Chapter 2: Theoretical Foundations: GVC Governance, Industrial Upgrading, and the Capability-Based View of Green Competitiveness
2.1 GVC Governance and Upgrading: Lead Firm Governance and Upgrading Trajectories
2.1.1 The governance turn in global value chain analysis
Global value chain research began as an exercise in mapping. Researchers traced the sequence of activities—design, component production, assembly, branding, retail—that together deliver a product to its final consumer, and asked where in the world each activity was performed and by whom. That descriptive project quickly ran into a theoretical problem. The same activity could be organized in radically different ways depending on the product and the chain. A garment assembled in Guangdong for a European fast-fashion retailer operated under a different set of commercial arrangements than a specialty steel component produced in Germany for a German machine-tool maker, even if both were "global" in the sense that inputs crossed borders. Mapping alone could not explain why the organizational forms differed, nor could it predict what would happen to a supplier that wanted to move from assembly into design.
The governance turn answered this question by making the lead firm the central analytical object. Gereffi's early work on buyer-driven and producer-driven commodity chains (Gereffi, 1994; Gereffi and Korzeniewicz, 1994) established the basic insight that power in a chain is not distributed evenly and that the location of power shapes the opportunities available to firms upstream. In buyer-driven chains—apparel, footwear, furniture, toys—retailers and branded marketers set the terms, and manufacturers compete for orders on price, delivery reliability, and increasingly on compliance with codes of conduct. In producer-driven chains—automobiles, aircraft, semiconductors—manufacturers with proprietary technology occupy the lead position and coordinate a tiered supplier network. The distinction was crude but generative. It told suppliers that the route to upgrading depended on who was buying and what they valued.
The framework was refined substantially in the 2000s. Humphrey and Schmitz (2002a, 2002b) distinguished four upgrading trajectories: process upgrading (producing more efficiently), product upgrading (producing more sophisticated goods), functional upgrading (taking on new functions in the chain, such as design or marketing), and inter-chain upgrading (moving the capabilities acquired in one chain into a new, higher-value chain). This taxonomy has proved durable because it separates the question "what are you getting better at?" from the question "who benefits from your improvement?" A supplier that undertakes process upgrading may capture none of the gains if the lead firm uses the cost reduction to demand lower prices. A supplier that undertakes functional upgrading may find the lead firm actively resistant, because the supplier is encroaching on functions the lead firm considers its own.
Gereffi, Humphrey, and Sturgeon (2005) took the next step with a typology of governance forms: market, modular, relational, captive, and hierarchy. The typology is generated by three variables—the complexity of transactions, the codifiability of information, and the competence of suppliers. Where transactions are simple and information codifiable, market governance prevails and suppliers face low switching costs. Where transactions are complex but supplier competence is high, relational governance prevails and coordination happens through repeated interaction and trust. Where supplier competence is low and the lead firm needs to specify what it wants in detail, captive governance prevails: the supplier is locked into the relationship because its capabilities were built to serve one customer's specifications. The captive form matters for our purposes because it describes the position of many Chinese export manufacturers in the 1990s and 2000s. It also describes the condition that makes compliance-based upgrading a trap: a supplier that invests only in meeting the lead firm's specified requirements deepens its dependence without building capabilities that transfer to other customers or to higher-value functions.
The governance literature has since multiplied. Global production network analysis (Henderson et al., 2002) emphasized that value chains are embedded in territorially specific institutional and social relations, and that the same chain can produce different outcomes in different places. Ponte and Sturgeon (2014) proposed a distinction between "governance as driving" (who sets the strategic direction) and "governance as coordination" (how activities are aligned), which helps explain cases where a lead firm sets strategy but delegates coordination to intermediaries. Gibbon, Bair, and Ponte (2008) reviewed the field and noted that governance typologies had become more numerous than the empirical cases could comfortably support. The proliferation is a sign of health in one sense—the phenomenon is complex—but it creates a problem for anyone trying to use the framework to analyze a specific industry's green transition. Which governance form applies, and what does it imply for a supplier's ability to build green capabilities?
The governance form that matters most for green capability building is not determined solely by transaction complexity, codifiability, and supplier competence in the abstract. It is determined by the interaction between those variables and the lead firm's sustainability strategy. A lead firm that treats environmental performance as a compliance checkbox will govern its suppliers through specification and audit—a form close to captive governance, even where the supplier is technically competent, because the lead firm wants verification rather than innovation. A lead firm that treats environmental performance as a source of product differentiation will govern through relational mechanisms, because it needs suppliers to solve problems it has not fully specified. The same supplier can therefore face different governance forms from different customers, and its green capability trajectory will depend on which customers dominate its order book. This is the first modification the book makes to the standard GVC governance typology, and it recurs throughout the empirical chapters.
2.1.2 Upgrading as a contested process
The upgrading literature has always contained a tension between the optimistic view that chain insertion is a ladder and the pessimistic view that it is a treadmill. The optimistic view, associated with the early Gereffi work and with the policy literature that followed, holds that participation in global value chains exposes suppliers to knowledge, standards, and market signals that they would not otherwise encounter, and that suppliers can use these exposures to climb. The pessimistic view, developed by Humphrey and Schmitz (2002a) and elaborated by later scholars, holds that lead firms design governance to capture the gains from upgrading, and that suppliers who attempt functional upgrading face retaliation—order withdrawal, price pressure, or the transfer of the order to a competing supplier.
The empirical record supports both views in different contexts. Bazan and Navas-Alemán's (2003) study of the Sinos Valley footwear cluster in Brazil found that firms inserted into global chains upgraded their production processes faster than firms selling domestically, but that firms selling domestically were more likely to undertake functional upgrading into design and branding. The global chain delivered process knowledge and withheld functional opportunity. Bair and Gereffi's (2001) study of Torreón's blue jeans industry found a similar pattern: full-package production represented a real advance over assembly, but the advance was bounded by the retailers' control over design and marketing. These findings have been replicated across apparel, footwear, electronics assembly, and horticulture.
