forms of activity that characterise many of the industries thatthe GCCs framework for instance was designed to analyse More impor-tantly perhaps our preference for a discourse of lsquoproductionrsquo places theanalytic emphasis on the social processes involved in producing goods and services and reproducing knowledge capital and labour powerNotwithstanding Marxrsquos denitive deconstruction and interrogation ofthe commodity (in Part I of the rst volume of Capital) the discourse of commodities has long been captured by orthodox economics ofwhatever paradigm As a consequence it has transmuted into a reiedlanguage shorn of its social content There is a need therefore to re-focus attention on the social circumstances under which commodities areproduced and consumed and thus avoid the ever-present danger of slip-ping into a perception of commodities as de-humanized building blocksinvolved in the making of other commoditiesThe metaphor of the chain gives the impression of an essentially linear

process of activities that ultimately result in a nal commodity ratherthan one in which the ows of materials semi-nished products designproduction nancial and marketing services are organized verticallyhorizontally and diagonally in complex and dynamic congurationsAdditionally the chain metaphor ndash consistent with a commodity dis-course ndash seems to have difculties incorporating due attention to theissues of the reproduction of labour power etc Furthermore the chainmetaphor works against the possibility of conceiving of the individual

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

444

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

rms incorporated into a production system has having room forautonomous action within that system in spite of the fact that suchautonomy is central to the possibilities for industrial upgrading and thussustained economic development As a consequence of these difcultieswe nd a discourse of networks to be more inclusive empiricallyadequate and thus more analytically fertileAdoption of a network discourse also delivers other potential benets

In particular as long as lsquoproductionrsquo is couched broadly to include inter-mediate and nal markets and as long as the dynamics of power andknowledge between actors and institutions are understood in a multidi-rectional and non-deterministic fashion then the GPN framework allowsfor far greater complexity and geographical variation in producer-consumer relations than the GCC approach for instance has so farachieved Specically it should facilitate our ability to reveal how cer-tain key knowledges lsquocirculatersquo between producers consumers and inter-mediaries rather than moving in a uni-directional manner a key insightof the expanding literature on lsquocommodity culturesrsquo (eg Cook and Crang1996 Jackson 1999) Moreover this approach should also allow morecomplex social geographies to be revealed in the sense that agents in avariety of locations can be seen to combine to inuence the productionprocess12Finally while it is now fashionable to term lsquoglobalrsquo phenomena and

practices that until recently would have been more likely to be termedlsquointernationalrsquo or lsquotransnationalrsquo our adoption of the former term isdriven by our concerns with analytical precision Specically the termslsquointernationalrsquo and lsquotransnationalrsquo derive from essentially state-centricdiscourses Thus while they incorporate notions of cross-border activityof many sorts they do not adequately express the way in which nonplace-specic processes penetrate and transform place-specic ones andvice versa They do not therefore help to deliver the imaginative sensi-bilities necessary to grasp the dialectics of global-local relations that arenow a pre-condition for the analysis of economic globalization and itsasymmetric consequencesThe global production network as proposed here is a conceptual

framework that is capable of grasping the global regional and localeconomic and social dimensions of the processes involved in many(though by no means all) forms of economic globalization13 Productionnetworks ndash the nexus of interconnected functions and operations throughwhich goods and services are produced distributed and consumed ndashhave become both organizationally more complex and also increasinglyglobal in their geographic extent Such networks not only integrate rms(and parts of rms) into structures which blur traditional organizationalboundaries ndash through the development of diverse forms of equity andnon-equity relationships ndash but also integrate national economies (or parts

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

445

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

of such economies) in ways which have enormous implications for their well-being At the same time the precise nature and articula-tion of rm-centred production networks are deeply inuenced by theconcrete socio-political contexts within which they are embedded The process is especially complex because while the latter are essentiallyterritorially specic (primarily though not exclusively at the level of the nation-state) the production networks themselves are not They lsquocut throughrsquo state boundaries in highly differentiated ways inuencedin part by regulatory and non-regulatory barriers and local socio-cultural conditions to create structures which are lsquodiscontinuouslyterritorialrsquo14The GPN framework explicitly recognizes that

l rms governments and other economic actors from different societiessometimes have different priorities vis-agrave-vis protability growtheconomic development etc (as was made clear for instance in thecommentary surrounding the East Asian crisis eg Chang 1998 andHenderson 1999) and consequently the production networkrsquos impli-cations for rm and economic development at each spatial locationcannot be lsquoread-offrsquo from the logic of the networkrsquos organisation andthe distribution of corporate power within it The GPN perspectivein other words accords a degree of relative autonomy to domesticrms governments and other economic actors (eg trade unionswhere relevant) whose actions potentially have signicant implica-tions for the economic and social outcomes of the networks in thelocations they incorporate

l input-output structures within the networks are centrally importantnot least because it is these that constitute the sites where value isgenerated and where the enormous variations in working conditionsthat exist around the world are delivered Consequently any workon intra and inter-rm networks must pay signicant attention tothese structures and their consequences

l an understanding of the lsquoterritorialityrsquo of production networks ndashnamely how they constitute and are re-constituted by the economicsocial and political arrangements of the places they inhabit ndash is centralto an analysis of the prospects for development at the local level

l the distinction between lsquoproducer-driverrsquo and lsquobuyer-drivenrsquo net-works is more uid than Gerefrsquos work allows for with combinationsof both in the same product areas and indeed in some cases (egauto components and consumer electronics) the same sector and

l in some sectors (pharmaceuticals and some electronics for example)technological alliances and licensing agreements are forms of inter-rm association that may have signicant developmental implicationsConsequently they require attention in their own right

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

446

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

Methodologically then the GPN perspective directs attention to

l the networks of rms involved in RampD design production and mar-keting of a given product and how these are organized globally andregionally

l the distribution of corporate power within those networks andchanges therein

l the signicance of labour and the processes of value creation andtransfer and

l the institutions ndash particularly government agencies but also in somecases trade unions employer associations and NGOs ndash that inuencerm strategy in the particular locations absorbed into the productionchain

l the implications of all of these for technological upgrading value-adding and capturing economic prosperity etc for the various rmsand societies absorbed into the chains

