Explaining governance in global value chains: A modular theory-building effort

ABSTRACT

In this article, we review the evolution and current status of global value chain (GVC) governance theory and take some initial steps toward a broader theory of governance through an exercise in ‘modular theory-building’. We focus on two GVC governance theories to which we previously contributed: a theory of linking and a theory of conventions. The modular framework we propose is built on three scalar dimensions: (1) a micro level – determinants and dynamics of exchange at individual value chain nodes; (2) a meso level – how and to what extent these linkage characteristics ‘travel’ upstream and downstream in the value chain; and (3) a macro level – looking at ‘overall’ GVC governance. Given space limitations, we focus only on the issue of ‘polarity’ in governance at the macro level, distinguishing between unipolar, bipolar and multipolar governance forms. While we leave a more ambitious analysis of how overall GVC governance is mutually constituted by micro/meso factors and broader institutional, regulatory and societal processes to future work, we provide an initial framework to which this work could be linked. Our ultimate purpose is to spur future efforts that seek to use and refine additional theories, to connect theories together better or in different modular configurations, and to incorporate elements at the macro level that reflect the changing constellation of key actors in GVC governance – the increasing influence of, for example, NGOs, taste and standard makers, and social movements in GVC governance.

KEYWORDS:

• Governance
• global value chains
• polarity
• linking mechanisms
• conventions
• modular theory building

ACKNOWLEDGEMENTS

Useful feedback on earlier versions of this article was received at the the workshop on `Value Chains, Production Networks, and the Geographies of Development: Emerging Challenges and Future Agenda', National University of Singapore (1–2 December, 2011) and at the convention of the Association of American Geographers (AAG), New York (25–29 February, 2012). The authors would also like to thank Gary Gereffi, John Humphrey and two anonymous referees for insightful and constructive comments.

Notes

¹Bair and Werner (2011) offer a ‘disarticulation’ perspective on GVCs that looks at the dynamics of exclusion and expulsion in GVCs (see also other contributions in the same special issue). Building on earlier work by Bair (2009), this perspective calls for a return to the long-range, macro historical roots of GVC analysis, with a focus on historical change, disjunctions, fragmentations and disarticulations. Disarticulation scholars argue that the GVC literature has developed a bias on inclusion dynamics and that attention should be paid to how links in the chain are forged, not only in material terms, but also ideologically and in relation to the creation of subjectivities. They examine the set of social relations that secure commodity production and related processes of exclusion (Bair and Werner, 2011) and highlight the social and spatial contours of production through everyday practices and struggles over the creation and appropriation of value (see also Neilson and Pritchard, 2009; Bair, 2011).

²This does not mean that any theory can be simply ‘plugged into’ our framework. GVC perspectives are generally ‘activity-oriented’, thus, for example, an accumulation perspective to understanding the global economy is less likely to fit in our framework of GVC governance.

³The original classification in Gibbon, Bair and Ponte (2008) mentioned governance as ‘driving’, as ‘coordination’ and as ‘normalization’. Here, we adopt the term ‘linking’, instead of coordination, to avoid confusion in Figure 2, where we will use both linkages and conventions to explain the nature of coordination at individual nodes of the value chain. We also use the term ‘normalizing’, instead of ‘normalization’, to use the same grammatical form for all three.

⁴Feenstra and Hamilton (2006) describe in detail the ways in which retailers gained power relative to manufacturers, beginning in the United States in the 1960s, a trend that continues to the present day. This ‘retail revolution’ has been a major factor in de-industrialization within the United States, as retailers increased overseas sourcing of apparel, electronics and consumer goods, in turn forcing manufacturers to move their own facilities offshore and increase sourcing in low-cost locations in East Asia. At the same time, it spurred ‘late’ industrialization and industrial upgrading, first in Japan, and later in Korea and Taiwan (Amsden, 1989; Wade, 1990; Evans, 1995).

⁵Global production network (GPN) scholars (Dicken et al., 2001; Henderson et al., 2002; Hess and Yeung, 2006; Coe, Dicken and Hess, 2008; Yeung, 2009) have been similarly motivated to respond to the perceived limitations of the GCC approach: that it was too structuralist and long-range historical in nature, ‘linearist’ (using the value-added chain for specific commodities as the unit of analysis), and focused on a narrow binomial approach to governance (buyer-driven vs. producer-driven) (Gereffi, 1994). Much has been made in the literature of the differences between the GVC and GPN approaches. The GPN perspective, rooted in the field of economic geography, is less concerned with the hierarchy of power in specific cross-border industries and more concerned with how business networks are embedded in, and intertwine with, institutions and the social and cultural contexts extant in specific locations. Because of this, the GPN literature tends to highlight complexity, messiness and variety in production networks (Dicken et al., 2001; Henderson et al., 2002; Hess and Yeung, 2006). However, many similarities are also apparent. Both sets of literature approach the ‘global’ in ‘bottom-up’ fashion, through generalization from field research on product- and even firm-specific experiences. They provide macro representations extrapolated from case studies highlighting the ideal–typical dynamics of inter-firm exchange in specific industries, rather than by examining anonymous trade and investment flows at the level of nation-states. The GVC and GPN literatures have a similar interest in mapping the spatial and organizational division of labour and understanding the qualitative aspects of how value is created, allocated and captured in specific industries and locations. They share a normative interest in the upgrading trajectories of less developed places and how value-creation processes and learning dynamics can benefit or exclude disadvantaged regions and actors. Finally, the more recent GVC literature has in a way been more GPN-like by paying increased attention to institutional, regulatory and standard-making processes (Neilson and Pritchard, 2009; Cattaneo, Gereffi and Staritz, 2010; Ponte, Gibbon and Vestergaard, 2011). It has strived to include in its analysis some of the insights of cultural political economy, it has reflected upon differences in business culture or normative expectations of how a ‘good business’ is organized (Gibbon and Ponte, 2005; Ponte, 2009; Bair and Werner, 2011), and it has been fine-tuning the territorial and contested dimensions of restructuring in global value chains (Neilson and Pritchard, 2009; Mahutga, 2012).

⁶It also drew in key concepts from a range of ‘outside’ literatures: the asset-specificity problem and the markets versus hierarchy frame from institutional economics, the social embeddedness of economic life from industrial sociology, the relationship between tacit knowledge sharing and spatial agglomeration from economic geography, the enabling role of technology and standards from technology management, and the pragmatic approach to firm-level learning from management and evolutionary economics. As such, the original GVC governance framework can be seen as a prior exercise in modular theory building.

⁷Other work on governance as normalization has been carried out through the lenses of governmentality (Gibbon and Ponte, 2008).

⁸This list is not meant to be exhaustive. Other contributions have examined a ‘connectivist’ convention (Boltanski and Chiapello, 1999) and a ‘green’ convention (Latour, 1998). However, much of the literature engaging with conventions actually uses only three typologies (market, industrial and domestic), sometimes four (including civic). Ponte (2009), however, has found it helpful to engage with all six conventions that parallel the six orders of worth in Boltanski and Thévenot (1991).

⁹These nodes are only meant to illustrate the analytic process; they are not meant to be exhaustive.

¹⁰Of course, the analysis does not need to start with the linkages between lead firms and first-tier suppliers. While these do tend to be important, the analysis can start at any node. The key element of GVC methods, however, is that research and analysis is not confined to a single node. Depending on the nature of the GVC, there may be more or fewer nodes than illustrated in Figure 2.