The treadmill metaphor captures something important about the relationship between upgrading and rent. In a competitive chain, any capability that can be codified and transferred becomes a commodity, and the rent associated with it dissipates. Suppliers therefore face a continuous pressure to upgrade into activities that are harder to codify—design, brand, customer relationships, proprietary process technology. The pressure is structural, not the result of ill will on either side. Lead firms are responding to the same competitive pressure, and their response is to keep the activities that generate rent inside the firm and to outsource the activities that do not. A supplier that wants to escape the treadmill must find a way to build capabilities that are not easily codified and transferred. This is where the capability-based view becomes relevant, and where green governance enters the story.
Green capabilities have an unusual property in this respect. Many of them are process-based and codifiable—energy management systems, emissions monitoring, waste sorting—and these are exactly the capabilities that lead firms can specify and audit, which means they become commodities quickly. But some green capabilities are systemic and hard to codify: the ability to redesign a product for disassembly, the ability to coordinate a closed-loop material flow across multiple suppliers, the ability to anticipate regulatory change and position a product portfolio ahead of it. These capabilities are closer to the relational and innovation-based end of the spectrum, and they are more likely to generate sustained rent. The distinction between codifiable and systemic green capabilities is central to the argument of this book, and I return to it in Chapter 4.
2.1.3 The compliance trap
A pattern appears repeatedly in the empirical chapters: a supplier invests in meeting environmental specifications, passes the audit, and receives nothing in return. I call it the compliance trap. The trap has three components. The lead firm specifies environmental requirements in detail—materials restrictions, energy efficiency thresholds, wastewater standards, carbon reporting formats. The supplier invests in meeting the specifications, often at substantial cost, and passes the audit. The investment does not translate into a higher price, a longer contract, or a new function. The supplier has incurred a cost without acquiring a capability that differentiates it from competitors who can meet the same specification, often by purchasing the same off-the-shelf solution from the same equipment vendor.
The trap is not a market failure in the conventional sense. It is a rational outcome of a governance form in which the lead firm needs verification rather than innovation. The lead firm's sustainability team is accountable for supply chain risk, not for supplier development. Its job is to ensure that suppliers meet a standard, and the cheapest way to do that is to specify the standard precisely and audit compliance. The supplier's sustainability manager is accountable for passing the audit at minimum cost. Both parties are behaving rationally, and the result is an equilibrium in which environmental performance improves incrementally while the supplier's competitive position does not change.
Escaping the trap requires a shift in the governance relationship, and the shift is not something a supplier can accomplish unilaterally. It requires either a customer that values differentiation over verification, or a regulatory environment that rewards capabilities the supplier can build ahead of the market, or a collective action arrangement—an industry association, a cluster-level institution—that allows suppliers to invest in shared capabilities without free-riding. The Chinese context provides examples of all three, and the variation across industries is one of the empirical puzzles this book addresses. Furniture firms in Shunde, for instance, have faced a customer base that is highly price-sensitive and a regulatory environment that has tightened rapidly; their green investments have been driven largely by compliance, and the trap has been difficult to escape. Technology-intensive firms in Shenzhen and Suzhou have faced customers who value product-level environmental performance—energy efficiency in consumer electronics, for example—and have been able to convert compliance investments into product differentiation. The difference is not a difference in managerial talent. It is a difference in the governance relationship and in the structure of demand.
The trap also has a temporal dimension that the literature has underplayed. Firms that enter the compliance cycle early may accumulate organizational knowledge that later becomes useful—a measurement infrastructure, a supplier network for green inputs, a relationship with a testing laboratory. Whether this knowledge compounds into a capability or remains a sunk cost depends on whether the firm has a use for it beyond the specific audit. A firm that only ever serves one compliance-driven customer will not compound. A firm that serves several customers with different environmental requirements, or that sells into a domestic market where green product attributes are beginning to matter, has a chance to convert the knowledge into something portable. The empirical chapters track this difference.
2.1.4 Rents, quasi-rents, and the appropriability of green value
The language of rent is useful here because it connects the governance literature to the capability literature. In the original formulation, rent is the return to a resource that is scarce and not easily imitated. Ricardian rent accrues to the owner of a superior factor of production; Schumpeterian rent accrues to the innovator who temporarily monopolizes a new product or process; relational rent, in Dyer and Singh's (1998) formulation, accrues to partners who invest in relation-specific assets and develop routines for knowledge sharing. Each type of rent has a different implication for upgrading. Schumpeterian rent is the most accessible to a supplier because it depends on innovation rather than on ownership of a scarce resource. Relational rent is accessible only through a relationship with a specific partner, which makes it vulnerable to the partner's defection. Ricardian rent is the least accessible to a supplier that does not own the scarce resource.
Green capabilities can generate all three types of rent, and the type matters for whether they support upgrading. A supplier that holds a proprietary low-carbon process technology can earn Schumpeterian rent until competitors imitate it. A supplier that has co-developed a closed-loop material system with a specific customer can earn relational rent as long as the relationship lasts. A supplier that owns a renewable energy asset—a rooftop solar array, a biomass plant—can earn a quasi-rent that is partly Ricardian, because the asset is location-specific and not easily replicated by a competitor at a different site. The first two types of rent are more likely to support functional upgrading, because they involve capabilities that the supplier can deploy with other customers. The third is more likely to support process upgrading within the existing relationship, because the asset is tied to a specific facility.