More specically the components of the GPN framework can be disag-gregated ndash for purposes of elaboration ndash by reference to Figure 1 Whilewe elaborate these components below it is worth noting here our viewof lsquotechnologyrsquo in the schema While some contributions recognize thecentral role of technological change and information and communica-tion technology (ICT) in shaping and transforming global networks weexclude lsquotechnologyrsquo as a separate category Instead ICT is rather seenas an inherent element of GPNs underlying the development and main-tenance of network connections Technology as one of globalizationrsquosdrivers inuences the processes of value creation in different places aswell as transforming the means by which power is exercised Addition-ally it affects the agentsrsquo possibilities of embedding in and disembeddingfrom particular networks and territoriesA similar view is taken of the notion of spatiality Specic spatial

congurations are an inherent characteristic of all networks each GPNcan be mapped by lsquoplacingrsquo its agents and sketching their mutual connec-tions By the same token every form of embeddedness always has anintrinsic spatial characterThere are however other aspects of spatiality to be considered Firstly

there is the issue of scalarity All GPNs have to be regarded as multi-scalar ranging from the local and regional to the national and globaland back again15 Such multi-scalar networks are built-up and trans-formed over time by a multiplicity of agents with asymmetrical inuenceand power This leads to another important facet of spatiality namelythe boundedness of network-based activities for instance within thepolitical space of the national (or in federal contexts sub-national) stateWhereas business agents are able to transcend political or other borders(cultural for instance) between territories most of the non-business

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

447

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

institutions are bounded ndash and thus restricted ndash by their spatial contextsat different geographic scales This of course has various implicationsfor development especially in terms of the distribution of power andvalue creation and capture

Conceptual categories

There are three principal elements on which the architecture of the GPNframework is raised The rst of these isValue by lsquovaluersquo we mean both Marxian notions of surplus value and

more orthodox ones associated with economic rent Thus we are interestedin the following matters

l The initial creation of value within each of the rms incorporated intoa given GPN The signicant issues here include the conditions underwhich labour power is converted into actual labour through the labourprocess and the possibilities for generating various forms of rent Inthe former the issues of employment skill working conditions andproduction technology are important as well as the circumstancesunder which they are reproduced (hence connecting these issues tobroader social and institutional questions) In the latter (see Kaplinsky1998 Geref 1999b) the issues are whether a given rm can gener-ate rents from (a) an asymmetric access to key product and process

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

448

Firmsndash Ownershipndash ldquoArchitecturerdquo

Institutionsndash Governmentalndash Quasi-governmentalndash Non-governmental

ConfigurationCoordination

Networks (BusinessPolitical)ndash ldquoArchitecturerdquondash Power configurationndash Governance

Sectorsndash Technologiesndash Products Markets

Development

StructuresValue

Valuendash Creationsndash Enhancementndash Capture

Embeddednessndash Territorialndash Network

Powerndash Corporatendash Collectivendash Institutional

Categories

Dimensions

Figure 1 A framework for GPN analysis

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

technologies (lsquotechnological rentsrsquo) (b) from particular organizationaland managerial skills such as lsquojust-in-timersquo production techniques and lsquototal quality controlrsquo etc (lsquoorganizational rentsrsquo) (c) various inter-rmrelationships that may involve the management of production link-ages with other rms the development of strategic alliances or themanagement of relations with clusters of small and medium sizedenterprises (lsquorelational rentsrsquo) or (d) from establishing brand-nameprominence in major markets (lsquobrand rentsrsquo) In certain sectors andcircumstances (e) additional rents may accrue to some rms as a con-sequence of the product scarcities created by protectionist trade poli-cies (lsquotrade-policy rentsrsquo) though this is another issue that connectsquestions of value creation to the institutional contexts (national andinternational in this case) within which rms operate

l The circumstances under which value can be enhanced The issuesinvolved here include (a) the nature and extent of technology trans-fers both from within and without the given production network (b) the extent to which lead and other major rms within the networkengage with supplier and subcontractors to improve the quality andtechnological sophistication of their products (c) as a consequencewhether demands for skill in given labour processes increase overtime (d) whether local rms can begin to create organizational rela-tional and brand rents of their own In all of these cases the nationalinstitutional inuences to which the rms are subject (governmentagencies trade unions employer associations for instance) may bedecisive for the possibilities of value enhancement16

l The possibilities that exist for value to be captured It is one thing forvalue to be created and enhanced in given locations but it may bequite another for it to be captured for the benet of those locationsThe pertinent issues here partly involve (a) matters of governmentpolicy but they also involve (b) questions of rm ownership and (c) the nature of corporate governance in given national contexts Inthe rst case the nature of property rights and thus laws governingownership structures and the repatriation of prots can be importantwhile in the second the extent to which rms are totally foreign ownedtotally domestically owned or involve shared equity as in joint-venturearrangements continues to be decisive as a long tradition in the polit-ical economy of development has argued and recent experience inBritain for instance has underlined17 In the third case the extent to which corporate governance is founded on stakeholder principlesrather than on shareholder dominance (and required by legal statute)can have important consequences for whether value generated in agiven location is retained there and indeed used to the benet of the commonweal18 The issue of value capture then underlines thesignicance of the national form of capitalism ndash and thus matters of

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

449

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

expectations rights and obligations ndash for questions of economic andsocial development

Power the source of power within the GPNs and the ways in whichit is exercised is decisive for value enhancement and capture and thusfor the prospects for development and prosperity19 There are three formsof power that are signicant here

l Corporate power Here we have in mind the extent to which the leadrm in the GPN has the capacity to inuence decisions and resourceallocations ndash vis-agrave-vis other rms in the network ndash decisively andconsistently in its own interests Our adoption of a network discourseimplies a rejection of a zero-sum conception of power in that leadrms rarely if ever have a monopoly on corporate power Ratherwhile power is usually asymmetrically distributed in production net-works lesser rms sometimes (and for contingent reasons) have suf-cient autonomy to develop and exercise their own strategies forupgrading their operations etc Additionally and at least in principlelesser rms incorporated into networks have the possibility of com-bining with other lesser rms to improve their collective situationwithin the GPN (as when SME clusters constituted as industrialdistricts are incorporated into GPNs)20

l Institutional power Our reference here is to the exercise of power by(a) the national and local state (in the latter case where the nationalstate is constituted as a federal polity) (b) international inter-stateagencies ranging from the increasingly integrated European Union on the one hand through to looser confederations such as ASEAN orNAFTA on the other (c) the lsquoBretton Woodsrsquo institutions (InternationalMonetary Fund World Bank) and the World Trade Organisation (d) the various UN agencies (particularly the ILO) and (e) the inter-national credit rating agencies (Moodys Standard and Poor etc) whichexercise a unique form of private institutional power The capacity to exercise power to inuence the investment and other decisions oflead companies and other rms integrated into GPNs is inevitablyasymmetric and varies both within and between these ve categoriesThus with regard to national states some of those in East Asia (par-ticularly South Korea and Taiwan but more recently China) have been perceived in recent decades as being among the most capable of inuencing private companies in the interests of industrializa-tion and development (among an enormous literature see Wade 1990and Henderson 1993) while states as disparate as those of Britain andIndonesia have been far less able to do so21 The power of the inter-state agencies is potentially considerable ndash particularly in the case ofthe EU ndash though elsewhere it remains weakly developed The powerof the Bretton Woods institutions while it can be considerable is

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

450

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

exercised indirectly and impacts on companies workforces and com-munities via the economic and social policies that national govern-ments are obliged to implement The power of the UN agencies is of much less signicance than any of the others in that its inuenceon rms is not merely indirect but it is also only moral and advisoryThe signicance of the credit rating agencies is potentially consider-able both directly for many lead companies and indirectly via theircredit risk assessments of national governments However as yet weknow little of the ways in which their inuence is exercised (but seeSassen 1999)

l Collective power By this form of power we understand the actions ofcollective agents who seek to inuence companies at particular loca-tions in GPNs their respective governments and sometimes inter-national agencies (most recently the IMF and WTO in particular)Examples of such collective agents include trade unions employersassociations and organizations that advance particular economic inter-ests (eg of small businesses) NGOs concerned with human rightsenvironmental issues etc These agencies may be nationally or locallyspecic or they may be internationally organized as with some tradeunions (eg the International Metal Workers) or human rights orga-nizations (eg Amnesty International) In most circumstances wheresuch agencies are engaged they attempt to exercise countervailingpower either directly on particular rms or groups of rms withingiven networks or indirectly on national governments or internationalagencies

Embeddedness GPNs do not only connect rms functionally and terri-torially but also they connect aspects of the social and spatial arrange-ments in which those rms are embedded and which inuence theirstrategies and the values priorities and expectations of managersworkers and communities alike The ways in which the different agentsestablish and perform their connections to others and the specics ofembedding and disembedding processes are to a certain extent basedupon the lsquoheritagersquo and origin of these agents Firms ndash be they TNCs orsmaller local enterprises ndash arise from and continue to be inuenced bythe institutional fabrics and social and cultural contexts of particularforms of capitalism (or in the case of Eastern Europe China etc priorto the 1980s particular forms of state socialism) in their countries oforigin While the nature of education training and labour systems andthe sources and organization of corporate nance are important of partic-ular signicance for rm development priorities and strategies are thenature of state policy and the legal framework (cf Zysman 1983 Hutton1995 Whitley 1999)Local companies that have emerged from particular social and insti-

tutional contexts evolve over time on the bases of trajectories that are

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

451

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

in part a reection of these contexts As many scholars have pointed outwith regard to the former state socialist societies of Eastern Europe thesetrajectories are lsquopath dependentrsquo and thus to some extent historicallyconstrained (for instance Stark 1992 Hausner et al 1995 Czaban andHenderson 1998) While it is important to recognize that such constraintsare not immutable and that their inuence may be waning ndash not leastbecause of globalization ndash it is also important to acknowledge that somelead rms when investing overseas may carry the institutional lsquobaggagersquoof their home bases with them But others might also tend to operate ator near the lowest common denominator that domestic policies and legalframeworks will allow22Among the different dimensions and aspects of embeddedness23 there

are two related forms of rm and network embeddedness that are ofinterest here The rst form territorial deals with the various GPN rmsrsquolsquoanchoringrsquo in different places (from the nation state to the local level)which affects the prospects for the development of these locations Thesecond form network embeddedness refers to the network structure the degree of connectivity within a GPN the stability of its agentsrsquo rela-tions and the importance of the network for the participants Both formsof course are the result of essentially social and spatial processes oflsquoembeddingrsquo

l Territorial embeddeness GPNs do not merely locate in particularplaces They may become embedded there in the sense that theyabsorb and in some cases become constrained by the economic activ-ities and social dynamics that already exist in those places Oneexample here is the way in which the GPNs of particular lead rmsmay take advantage of clusters of small and medium enterprises (withtheir decisively important social networks and local labour markets)that pre-date the establishment of subcontracting or subsidiary oper-ations by such rms Moreover the location of lead rms in particularplaces might generate a new local or regional network of economicand social relations involving existing rms as well as attracting newones Embeddedness then becomes a key element in regional econ-omic growth and in capturing global opportunities (Harrison 1992Amin and Thrift 1994)24 The resulting advantages in terms of valuecreation etc may result in spatial lsquolock-inrsquo for those rms with knock-on implications for other parts of that rmrsquos GPN (see Grabher 1993and Scott 1998) Similarly national and local government policies(training programmes tax advantages etc) may function to embedparticular parts of the GPN in particular cities or regions in order tosupport the formation of new nodes in global networks or what Hein(2000) describes as lsquonew islands of an archipelago economyrsquo But thepositive effects of embeddedness in a particular place cannot be takenfor granted over time For example once a lead rm cuts its ties within

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

452

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

a region (for instance by disinvestment or plant closure) a processof disembedding takes place (Pike et al 2000 60ndash1) potentially under-mining the previous base for economic growth and value captureFrom a development point of view then the mode of territorialembeddeness or the degree of a GPN rmrsquos commitment to a partic-ular location is an important factor for value creation enhancementand capture

l Network embeddeness GPNs are characterized not only by their terri-torial embeddedness but also by the connections between networkmembers regardless of their country of origin or local anchoring inparticular places It is most notably the lsquoarchitecturersquo durability andstability of these relations both formal and informal which deter-mines the agentsrsquo individual network embeddedness (actor-networkembeddedness) as well as the structure and evolution of the GPN asa whole While the former refers to an individualrsquos or rmrsquos rela-tionships with other actors the latter consists not only of businessagents involved in the production of a particular good or service butalso takes the broader institutional networks including non-businessagents (eg government and non-government organizations) intoaccount Network embeddedness can be regarded as the product ofa process of trust building between network agents which is impor-tant for successful and stable relationships Even within intra-rmnetworks where the relationships are structured by ownership inte-gration and control trust between the different rm units and thedifferent stakeholders involved might be a crucial factor such as inthe case of joint ventures (Yeung 1998)