The Chinese policy environment has made the third type of rent unusually accessible. Feed-in tariffs, renewable portfolio standards, and provincial subsidies for distributed solar have made it profitable for manufacturers to install rooftop solar, and the payback periods have been short enough that the investment is justified on cost grounds alone. This is a real achievement, and it has reduced the carbon intensity of Chinese manufacturing substantially. But it has not, by itself, built the capabilities that would allow a supplier to climb the value chain. The solar array is an asset, not a capability. It generates a quasi-rent, not a competitive advantage that transfers across products or customers. The distinction between asset-based and capability-based green competitiveness is one of the central analytical distinctions in this book, and I develop it in the next section.
2.2 Capability-Based View of Green Competitiveness
2.2.1 From the resource-based view to dynamic capabilities
The resource-based view of the firm, as formulated by Wernerfelt (1984) and Barney (1991), holds that sustained competitive advantage derives from resources that are valuable, rare, imperfectly imitable, and non-substitutable. The view was a corrective to the industry-organization tradition, which explained firm performance primarily by industry structure, and it had the virtue of directing attention to what firms possess rather than only to what markets do. Its weakness was that it was static. A resource that is valuable today may be worthless tomorrow, and the view did not explain how firms adapt.
Teece, Pisano, and Shuen (1997) addressed the weakness with the concept of dynamic capabilities: the firm's ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments. The concept shifted the unit of analysis from resources to routines and processes. A firm's advantage lies not in what it owns but in how it learns, how it coordinates, and how it transforms. Eisenhardt and Martin (2000) refined the concept by distinguishing dynamic capabilities in moderately dynamic markets, where they resemble traditional routines, from dynamic capabilities in high-velocity markets, where they are simple, experimental, and improvisational. The distinction matters for green capability because environmental regulation and customer expectations have been changing at different speeds in different industries. Furniture has faced a moderately dynamic regulatory environment—standards have tightened predictably, and firms have had time to adapt. Electronics has faced a faster-moving environment—materials restrictions, energy efficiency standards, and customer expectations have shifted more abruptly—and firms have had to develop more experimental approaches.
The capability literature has been extended in several directions that are relevant here. Ambrosini and Bowman (2009) distinguished "capabilities" from "resources" and argued that the distinction is often blurred in empirical work; they proposed that capabilities be understood as the routines through which resources are deployed. Helfat and Peteraf (2003) introduced the concept of the capability lifecycle, which traces a capability through founding, development, and maturity, and which helps explain why some firms fail to renew capabilities that were once strong. The lifecycle concept is useful for analyzing green capabilities because it suggests that a capability built for compliance in one regulatory regime may become a liability in another. A firm that has invested heavily in an end-of-pipe wastewater treatment system, for example, may find that the system is obsolete when the regulatory standard shifts from concentration-based to total-load-based limits, and the firm's capability in operating the old system does not transfer.
The most relevant extension for our purposes is the literature on absorptive capacity. Cohen and Levinthal (1990) defined absorptive capacity as the firm's ability to recognize the value of new information, assimilate it, and apply it to commercial ends. The concept explains why some firms benefit from exposure to external knowledge while others do not: the difference lies in prior related knowledge, which allows the firm to recognize the value of new information and to integrate it. Zahra and George (2002) distinguished potential absorptive capacity (acquisition and assimilation) from realized absorptive capacity (transformation and exploitation), and argued that firms often have more potential than realized capacity. The distinction is useful for analyzing green capability in Chinese manufacturing because many firms have been exposed to green knowledge through their customers and through government programs, but fewer have been able to transform that exposure into capabilities that affect their competitive position.
2.2.2 Natural-resource-based view and the green capability literature
The application of the resource-based view to environmental performance began with Hart (1995), who proposed a natural-resource-based view of the firm. Hart argued that environmental constraints create three strategic capabilities: pollution prevention, product stewardship, and sustainable development. Pollution prevention is about reducing emissions and waste, and it is primarily a process capability. Product stewardship is about reducing the life-cycle environmental impact of products, and it requires coordination across the value chain. Sustainable development is about addressing the social and environmental needs of stakeholders, and it requires a long-term strategic orientation. Hart's framework was important because it connected environmental performance to the resource-based view's logic of competitive advantage, and because it distinguished capabilities at different levels of the value chain.
The framework has been tested extensively. Russo and Fouts (1997) found that environmental performance was positively associated with financial performance in a sample of US firms, and that the association was stronger in high-growth industries. Judge and Douglas (1998) found that the integration of environmental issues into strategic planning was associated with financial performance. Melnyk, Sroufe, and Calantone (2003) found that firms with formal environmental management systems had better operational performance than firms without them, but that the relationship was mediated by the firm's approach to innovation. The literature has been criticized for relying on self-reported environmental performance and for failing to control adequately for industry effects, but the broad finding—that environmental capability can be a source of competitive advantage under certain conditions—has held up.
The conditions matter. Hart and Dowell (2011) revisited the natural-resource-based view and argued that the relationship between environmental performance and competitive advantage depends on the firm's ability to create value that customers are willing to pay for, and on the firm's ability to prevent imitation. A firm that reduces emissions by installing a widely available technology creates value for society but not for itself, because competitors can do the same thing. A firm that develops a proprietary process for recycling a difficult material creates value that it can capture, at least temporarily. The distinction between value creation and value capture is central to the analysis of green competitiveness, and it is often neglected in the policy literature, which tends to assume that any environmental improvement is good for the firm.
The literature on green innovation has added a further distinction between incremental and radical green innovation. Rennings (2000) argued that environmental innovation has a "double externality" problem: it produces knowledge spillovers like any innovation, and it produces environmental benefits that are not priced. The double externality means that firms underinvest in green innovation relative to the social optimum, which justifies policy intervention. The argument is important because it explains why green capability building is not simply a matter of managerial choice. Firms face a structural disincentive to invest in green innovation, and the disincentive is stronger for radical innovations than for incremental ones. The policy environment—carbon pricing, standards, subsidies—changes the incentive, and the change is what makes green capability building economically rational for firms that would otherwise free-ride.