Conceptual dimensions

The categories sketched above are lsquoenergizedrsquo and lsquoliversquo through anumber of conceptual dimensions These constitute the frameworksthrough which value is created power exercised or institutional embed-dedness etc given concrete effect in terms of particular initiatives andpolicies There are four broad dimensions that are of signicance

FirmsOne rm is clearly not the same as another Firms even within the samesector differ in terms of their strategic priorities their attitudes to labourrelations the nature of their relations with suppliers etc As a conse-quence one would expect that while there may be similarities betweenthe ways in which rms in the same sector operate (generate value exer-cise their power over suppliers etc) there will still be important rm-specic differences not least in terms of the locations where lead rmsdecide to invest or establish supplier and subcontractor connections

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

453

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

These differences may stem from the nature of ownership (equity arrange-ments andor lsquonationalityrsquo) managerial whim or they may derive fromvalues embodied in the rmrsquos evolution25 Whatever the source of thesedifferences it is likely that they have implications for the ways in whichtheir GPNs are constructed (if they are lead rms) or for the ways inwhich they participate (seek to develop and exercise autonomy forinstance) in other rmrsquos GPNs (if they are suppliers and subcontractors)

SectorsWhile GPNs have characteristics that are rm-specic rms that oper-ate in the same sector are likely to create GPNs that have some degreeof similarity The reasons for this are that similar technologies productsand market constraints are likely to lead to similar ways of creating com-petitive advantage and thus broadly similar GPN architectures Thus forour purposes a sector need to be dened by criteria other than merestatistical classication Besides being a unique structure of competitionand technology rms in the same economic sector usually share a com-mon lsquolanguagersquo and a particular communication structure specic to thatsector (Hess 1998) A sector not only includes a range of companiesfrom the sectorrsquos leading producers to suppliers of different elementsincluding service functions but its governance structure is often com-plemented by purpose-built organizations such as industrial pressuregroups (for instance employer and labour associations) vocational train-ing institutions or others These sectoral particularities create sector-specic regulational environments were particular issues are addressedby government policies at different scales Examples of these include the supra-national multi-bre-agreement for the textiles and clothing sec-tor and national lsquosectorrsquo policies to foster innovation and competitive-ness (as is the case of some Asian countriesrsquo automobile and electronicsindustry policies)

NetworksIt is within the various networks that particular issues of governance ariseAs the ways in which power is mobilized and exercised is likely to varyfor a combination of rm and sector-specic reasons it is reasonable toexpect that the architecture of governance is likely to exhibit considerablevariation As a consequence there is likely to be signicant variation forinstance in the extent to which secondary rms in a given network arecapable of exercising a degree of autonomy that would allow them tomove into higher value-added activities with their more positive impli-cations for economic development Pending much more research that isopen to such variations it is premature to move towards a conceptualclosure of network governance structures

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

454

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

InstitutionsIn principle the institutional arrangements impact both locally andglobally on the GPNs26 they can be of considerable importance in thegeneration of value locally in its enhancement and in its capture Addi-tionally they can be of utmost signicance in setting standards (includ-ing the moral tone) for labour relations working conditions and wagelevels They are in other words central to the question of whether GPNscan deliver sustained economic and social development in the locationsthey incorporate It is important to recognize of course that the conse-quences which institutions have for GPNs and their local and interna-tional operations and implications can be positive or negative In thelatter case the institutional fabric of post-socialist Russia for instanceseems to be a case in point for all but criminal networks (see Castells2000b chapters 1 and 3) as are some of the recent policy decisions of theIMF (in relation to the East Asian crisis for instance) and WTO

GPN categories and dimensions a stylised example

As an indication of how GPNs might be visualized and analysed wedevelop below a mapping technique that allows us to highlight andcompare their main elements and linkages We apply this to a stylisedexample27 in order to underline the potential signicance of the GPNframework for the analysis of inter-organizational connections and theirrelation to economic development in the regions states and localitiesaffected by the GPN in questionIn Figure 2 we have sketched a GPN operating in four lsquoregionsrsquo and

composed of different types of rms and involving organizations of var-ious scopes from local inuence to global power In each of the regionsbe they geographic arenas (such as Eastern Europe) economic blocs (suchas the European Union) nation states or sub-national territories value iscreated and captured but to varying extents Region A for exampleshows high degrees of value generation and capture without containingmuch of the networkrsquos material ows This could be the case for a leadrm with strong RampD activities design marketing and other servicesretained in its home country while dispersing its lower value-addedproduction processes to other countries In contrast Region C is a loca-tion for value creation but it is unable to capture much of it as a resultfor instance of external ownership of many rst and second tier sup-pliers there and the prot transfers to the respective corporate head-quarters outside of the regionTwo examples of low value creation are Regions B and D The latter

shows little or no capacity to capture whatever value that is being createdwithin the region as can often be the case in pure lsquobranch plantrsquo circum-stances Neither the low value added in terms of products or technology

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

455

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

nor the enhancement of skills (value in the form of knowledge) isunusual in this type of situation The positive implications for develop-ment of being integrated in a GPN under such circumstances thereforeare rather limited In Region B on the other hand though not muchvalue is created most of it is captured within the region In this casethe ability to capture value is enhanced by the non-rm institutionsshown in Figure 1 This ability of course strongly relates to questionsof power and its distributionThe power exercised within the GPN can be shown as non-material

ows between different agents (rms as well as other organizations) The

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

456

stateType

quasi-state

non-state

globalInfluenceReach

Non-material

Colour

Institutions

Firms

RelationsFlows

SpacePlace

national

localregional

Final ProducerAssembler (focal firm)

1st TierNetwork Member

Sector x

Powerinformation(Direction Intensity)

SubsidiaryBranch Plant

OEM

Key

DistributorRetail

2nd TierNetwork Member Sector y

Material

Value Capture

Value Creation

InputOutput(QuantityImportance)

Dominated Relationship

Mutual Relationship

low high

low

high

Region A

Region B Region D

Region C

Eg IMF

Eg Labour Union

Eg City Council

Focal Firm

Eg Profit Transfer Legislation

Eg National MoF

Eg Joing RampD for Specific Systems

Eg National Environmental Group Eg OEM

Focal Firm Different Sector

Figure 2 Mapping global production networks ndash a stylised example

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

value generation and capture process in Region B for instance is heavilydetermined by the power of global institutions (eg IMF) and nationalgovernment agencies here represented as the ministry of nance (MoF)inuencing the prot transfers of foreign owned subsidiaries An actualcase in point would be the GPN of foreign car assemblers in China wheregovernment power is used to inuence the localization of productionand therefore the value generation and capture process including aspectsof employment skills and technology transferThe corporate power of some rms over their regional environment

on the other hand is exemplied by Region Arsquos lead rm affecting thelocal administration as shown in Figure 2 while collective power isexercised by the labour union There are plenty of examples where acompany or a group of rms are able to shape the institutional andregional environment in its favour especially on the local and nationallevels as in economically weak regions eager to attract or retain externalinvestmentThe territorial embeddedness of the network under consideration is not

immediately educible but can be represented by the density and inten-sity of localregional or national connections between the differentagents Firms and organizations in Region D for instance have only fewand rather weak relations with each other There then territorial embed-dedness is limited Linkages to other agents outside the region on theother hand are comparatively strong indicating a high degree of networkembeddednessIn sum the technique of mapping the GPN demonstrated here pro-

vides the possibility of visualizing the economic and social agents aswell as highlighting the structural and spatial dimensions of networkssectors and the linkages between them It allows us to visualize theGPNrsquos implications for development in different places within the GPNsrsquoterritorial reach and the main agents responsible for these implicationsWhat cannot be shown of course is the evolution of the GPN over time(path dependency) and structural preconditions shaping it (such asdifferent national capitalisms or national modes of regulation) Never-theless what we have here is a scale-transcending model of globalproduction networks which provides a sense of their likely implicationsfor economic and social development

CONCLUSION

In this article we have outlined a conceptual framework for mappingand analysing certain aspects of economic globalization ndash those relatedto production and consumption ndash and their developmental consequencesIn so doing we have foregrounded the ways in which companies orga-nize and control their global operations the ways in which they are (or

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

457

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

can be) inuenced by states trade unions NGOs and other institutionsin particular locations and the implications that the resulting combina-tions of agents and processes might have for industrial upgrading highervalued added etc and ultimately for the prospects of poverty reductionandor generalized prosperity in those locationsThe framework we have proposed ndash that of the global production

network ndash is an explicit attempt to break with state-centric conceptual-isations on the one hand and signicantly extend the analytic and policyutility of cognate formulations on the other The proof of success how-ever will depend on whether the GPN framework stimulates researchthat delivers analyses that are both empirically and theoretically richerthan at present More importantly however it will depend on whetherthe framework helps to produce research that contributes more effec-tively to the task of improving the human condition in the age ofeconomic and geo-political turbulence in which we now exist

ACKNOWLEDGEMENTS

This article draws on work conducted under the auspices of the Econ-omic and Social Research Council project Making the Connections GlobalProduction Networks in Europe and East Asia (Grant R 000 238535) We are grateful to the ESRC for their support and to Dieter Ernst JohnHumphrey Alisdair Rogers three anonymous referees and the partici-pants at the Annual Conference of the Global Studies Association(Manchester July 2001) for their comments on an earlier version

NOTES

1 For the purposes of this article our intellectual engagement is largely withcontributions to the political economy of development be they in lsquosociolog-icalrsquo lsquoeconomicrsquo lsquogeographicrsquo or lsquopolitical sciencersquo guises

2 We have in mind here the arguments of the supposed panacea of lsquofreersquomarkets as development tools on the one hand through to the stress onstate industrial initiatives on the other as well as those that view the state-market relation as symbiotic for development purposes In all of these caseshowever the analytical weight tends to be placed on the nature and appli-cation of state economic policy (cf Evans 1992)

3 The few notable monographs here (such as Geref 1983 Henderson 1989Doner 1991 Sklair 1993) only serve to underline the general rule

4 We do not mean to deny the relevance of some state-centric contributionsto the analysis of globalization and its problems and how the latter mightbe resolved Some of the work on the East Asian crisis for instance (egChang 1998 Henderson 1999 Weiss 1999) are cases in point

5 See for example Geref (1995 1999a) Geref and Korzeniewicz (1994) Sklair(1995) and the IDS Bulletin (323 2001) which focuses entirely on globalvalue chains

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

458

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

6 On the former see for instance Jessoprsquos (2001) collection of some of theseminal contributions On the latter see Storper and Salais (1997 particularlychapter 10)

7 Hardly any work has been done for instance on households states and thereproduction of labour power from within a GCCs perspective

8 See for instance the work on the Brazilian lsquoreserved marketrsquo for personalcomputers (Evans 1986 Schmitz and Hewitt 1992)

9 See for instance the essays collected in Geref and Korzeniewicz (1994) andthe special issue of the IDS Bulletin (323 2001) See also Clancy (1998) Dolanand Humphrey (2000) Bonacich and Appelbaum (2000) and Kaplinsky (2000)among others The ILOrsquos research institute the International Institute ofLabour Studies sponsored a programme on lsquoglobal commodity chainsrsquo in the late 1990s The continuing media attention to the exploitative work-ing conditions evident in the supplier companies integrated into the chainsof the likes of Nike and Gap for instance underlines the utility of the GCCframework for agencies such as the ILO

10 Specically it implies rejection of the term lsquoglobalrsquo as a simplistic geograph-ical construct (see our later discussion) Similarly economic lsquoglobalizationrsquocomes to refer to the extension of functionally integrated (and thus sociallyrelational) economic activities across national boundaries (cf Dicken 19985) The implication of this for the conceptualisation of GPNs is that theycome to be seen as dynamic typologies which potentially change shape andscope over time

11 Though he had previously worked with the notion of the lsquointernationalproduction networkrsquo Ernst rst used the term lsquoglobal production networkrsquoin a conference paper of 1999 (Ernst 1999) Our rst attempt to elaborate aGPN framework appeared in a research proposal that same year (Dickenand Henderson 1999)

12 See for instance Hughesrsquo (2000) study of the cut-ower trade13 It is unlikely to be of particular help for instance for the analysis of some

forms of nance capital such as bank loans and portfolio investment14 For a discussion of regional politics and production networks see Cabus

and Hess (2000)15 In other works a continuum of scales (see Swyngedouw 1997 Dicken and

Malmberg 2001)16 There is a growing literature that addresses these concerns with respect to

differing lsquoqualitiesrsquo of foreign direct investment See for instance Turok(1993 Amin et al (1994) and Young et al (1994)