2.2.3 Green capabilities as ordinary and dynamic capabilities
Some green capabilities are ordinary capabilities in Teece's (2007) sense: they allow the firm to perform a defined task efficiently, and they can be bought or imitated. An energy management system certified to ISO 50001 is an ordinary capability. So is a wastewater treatment plant that meets a specified standard. These capabilities are necessary for compliance, and they can be a source of advantage if the firm can operate them more cheaply than competitors, but the advantage is not sustainable because the capability is available to anyone who can pay for it.
Other green capabilities are dynamic capabilities: they allow the firm to sense and seize opportunities that are not yet fully specified, and to reconfigure its resource base as the regulatory and market environment changes. A capability for anticipating regulatory change and positioning a product portfolio ahead of it is a dynamic capability. So is a capability for coordinating a closed-loop material flow across multiple suppliers, because it requires the firm to sense what materials are available, to seize the opportunity to redesign a product around them, and to reconfigure its supplier relationships accordingly. These capabilities are harder to imitate because they are embedded in organizational routines and relationships, and they are more likely to generate sustained rent.
The distinction has a practical implication for managers. A firm that invests only in ordinary green capabilities will pass audits and meet standards, but it will not escape the compliance trap. A firm that invests in dynamic green capabilities will be able to anticipate change and to position itself ahead of competitors. The two types of investment are not mutually exclusive—a firm needs ordinary capabilities to operate, and it needs dynamic capabilities to adapt—but they have different returns and different risk profiles. Ordinary capabilities have a predictable return and a low risk of failure. Dynamic capabilities have an uncertain return and a higher risk of failure, but they are the only route to sustained advantage.
The Chinese evidence suggests that the balance between the two types of investment varies systematically across industries. Labor-intensive industries, which face price-sensitive customers and a regulatory environment that has tightened through specification, have invested primarily in ordinary capabilities. Technology-intensive industries, which face customers who value product performance and a regulatory environment that has created markets for new products, have invested more in dynamic capabilities. Capital-intensive industries, which face long asset lives and a regulatory environment that is shifting from intensity-based to absolute caps, are in an intermediate position: they have invested heavily in ordinary capabilities for process efficiency, and they are beginning to invest in dynamic capabilities for asset transition. The variation is the empirical core of Chapters 5 through 7, and the theoretical framework developed in this chapter is meant to explain it.
2.2.4 The co-evolution of institutions and capabilities
The capability literature has traditionally treated the institutional environment as a backdrop. The firm builds capabilities, and the environment selects among firms. This treatment is inadequate for analyzing green capability in China, where the state has been an active participant in capability building through industrial policy, standard setting, and public investment. The co-evolutionary perspective, developed by Nelson (1993) and elaborated by Murmann (2003), offers a better framework. It holds that technologies, institutions, and firm capabilities co-evolve, with each shaping the others over time. The co-evolutionary perspective is useful because it explains why the same firm can succeed in one institutional environment and fail in another, and why the same institutional environment can produce different outcomes in different industries.
The Chinese green transition is a case of co-evolution on a compressed timescale. The central government set the dual carbon targets in 2020, and the policy architecture—carbon markets, energy intensity targets, subsidy programs—was built in the years that followed. Firms responded to the policy signals, and their responses shaped the next round of policy. The co-evolution has been faster in some industries than in others. Solar and electric vehicles have seen rapid co-evolution, with policy creating markets, firms investing in capacity, costs falling, and policy being adjusted in response. Furniture and other traditional industries have seen slower co-evolution, with policy tightening through standards and firms responding defensively. The difference is partly a matter of technology—solar and batteries have steep learning curves, furniture does not—and partly a matter of the political economy of the industry. Industries with strong lobbying capacity and high employment have been able to shape the pace of regulation; industries with weaker political capacity have faced regulation as an external shock.
The co-evolutionary perspective also helps explain the cluster dimension of green capability. Industrial clusters are not simply agglomerations of firms; they are institutional systems that include trade associations, technical schools, testing laboratories, and local government agencies. The cluster's institutional system shapes what capabilities firms can build and at what cost. A cluster with a strong testing laboratory can support product-level environmental certification at low cost; a cluster without one forces firms to send samples to distant labs, which raises the cost and slows the certification process. The cluster dimension is the subject of Chapter 9, but it enters the theoretical framework here because it is part of the explanation for why green capability building has been uneven across regions and industries.
2.2.5 Measuring green capability: a note on method
The empirical analysis in this book relies on a mixed-methods design, and the measurement of green capability is one of the methodological challenges. Certification offers one proxy. ISO 14001 certification, ISO 50001 certification, and participation in voluntary programs such as the Carbon Disclosure Project are observable and comparable across firms. The limitation is that certification measures the presence of a management system, not the quality of its implementation. A firm can be certified and still have poor environmental performance if the system is a paper exercise. The literature on decoupling—Bromley and Powell (2012) is the canonical statement—has documented the gap between formal structures and actual practices, and the gap is likely to be larger in firms that face strong audit pressure and weak enforcement.
Emissions or energy intensity offers a second proxy. These measures are continuous and comparable, and they capture outcomes rather than formal structures. The limitation is that they are affected by factors outside the firm's control—the product mix, the age of the capital stock, the local energy grid—and that they measure performance rather than capability. A firm with a favorable product mix can have low emissions intensity without having strong capabilities.