17 We have in mind the continuing dis-investment in British subsidiaries (withknock-on effects for local suppliers) by foreign companies Since 1998 thesehave included at a minimum Siemens Samsung LG and Motorola (in elec-tronics) BMW Ford and General Motors (in automobiles) and Corus (steel)

18 Germany on the one hand and Britain and the USA on the other consti-tute polar opposites in this sense In the latter shareholders have supremepower over the disposal of prots and assets while in the former ownersare obliged to consider the interests of other stakeholders and the workforcein particular (Lane 1989) Indeed in Germany property holders have a consti-tutional obligation to exercise their rights in the interests of the public good(Hutton 2001)

19 Although not theorized in terms of power Humphrey and Schmidtrsquos (2001)discussion of the governance structures of lsquovalue chainsrsquo is an importantcomplement at this point to our work

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

459

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

20 Castells develops ideas similar to these with regard to the exercise ofeconomic and foreign policy by national states absorbed into lsquonetwork statesrsquo(of which the European Union is the prototype) See Castells (2000b chapter5) and also Carnoy and Castells (2001)

21 This is obviously not the place to explain such discrepancies except to mark that the answers seem to lie in a combination of political will (or itsabsence) and differing institutional capacities for economic governance Forthe British and Indonesian cases see Hutton (1995) and Hill (1996) respec-tively For more general and theoretical accounts of the relation between statecapacities and economic development see Evans (1995) and Evans and Rauch(1999)

22 Japanese companies for instance have never offered lsquopermanent employ-mentrsquo contracts to employees in their foreign subsidiaries Similarly Germancompanies though required by German and EU legislation to consult exten-sively with employees before instituting redundancy programmes have neverdone so in countries where such laws do not apply Recent disinvestments in Britain by Siemens and BMW are cases in point

23 As Oinas (1997) Markusen (1999) and Pike et al (2000) among others havepointed out the notion of embeddedness still remains rather vague andtherefore in need of conceptual improvement However its importance forthe understanding of economic organization is widely acknowledged evenby critical voices (see for example Sayer 2000)

24 There is also a downside The nature of local networks and socio-economicrelations may under certain circumstances generate an inability to captureglobal opportunities and lead to regional economic downturn (Oinas 199726) Strong embeddedness therefore is not necessarily a lsquogoodrsquo or positivequality of networks or its agents

25 Examples in Britain for instance include the ethical stance of such com-panies as the Co-operative Bank and the Body Shop

26 From the perspective of actor-network theory GPNs would themselves beinstitutions However this is not the position adopted here Rather institu-tions are perceived as social and political formations ndash be they sub-nationalnational or international ndash with attendant histories values and cultural prac-tices which have consequences for how GPNs are formed and develop overtime

27 This example is generalized and therefore by no means comprehensive realGPNs of course have far more linkages and agents than could be sketchedhere For instance due to graphic restrictions the intra-rm network owsof the lead company are not shown in Figure 2

REFERENCES

Amin A and Thrift N (1994) lsquoHolding down the globalrsquo in A Amin and NThrift (eds) Globalization Institutions and Regional Development in EuropeOxford University Press pp 257ndash60

Amin A Bradley D Howells J Tomaney J and Gentle C (1994) lsquoRegionalincentives and the quality of mobile investment in the less favoured regionsof the ECrsquo Progress in Planning 41(1) 1ndash112

Armstrong W and McGee TG (1985) Theatres of Accumulation New YorkMethuen

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

460

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

Bonacich E and Appelbaum RP (2000) Behind the Label Inequality in the LosAngeles Garments Industry Berkeley and Los Angeles University of CaliforniaPress

Boyer R and Drache D (eds) (1996) States against Markets The Limits of Global-ization London Routledge

Cabus P and Hess M (2000) lsquoRegional politics and economic patterns ldquoglocal-isationsrdquo and the network enterprisersquo BELGEO 1 79ndash101

Carnoy M and Castells M (2001) lsquoGlobalization the knowledge society andthe Network State Poulantzas at the milleniumrsquo Global Networks 1(1) 1ndash18

Castells M (2000a) The Rise of the Network Society (2nd Edition) Oxford BlackwellCastells M (2000b) End of Millennium (2nd Edition) Oxford BlackwellCoates D (2000) Models of Capitalism Cambridge Polity PressChang H-J (1998) lsquoKorea the misunderstood crisisrsquo World Development 26(8)

1555ndash61Clancy M (1998) lsquoCommodity chains services and development theory and

preliminary evidence from the tourism industryrsquo Review of InternationalPolitical Economy 5(1) 122ndash48

Cook I and Crang P (1996) lsquoThe world on a plate culinary culture displace-ment and geographical knowledgesrsquo Journal of Material Culture 1 131ndash53

Czaban L and Henderson J (1998) lsquoGlobalisation institutional legacies andindustrial transformation in Eastern Europersquo Economy and Society 27(4)585ndash613

Dicken P (1998) Global Shift (3rd Edition) London Paul ChapmanDicken P and Henderson J (1999) lsquoMaking the connections global production

networks in Britain East Asia and Eastern Europersquo A research proposal tothe Economic and Social Research Council (July)

Dicken P Kelly PF Olds K and Yeung H W-C (2001) lsquoChains and networksterritories and scales towards a relational framework for analysing the globaleconomyrsquo Global Networks 1(2) 89ndash112

Dicken P and Malmberg A (2001) lsquoFirms in territories a relational perspec-tiversquo Economic Geography 77 345ndash63

Dolan C and Humphrey J (2000) lsquoGovernance and trade in fresh vegetablesthe impact of UK supermarkets on the African horticulture industryrsquo Journalof Development Studies 37(2) 147ndash76

Doner RF (1991) Driving a Bargain Berkeley and Los Angeles University ofCalifornia Press

Ernst D (1999) lsquoGlobalization and the changing geography of innovation sys-tems A policy perspective on global production networksrsquo Paper presentedat the Workshop on the Political Economy of Technology in DevelopingCountries Brighton October 8ndash9

Ernst D (2000) lsquoGlobal production networks and the changing geography ofinnovation systems implications for developing countriesrsquo East-West CentreWorking Paper 9 Honolulu

Ernst D and Kim L (2001) lsquoGlobal production networks knowledge diffusionand local capability formation a conceptual frameworkrsquo Paper presented atthe Nelson amp Winter Conference Aalborg 12ndash15 June