The case studies in this book use neither proxy alone. They combine multiple indicators into a capability profile: the firm's history of environmental investment, its organizational arrangements for environmental management (whether there is a dedicated function, whether it reports to the CEO, whether it has a budget), its participation in standard-setting and industry associations, its patent activity in environmental technologies, and its relationships with customers and regulators on environmental issues. The profile is not a single number, and it is not intended to support precise comparisons across firms. It is intended to support the identification of patterns and the tracing of mechanisms.
The mixed-methods design is a response to the limitations of each approach. The econometric analysis in Chapter 4 uses certification and emissions intensity as proxies, with controls for industry, size, ownership, and region. The case studies in Chapters 5 through 8 use the capability profile to trace how firms built capabilities over time and how the capabilities affected their position in the value chain. The combination is not a substitute for a perfect measure, but it is sufficient to support the book's central claim: that green capability is an endogenous driver of upgrading, not simply a response to regulation.
2.3 Integrated Framework for Analysis
2.3.1 Why an integrated framework is needed
The three literatures reviewed above—GVC governance, industrial upgrading, and the capability-based view—each explain part of the phenomenon this book addresses, and none explains it fully. The GVC literature explains why lead firms govern suppliers the way they do, and why the governance form shapes upgrading opportunities. It does not explain why some suppliers build capabilities that allow them to escape the governance constraints while others do not. The upgrading literature explains the trajectories that suppliers follow—process, product, functional, inter-chain—and the obstacles they encounter. It does not explain the microfoundations of capability building, or why the same trajectory produces different outcomes in different firms. The capability literature explains how firms build and deploy capabilities, and why some capabilities generate sustained rent while others do not. It does not explain how the value chain context shapes the returns to capability building, or why the same capability can be valuable in one chain position and worthless in another.
The integrated framework combines the three. Its logic runs as follows. The governance form of a value chain determines the demand for green capabilities: what lead firms require, what they are willing to pay for, and what they allow suppliers to do. The supplier's existing capability base determines its ability to respond to that demand. The interaction between demand and capability determines the upgrading outcome: whether the supplier merely complies, or whether it builds capabilities that allow it to move into higher-value functions. The outcome feeds back into the governance relationship: a supplier that builds differentiated capabilities may be able to negotiate a different governance form, while a supplier that only complies remains in the captive or market form. The feedback loop is the mechanism through which green capability becomes an endogenous driver of upgrading.
Five claims follow from this logic, and the empirical chapters develop them. The first concerns demand: the green capability demand a supplier faces is a function of the lead firm's sustainability strategy and the governance form of the chain. Lead firms that treat sustainability as a compliance issue demand verifiable compliance and govern through specification and audit; lead firms that treat sustainability as a differentiation issue demand innovation and govern through relational mechanisms. The second concerns response: the supplier's ability to meet that demand depends on its existing capability base, its absorptive capacity, and its access to complementary assets such as finance, technology, and skilled labor. Suppliers with stronger bases and better access are more likely to convert demand into capability. The third concerns the outcome: when demand is for compliance and the supplier's capability is ordinary, the result is compliance without upgrading—the compliance trap. When demand is for innovation and the supplier's capability is dynamic, the result is functional or inter-chain upgrading. The intermediate cases, compliance demand with dynamic capability and innovation demand with ordinary capability, produce partial outcomes that depend on the supplier's ability to shape the relationship. The fourth concerns institutions: carbon pricing, standards, and subsidies affect what lead firms demand and what suppliers can build, and cluster institutions affect the cost and speed of capability building. Because institutions and capabilities co-evolve, the outcomes in one period shape the institutional environment in the next. The fifth concerns factor intensity: labor-intensive industries face compliance-dominated demand and ordinary capability bases, and their green investments are driven by cost and regulation; technology-intensive industries face innovation-oriented demand and dynamic capability bases, and their green investments are driven by product differentiation; capital-intensive industries face a mix of compliance and innovation demand, and their green investments are driven by asset transition. Chapters 5, 6, and 7 develop these pathways in turn.
2.3.2 The Shunde vignette: a framework in miniature
The framework is abstract, and abstractions are easier to follow when they are grounded in a concrete case. I return to the Shunde furniture cluster, which was introduced in Chapter 1 and which will be analyzed in detail in Chapter 8. The vignette is not a test of the framework; it is an illustration of how the framework's components fit together.
Shunde is a district in Foshan, Guangdong, with a furniture industry that grew rapidly in the 1990s and 2000s and that now accounts for a significant share of China's furniture exports. The cluster's firms are mostly small and medium-sized, and they sell primarily to overseas buyers—retail chains, importers, and branded marketers. The buyers set the specifications, including environmental specifications, and the firms compete on price and delivery. The governance form is close to captive: the buyers specify what they want in detail, the firms' capabilities are built around the buyers' requirements, and switching costs are high for the firms and low for the buyers.
The environmental specifications have tightened over the past decade. Formaldehyde emission standards for wood panels have become stricter, first in the export markets and then in China. VOC emission standards for coatings have tightened. Buyers have begun to ask for carbon footprint data, and some have asked for evidence of sustainable wood sourcing. The firms have responded by investing in low-formaldehyde adhesives, water-based coatings, and chain-of-custody certification. The investments have been substantial—a mid-sized firm might spend several million yuan on a coating line—and they have been driven by the need to meet the buyers' specifications.
The outcome has been compliance without upgrading. The firms have met the standards, and they have retained their orders, but their position in the chain has not changed. The buyers have not offered higher prices, longer contracts, or new functions. The firms have not developed capabilities that they can deploy with other customers or in other products. The investments have raised the fixed cost of doing business, which has accelerated the consolidation of the cluster and the relocation of the most price-sensitive segments to inland provinces and to Southeast Asia. The firms that have done best are those that have combined compliance with a move into domestic branding or into higher-value product segments—a functional upgrading strategy that depends on capabilities that are not specific to the environmental investment.
The vignette illustrates the framework's logic. The governance form (captive) generated a demand for compliance rather than innovation. The firms' capability base (ordinary, built around the buyers' specifications) allowed them to comply but not to differentiate. The interaction produced the compliance trap. The firms that escaped the trap did so by changing the governance relationship—moving into domestic branding, where the customer values different things—rather than by building green capabilities that the existing buyers would pay for. The institutional environment (Chinese standards, export market standards) shaped the demand, and the cluster's institutions (testing labs, trade associations) shaped the cost of compliance. The factor intensity of the industry (labor-intensive, low margins) limited the resources available for dynamic capability building.
The vignette also illustrates the framework's limits. It does not explain why some firms in the cluster escaped the trap and others did not. That explanation requires attention to firm-level factors—the entrepreneur's prior experience, the firm's access to finance, its position in the domestic market—that are not captured by the framework's variables. The case studies in Chapter 8 address these factors, and the framework is meant to be a guide to the case analysis rather than a substitute for it.
2.3.3 The GUM model and its relevance
The framework developed in this book draws on an earlier model that I developed for analyzing the upgrading of local industrial clusters from a GVC perspective. The model, which I called the GUM model, identifies four basic factors—cluster learning, cluster innovation, insertion in GVC, and social capital—and two auxiliary factors—upgrading-oriented public policy and dominant enterprises in the cluster. The factors interact through a bidirectional mechanism, and the interaction produces upgrading in four dimensions: technical capability, innovation capability, external linkage capability, and social capital capability.
The GUM model was developed for a different purpose—analyzing cluster-level upgrading—but it is relevant here for two reasons. First, it emphasizes that upgrading is a systemic outcome, not a firm-level outcome. The cluster's learning and innovation systems, its external linkages, and its social capital determine what individual firms can do. This emphasis is consistent with the co-evolutionary perspective developed above, and it explains why green capability building has been uneven across clusters even within the same industry. Second, the model's distinction between basic and auxiliary factors is useful for analyzing the role of policy and lead firms. Policy and dominant enterprises are auxiliary in the sense that they shape the system without being part of its core; they can accelerate or retard upgrading, but they cannot substitute for the cluster's learning and innovation capacity. The distinction helps explain why policy interventions that focus only on providing subsidies or attracting lead firms have had limited success in building green capabilities: they address the auxiliary factors without strengthening the basic ones.
The GUM model's limitation, for our purposes, is that it treats environmental performance as one of many upgrading dimensions rather than as a distinct capability with its own logic. The framework developed in this chapter addresses the limitation by specifying the mechanisms through which green capability demand is generated, how firms respond, and why the response produces different outcomes. The two frameworks are complementary: the GUM model explains the cluster-level conditions that shape what firms can do, and the integrated framework explains the firm-level and chain-level mechanisms that determine whether green capability building translates into upgrading.
2.3.4 The framework's scope conditions
The framework is not universal. It applies under conditions that should be stated explicitly, so that readers can judge whether the cases in this book are within its scope and whether it can be extended to other contexts.
The value chain must be governed by lead firms that have some discretion over supplier selection and over the terms of trade. In commodity markets where suppliers sell at arm's length and buyers have no ability to specify production conditions, the framework's demand-side mechanism does not operate. The same is true in chains where the state is the only buyer and the state's procurement rules do not include environmental criteria.
The supplier must have some ability to invest in capabilities. Firms that are financially distressed, or that operate at the margin of subsistence, cannot make the investments that green capability building requires. The framework is therefore more applicable to firms with positive margins and access to finance than to firms in survival mode. This condition matters for the labor-intensive industries analyzed in Chapter 5, where many firms are small and financially constrained.
The institutional environment must provide some signal about future environmental requirements. If the regulatory environment is entirely unpredictable, firms cannot form expectations about the returns to green investment, and the framework's co-evolutionary mechanism does not operate. The Chinese environment has been unpredictable in some respects—the timing and stringency of local enforcement have varied—but the overall direction has been clear since the dual carbon announcement, and firms have been able to form expectations about the direction of change even when they could not predict the timing.
The industry's technology must have some scope for environmental improvement. Industries with mature, standardized technologies have less scope for green innovation than industries with rapidly changing technologies. The framework is therefore more applicable to technology-intensive industries than to industries where the technology is stable and the environmental impact is determined by the scale of production.
These conditions are not binary. They are matters of degree, and the framework can be applied with appropriate adjustments to cases that meet them partially. The empirical chapters discuss the conditions in each industry and note where the framework's predictions are weaker.
2.3.5 What the framework does not explain
A framework is defined as much by what it excludesas by what it includes. Four exclusions deserve to be stated, both because they help readers judge the book's claims and because they mark directions for future work.
The framework does not explain the behavior of lead firms. It treats the lead firm's sustainability strategy as an input—something that varies across chains and that shapes the demand for supplier capabilities—but it does not explain why some lead firms adopt differentiation strategies while others adopt compliance strategies. That question belongs to the literature on corporate strategy and on the political economy of retail, and it is beyond the scope of this book. The silence is a real limitation, because the lead firm's strategy is not exogenous to the supplier's behavior: a supplier that develops a compelling green product can sometimes persuade a lead firm to change its strategy. The case studies in Chapters 6 and 7 provide examples of this reverse influence, but the framework does not theorize it.
Nor does the framework explain the macro-level determinants of the dual carbon policy. It treats the policy architecture as a context that shapes firm behavior, but it does not explain why the Chinese government adopted the targets it did, or why the policy instruments were chosen. That question belongs to the literature on Chinese political economy and on climate policy, and it is addressed in Chapter 3. Treating policy as context is a simplification, because policy is itself shaped by firm behavior—industry associations lobby, firms comply or resist, and the policy adjusts. The co-evolutionary perspective acknowledges this, but the framework does not model it formally.