Evans P (1986) lsquoState capital and the transformation of dependence the Braziliancomputer casersquo World Development 14(7) 791ndash808

Evans P (1992) lsquoThe state as problem and solution predation embeddedautonomy and structural changersquo in S Haggard and R Kaufman (eds) ThePolitics of Economic Adjustment International Constraints Distributive Conictsand the State Princeton Princeton University Press pp 139ndash81

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

461

Evans P (1995) Embedded Autonomy Princeton Princeton University PressEvans P and Rauch J (1999) lsquoBureaucracy and growth a cross-national analysis

of the effects of ldquoWeberianrdquo state structures on economic growthrsquo AmericanSociological Review 64(4) 748ndash65

Froumlbel F Heinrichs J and Kreye O (1980) The New International Division ofLabour Cambridge Cambridge University Press

Geref G (1983) The Pharmaceutical Industry and Dependency in the Third WorldPrinceton Princeton University Press

Geref G (1994) lsquoThe organisation of buyer-driven global commodity chainshow US retailers shape overseas production networksrsquo in G Geref and MKorzeniewicz (eds) Commodity Chains and Global Development WestportPraeger pp 95ndash122

Geref G (1995) lsquoGlobal production systems and third world developmentrsquo inB Stallings (ed) Global Change Regional Response New York CambridgeUniversity Press pp 100ndash42

Geref G (1999a) lsquoInternational trade and industrial upgrading in the apparelcommodity chainrsquo Journal of International Economics 48(1) 37ndash70

Geref G (1999b) lsquoA commodity chains framework for analysing global indus-triesrsquo mimeo Department of Sociology Duke University

Geref G and Korzeniewicz M (eds) (1994) Commodity Chains and GlobalCapitalism Westport Praeger

Grabher G (1993) lsquoThe weakness of strong ties the lock-in of regional devel-opment in the Ruhr arearsquo in G Graher (ed) The Embedded Firm On theSocio-Economics of Inter-rm Relations London Routledge pp 255ndash78

Harrison B (1992) lsquoIndustrial districts old wine in new bottlesrsquo Regional Studies26 469ndash83

Harvey D (1969) Explanation in Geography London Edward ArnoldHausner J Jessop B and Nielsen K (eds) (1995) Strategic Choice and Path

Dependency in Post-Socialism Aldershot Edward ElgarHein W (2000) lsquoDie Oumlkonomie des Archipels und das versunkene Landrsquo E+Z

41(11) 304ndash7Henderson J (1989) The Globalisation of High Technology Production London

RoutledgeHenderson J (1993) lsquoThe role of the state in the economic transformation of

East Asiarsquo in C Dixon and D Drakakis-Smith (eds) Economic and SocialDevelopment in Pacic Asia London Routledge pp 85ndash114

Henderson J (1999) lsquoUneven crises institutional foundations of East Asianeconomic turmoilrsquo Economy and Society 28(3) 327ndash68

Henderson J and Castells M (eds) (1987) Global Restructuring and TerritorialDevelopment London Sage Publications

Hess M (1998) Glokalisierung industrieller Wandel und Standortstruktur MunichVVF

Heyzer N (1986) Working Women in Southeast Asia Milton Keynes OpenUniversity Press

Hill H (1996) Southeast Asiarsquos Emerging Giant Indonesian Economic Policy andDevelopment since 1996 Cambridge Cambridge University Press

Hughes A (2000) lsquoRetailers knowledge and changing commodity networks thecase of the cut ower tradersquo Geoforum 31 175ndash90

Humphrey J and Schmitz H (2001) lsquoGovernance in global value chainsrsquo IDSBulletin 32(3) 19ndash29

Hutton W (1995) The State Wersquore In London Jonathan CapeHutton W (2001) lsquoWhy the Germans are right about usrsquo The Observer 27 May

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

462

Jackson P (1999) lsquoCommodity cultures the trafc in thingsrsquo Transactions of theInstitute of British Geographers 24 95ndash108

Jenkins R (1987) Transnational Corporations and Uneven Development LondonMethuen

Jessop B (ed) (2001) Regulation Theory and the Crisis of Capitalism Volume I ndashThe Parisian School Cheltenham Edward Elgar

Kaplinsky R (1998) lsquoGlobalisation industrialisation and sustainable growth thepursuit of the nth rentrsquo Discussion Paper 365 Institute of DevelopmentStudies University of Sussex

Kaplinsky R (2000) lsquoSpreading the gains from globalisation what can be learnedfrom value chain analysisrsquo Working Paper 110 Institute of DevelopmentStudies University of Sussex

Lane C (1989) Industry and Society in Europe Cheltenham Edward ElgarLaw J (1999) lsquoAfter ANT complexity naming and typologyrsquo in J Law J and

J Hassard (eds) Actor-network Theory and After Oxford Blackwell pp 1ndash14Lenz B (1997) lsquoDas Filiegravere-Konzept als Analyseinstrument der organisatorischen

und raumlumlichen Anordnung von Produktions- und DistributionsprozessenrsquoGeographische Zeitschrift 85(1) 20ndash33

Markusen A (1999) lsquoFuzzy concepts scanty evidence policy distance the casefor rigor and policy relevance in critical regional studiesrsquo Regional Studies33(9) 869ndash84

Murdoch J (1998) lsquoThe spaces of actor-network theoryrsquo Geoforum 29(4) 357ndash74Mitter S and Rowbotham S (eds) (1995) Women Encounter Technology London

RoutledgeOinas P (1997) lsquoOn the socio-spatial embeddedness of business rmsrsquo Erdkunde

51(1) 23ndash32Pike A Lagendijk A and Vale M (2000) lsquoCritical reections on ldquoembedded-

nessrdquo in economic geography the case of labour market governance andtraining in the automotive industry in the North-East region of Englandrsquo inA Giunta A Legendijk and A Pike (eds) Restructuring Industry and TerritoryThe Experience of Europersquos Regions London The Stationery Ofce pp 59ndash82

Porter M (1985) Competitive Advantage Creating and Sustaining Superior Perform-ance London Macmillan

Porter M (1990) The Competitive Advantage of Nations London MacmillanRaikes P Jensen MF and Ponte S (2000) lsquoGlobal commodity chain analysis

and the French liere approach comparison and critiquersquo Economy and Society29(3) 390ndash417

Sassen S (1999) lsquoServicing the global economy recongured states and privateagentsrsquo in K Olds P Dicken PK Kelly L Kong and H W-C Yeung (eds)Globalisation and the Asia-Pacic London Routledge pp 149ndash62