The distributional consequences of green upgrading fall outside the framework as well. It focuses on whether firms build capabilities and whether the capabilities translate into upgrading, not on who gains and who loses from the process. The distributional question is important—green upgrading can displace workers, concentrate market power, and widen regional disparities—and it is addressed in the concluding chapter. The silence on distribution is a deliberate choice, made to keep the analysis tractable, but readers should not infer that the book is indifferent to the question.
Finally, the framework does not explain the environmental outcomes of green upgrading. It explains how firms build capabilities and how the capabilities affect their position in the value chain, but it does not measure whether the capabilities reduce emissions, resource use, or other environmental impacts. The omission is partly a matter of data—firm-level environmental outcomes are difficult to measure—and partly a matter of scope. The book's concern is with the relationship between green capability and competitive advantage, not with the environmental effectiveness of the green transition. The two questions are related but distinct, and conflating them would obscure the book's argument.
2.3.6 The framework in relation to alternative approaches
Three alternative approaches deserve consideration, both because they illuminate aspects of the phenomenon that the framework captures less well and because comparing them clarifies what the framework adds.
The institutionalist approach explains firm behavior primarily by the institutional environment in which firms operate. It has a long pedigree in economic sociology and in comparative political economy, and it has been applied extensively to environmental regulation. Its central claim is that firms' environmental practices are shaped by the regulatory regime, the legal system, and the norms of the business community, and that firm-level differences matter less than institutional differences. The approach explains cross-national variation in environmental performance well—why firms in Germany and Japan adopted cleaner production methods earlier than firms in countries with weaker regulation—and it has been used to good effect in studies of the Chinese environmental state. Its limitation, for our purposes, is that it does not explain why firms within the same institutional environment perform differently. The furniture cluster in Shunde and the furniture cluster in Ganzhou operate under the same national regulatory regime, but their green trajectories have diverged. The institutionalist approach can explain the divergence only by pointing to local institutional differences, which begs the question of why the local institutions differ.
The global commodity chain approach in its original formulation emphasizes the structural power of lead firms and the dependent position of suppliers. It is closely related to the GVC governance literature reviewed above, but it places more weight on the structural constraints that limit supplier agency. Its central claim is that suppliers in captive or market governance forms face a ceiling on their upgrading, and that the ceiling is set by the lead firms' interest in maintaining control over high-value activities. The approach explains why so many suppliers fail to upgrade despite substantial investments in capability building, and it is consistent with the compliance trap described above. Its limitation is that it does not explain the cases where suppliers do escape the ceiling. The Chinese evidence includes firms that have moved from assembly to design, from OEM to OBM, and from domestic supply to global brand ownership, and the structural approach has difficulty accounting for these cases without treating them as exceptions.
The entrepreneurial approach explains firm behavior primarily by the characteristics of the entrepreneur—vision, risk tolerance, networks, prior experience. It has been influential in the literature on Chinese private enterprise, and it has been used to explain why some firms in the same cluster and the same institutional environment pursue different strategies. Its central claim is that firm strategy is a matter of managerial choice, and that the choice is shaped by the entrepreneur's background and by the resources they can mobilize. The approach explains firm-level heterogeneity well, and it is consistent with the case study evidence in this book. Its limitation is that it does not explain why the distribution of entrepreneurial characteristics varies across industries and regions, or why some institutional environments reward entrepreneurship more than others.
The framework developed in this chapter combines the strengths of the three while avoiding their limitations. It takes the institutional environment seriously, but treats it as a context that shapes the demand for capabilities rather than as a determinant of firm behavior. It takes the structural power of lead firms seriously, but treats the governance form as a variable that can change as a result of supplier capability building. It takes entrepreneurial agency seriously, but situates the entrepreneur's choices within the constraints and opportunities created by the value chain and the institutional environment. The combination is not a synthesis in the sense of a grand unified theory. It is a framework that specifies the mechanisms through which the three sets of factors interact, and that generates propositions testable against the empirical evidence.
The comparison also clarifies the framework's distinctive contribution. The institutionalist approach explains why Chinese firms have improved their environmental performance; the structural approach explains why the improvement has not translated into upgrading for most firms; the entrepreneurial approach explains why some firms have upgraded. The framework explains why the translation from environmental performance to upgrading depends on the interaction between the governance form and the firm's capability base, and why the interaction produces different outcomes in different industries. The explanation is not a substitute for the three approaches; it is a way of integrating their insights into a single analytical structure.
2.3.7 The framework and the structure of the book
The framework organizes the empirical chapters that follow.
Chapter 3 analyzes the dual carbon policy architecture and the new competitive landscape it has created. The chapter applies the framework's institutional component: it shows how carbon pricing, standards, and subsidies have changed the demand for green capabilities and the returns to capability building. The chapter also introduces the concept of regulatory anticipation—the firm's ability to form expectations about future policy—which is a component of dynamic green capability.
Chapter 4 develops the concept of green governance capability and proposes a measurement approach. The chapter applies the framework's capability component: it distinguishes ordinary from dynamic green capabilities, specifies the organizational and relational conditions that support each, and presents the econometric evidence on the determinants of capability building. The chapter also tests the framework's claim that capability building is endogenous—that it is driven by the interaction between demand and the firm's existing base, not simply by regulation.
Chapters 5, 6, and 7 apply the framework to labor-intensive, technology-intensive, and capital-intensive industries respectively. Each chapter tests the framework's proposition that factor intensity shapes the pathway through which green capability affects upgrading. The labor-intensive chapter focuses on the compliance trap and on the limits of cost-based upgrading. The technology-intensive chapter focuses on innovation-driven green competitiveness and on the learning curve. The capital-intensive chapter focuses on asset transition and on the governance of low-carbon process substitution.