Sayer A (2000) lsquoMarkets embeddedness and trust problems of polysemy andidealismrsquo mimeo Department of Sociology University of Lancaster

Schmitz H and Hewitt T (1992) lsquoAn assessment of the market reserve for theBrazilian computer industryrsquo in H Schmitz and J Cassiolato (eds) Hi-Techfor Industrial Development Lessons from the Brazilian Experience in Electronicsand Automation London Routledge pp 21ndash52

Scott AJ (1998) Regions and the World Economy The Coming Shape of GlobalProduction Competition and Political Order Oxford Oxford University Press

Sklair L (1993) Assembling for Development (2nd Edition) San Diego Universityof California Center of US-Mexico Relations

Sklair L (1995) Sociology of the Global System (2nd Edition) London Prentice Hall

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

463

Jackson P (1999) lsquoCommodity cultures the trafc in thingsrsquo Transactions of theInstitute of British Geographers 24 95ndash108

Jenkins R (1987) Transnational Corporations and Uneven Development LondonMethuen

Jessop B (ed) (2001) Regulation Theory and the Crisis of Capitalism Volume I ndashThe Parisian School Cheltenham Edward Elgar

Kaplinsky R (1998) lsquoGlobalisation industrialisation and sustainable growth thepursuit of the nth rentrsquo Discussion Paper 365 Institute of DevelopmentStudies University of Sussex

Kaplinsky R (2000) lsquoSpreading the gains from globalisation what can be learnedfrom value chain analysisrsquo Working Paper 110 Institute of DevelopmentStudies University of Sussex

Lane C (1989) Industry and Society in Europe Cheltenham Edward ElgarLaw J (1999) lsquoAfter ANT complexity naming and typologyrsquo in J Law J and

J Hassard (eds) Actor-network Theory and After Oxford Blackwell pp 1ndash14Lenz B (1997) lsquoDas Filiegravere-Konzept als Analyseinstrument der organisatorischen

und raumlumlichen Anordnung von Produktions- und DistributionsprozessenrsquoGeographische Zeitschrift 85(1) 20ndash33

Markusen A (1999) lsquoFuzzy concepts scanty evidence policy distance the casefor rigor and policy relevance in critical regional studiesrsquo Regional Studies33(9) 869ndash84

Murdoch J (1998) lsquoThe spaces of actor-network theoryrsquo Geoforum 29(4) 357ndash74Mitter S and Rowbotham S (eds) (1995) Women Encounter Technology London

RoutledgeOinas P (1997) lsquoOn the socio-spatial embeddedness of business rmsrsquo Erdkunde

51(1) 23ndash32Pike A Lagendijk A and Vale M (2000) lsquoCritical reections on ldquoembedded-

nessrdquo in economic geography the case of labour market governance andtraining in the automotive industry in the North-East region of Englandrsquo inA Giunta A Legendijk and A Pike (eds) Restructuring Industry and TerritoryThe Experience of Europersquos Regions London The Stationery Ofce pp 59ndash82

Porter M (1985) Competitive Advantage Creating and Sustaining Superior Perform-ance London Macmillan

Porter M (1990) The Competitive Advantage of Nations London MacmillanRaikes P Jensen MF and Ponte S (2000) lsquoGlobal commodity chain analysis

and the French liere approach comparison and critiquersquo Economy and Society29(3) 390ndash417

Sassen S (1999) lsquoServicing the global economy recongured states and privateagentsrsquo in K Olds P Dicken PK Kelly L Kong and H W-C Yeung (eds)Globalisation and the Asia-Pacic London Routledge pp 149ndash62

Sayer A (2000) lsquoMarkets embeddedness and trust problems of polysemy andidealismrsquo mimeo Department of Sociology University of Lancaster

Schmitz H and Hewitt T (1992) lsquoAn assessment of the market reserve for theBrazilian computer industryrsquo in H Schmitz and J Cassiolato (eds) Hi-Techfor Industrial Development Lessons from the Brazilian Experience in Electronicsand Automation London Routledge pp 21ndash52

Scott AJ (1998) Regions and the World Economy The Coming Shape of GlobalProduction Competition and Political Order Oxford Oxford University Press

Sklair L (1993) Assembling for Development (2nd Edition) San Diego Universityof California Center of US-Mexico Relations

Sklair L (1995) Sociology of the Global System (2nd Edition) London Prentice Hall

HENDERSON ET AL GLOBAL PRODUCTION NETWORKS

463

Stark D (1992) lsquoPath dependence and privatization strategies in East-CentralEuropersquo East European Politics and Societies 6(1) 17ndash51

Storper M and Salais R (1997) Worlds of Production The Action Frameworks ofthe Economy CambridgeMass Harvard University Press

Sturgeon TJ (2001) lsquoHow do we dene value chains and production networksrsquoIDS Bulletin 32(3) 9ndash18

Swyngedouw E (1997) lsquoHoming in and spacing out re-conguring scalersquo Paperpresented at the Deutscher Geographentag Conference Bonn 6ndash11 October

Turok I (1993) lsquoInward investment and local linkages how deeply embeddedis ldquoSilicon Glenrdquorsquo Regional Studies 27(5) 401ndash18

Wade R (1990) Governing the Market Princeton Princeton University PressWeiss L (1999) lsquoState power and the Asian crisisrsquo New Political Economy 4(3)

317ndash42Whatmore S and Thorne L (1997) lsquoNourishing networks alternative geogra-

phies of foodrsquo in DJ Goodman and MJ Watts (eds) Globalizing FoodAgrarian Questions and Global Restructuring London Routledge pp 287ndash304

Whitley R (1996) lsquoBusiness systems and global commodity chains competingor complementary forms of economic organizationrsquo Competition and Change1(4) 411ndash25

Whitley R (1999) Divergent Capitalisms Oxford Oxford University PressYeung H W-C (1998) lsquoThe social-spatial constitution of business organisation

a geographical perspectiversquo Organization 5(1) 101ndash28Young S Hood N and Wilson A (1994) lsquoTargeting policy as a competitive

strategy for European inward investment agenciesrsquo European Urban andRegional Studies 1(2) 143ndash59

Zysman J (1983) Governments Markets and Growth Financial Systems and thePolitics of Industrial Change Oxford Robertson

REVIEW OF INTERNATIONAL POLITICAL ECONOMY

464