Chapter 8 applies the framework to the furniture industry in detail, using the Shunde vignette as a starting point and extending the analysis to the industry's export competitiveness, its exposure to carbon border adjustment, and its relocation dynamics. The chapter tests the framework's predictions about the conditions under which compliance-based green investment translates into upgrading, and it identifies the conditions under which it does not.
Chapter 9 applies the framework's cluster and institutional components. It analyzes the role of cluster institutions—testing laboratories, trade associations, technical schools—in reducing the cost of capability building, and it analyzes the multi-level governance arrangements that connect national policy to local implementation. The chapter tests the framework's proposition that cluster institutions shape the returns to capability building, and it identifies the institutional configurations that have supported successful upgrading.
Chapter 10 returns to the framework and assesses its explanatory power. It reviews the evidence for each proposition, identifies the cases where the framework's predictions were not borne out, and proposes modifications. The chapter also discusses the policy implications: what governments can do to support the transition from compliance to capability, and what they should not do.
2.3.8 The framework's testable implications
A framework that cannot be tested is not useful. Five implications follow from the argument, and the empirical chapters assess the evidence for each.
The relationship between environmental regulation and green capability building should be stronger in industries where regulation creates markets for new products than in industries where it only tightens standards for existing products. The implication follows from the claim that demand for innovation, not demand for compliance, drives dynamic capability building. The empirical test is whether firms in technology-intensive industries have built more dynamic capabilities than firms in labor-intensive industries, controlling for the stringency of regulation.
The relationship between green capability and upgrading should be stronger for firms that have access to relational governance than for firms in captive or market governance. The implication follows from the claim that the governance form determines whether capability building translates into a change in the firm's position in the chain. The empirical test is whether firms with relational customers have been more likely to undertake functional upgrading than firms with captive customers.
The relationship between green investment and financial performance should be conditional on the firm's capability base. Firms with strong capability bases should be able to convert green investment into cost savings or price premiums; firms with weak capability bases should see the investment as a cost. The implication follows from the distinction between ordinary and dynamic capabilities. The empirical test is whether the returns to green investment vary with the firm's capability profile.
The relationship between cluster institutions and green capability building should be stronger for small firms than for large firms. The implication follows from the claim that cluster institutions reduce the cost of capability building, and that the reduction matters more for firms that cannot afford to build the capabilities on their own. The empirical test is whether small firms in clusters with strong institutions have built more capabilities than small firms in clusters with weak institutions.
The relationship between carbon pricing and relocation should be conditional on the firm's capability base and on the industry's factor intensity. Firms with strong capability bases should be able to absorb the cost of carbon pricing; firms with weak bases should be more likely to relocate. The implication follows from the claim that the effect of policy on firm behavior depends on the firm's ability to respond. The empirical test is whether relocation rates vary with capability profiles, controlling for the carbon intensity of production.
These implications are not exhaustive. They are the ones the empirical chapters test most directly, and the ones the concluding chapter returns to. The framework's value will be judged by whether the implications are borne out, and by whether the cases that depart from them can be explained by the framework's scope conditions or by factors the framework does not capture.
2.3.9 The framework and the book's argument
The framework is a means, not an end. The book's argument is that green governance capability is an endogenous driver of industrial upgrading and value chain climbing, not merely a passive product of regulatory compliance. The framework supports the argument by specifying the mechanisms through which the endogeneity operates: the interaction between the governance form and the capability base, the distinction between ordinary and dynamic capabilities, the role of cluster institutions and of the co-evolution of institutions and capabilities, and the conditioning role of factor intensity.
The argument is a claim about causality, and causality claims require evidence. The empirical chapters provide it through a combination of case studies and econometric analysis. The case studies trace the mechanisms in operation and show how firms built capabilities over time and how the capabilities affected their position in the chain. The econometric analysis tests the framework's implications across larger samples and assesses the robustness of the case findings. The combination is not a definitive test—no single book can provide one—but it is sufficient to establish the argument's plausibility and to identify the conditions under which it holds.
The argument is also a claim about policy. If green capability is endogenous, then policies that focus only on compliance—standards, audits, penalties—will produce compliance without upgrading. Policies that support capability building—public investment in testing and R&D infrastructure, cluster-level institutions, demand-side incentives for green products—will produce upgrading. The policy implication is developed in the concluding chapter, and it is one of the book's contributions to the policy debate on the dual carbon transition.
The argument is finally a claim about the future of Chinese manufacturing. If the compliance trap is the default outcome for firms that face compliance-dominated demand and ordinary capability bases, then the green transition will accelerate the consolidation of traditional industries and the relocation of the most price-sensitive segments. The firms that survive will be those that build dynamic capabilities or that change their governance relationships. The claim is not a prediction—the future depends on policy choices that have not yet been made—but it is a statement about the structural pressures that the transition has created.
What the framework leaves unresolved is the question that opened the Shunde vignette and that runs through the empirical chapters: why, within a single cluster facing the same buyers and the same tightening standards, did a minority of firms escape the compliance trap while the majority did not? The framework's variables—governance form, capability base, institutional environment, factor intensity—are constant across most of the cluster, which means the answer must lie in firm-level factors the framework treats as residual: the timing of the firm's entry into the domestic market, the entrepreneur's prior experience with branded sales, the firm's access to finance at the moment the standards tightened. Chapter 8 takes up these factors directly. The theoretical work of this chapter is to have specified what the framework can and cannot explain, so that the case evidence can be read against a clear standard rather than against a general expectation that capabilities matter.