Chapter 3: Green Governance Capability: Concept, Dimensions and Endogenous Mechanisms

In the winter of 2021, I spent a week in Houjie, a town of about 400,000 people in Dongguan, Guangdong, where roughly 300 furniture factories sit along a single stretch of industrial road. The town is the hinge on which much of China's upholstered furniture exports turn. I had gone to talk to factory owners about carbon. Almost every owner I spoke with could tell me, to the yuan, what a container of sofas cost to ship to Rotterdam. Very few could tell me the carbon content of that sofa. One owner, a second-generation manager who had taken over his father's plant in 2016, put it plainly: "If a European buyer asks for a carbon number, I call my consultant. I don't know where it comes from. The consultant gives me a document. The document is what the buyer wants." He was not being evasive. He was describing an organizational fact: his firm had the accounting capability to price a sofa but not to measure it.

That gap—between the capacity to price a product and the capacity to account for its environmental footprint—is the subject of this chapter. I call the second capacity green governance capability, and the central claim I want to defend is that it is an organizational asset firms accumulate, that it varies systematically across sectors, and that under identifiable conditions it becomes a source of competitive advantage rather than a cost center. The chapter has three parts. The first defines the concept and its dimensions. The second explains the mechanisms through which green governance capability drives upgrading from within the firm and the cluster. The third connects the concept to the competitiveness debate, including the limits of what the evidence can support.

What this chapter does not argue needs stating at the outset. It does not argue that green governance capability is always good for firms, or that firms with more of it always outperform. The relationship between environmental management and financial performance is one of the most heavily studied in the management literature, and the honest summary is that the correlation is positive but small, heterogeneous across contexts, and vulnerable to reverse causality (Waddock and Graves, 1997; Wu, 2006; McWilliams and Siegel, 2000). Nor does the chapter argue that Chinese firms have already acquired this capability at scale. The evidence suggests the opposite: capability is thin, unevenly distributed, and concentrated in firms with particular structural characteristics. What the chapter does argue is that the capability is real, measurable enough to distinguish across firms, and analytically useful for explaining why some firms climb the value chain under carbon constraints while others are pushed out of it.

3.1 Conceptualizing Green Governance Capability

3.1.1 Why "Governance" and Not "Management"

The term governance is doing work here, and the choice over the more familiar environmental management is deliberate. Environmental management, as used in the ISO 14001 literature and the corporate environmentalism literature, refers to the set of practices, procedures and systems a firm uses to control its environmental impact. It is a functional category: a firm either has an environmental management system or it does not. Governance refers to the allocation of decision rights, the design of incentive structures, and the mechanisms of accountability through which environmental considerations enter the firm's strategic and operational decisions. A firm can hold a fully certified environmental management system and still make decisions that ignore carbon, because the system sits in a staff function disconnected from capital allocation. Governance is about whether environmental considerations are endogenous to the firm's decision-making, not bolted on.

I borrow the term from two literatures that rarely speak to each other. The first is corporate governance, where governance refers to the mechanisms—boards, ownership structures, incentive contracts—that align managerial action with stakeholder interests. The second is the global value chain literature, where governance refers to the authority and power relations that determine how activities are coordinated across firm boundaries (Gereffi, Humphrey and Sturgeon, 2005). Green governance capability sits at the intersection: it is the firm's capacity to coordinate environmental performance across its internal functions and across its external supply chain relationships, under conditions of incomplete information and misaligned incentives.

The framing has a practical implication. If green governance were simply environmental management, the policy prescription would be straightforward: provide more training, subsidize certification, disseminate best practice. If it is governance, the prescription is harder: it requires changing who decides, who bears the cost, and who is held accountable. That is why the capability is unevenly distributed and why it is path-dependent—it is embedded in organizational routines and authority structures that change slowly.

3.1.2 Decomposing the Capability

Any decomposition of an organizational capability into components is a modeling choice, and the choice should be justified by what it buys analytically. I use two dimensions rather than three or four because the empirical cases I have studied sort cleanly along two axes: whether the firm can produce a reliable number, and whether anyone inside the firm is answerable for that number. The first axis is measurement; the second bundles what I earlier considered separating into coordination and accountability. I keep them together because in practice they fail together. A firm that cannot make the carbon objective bind on procurement decisions almost never has a compensation system that rewards emissions reductions, and a firm that has the compensation system has already solved the coordination problem that made the system possible. Treating them as separate dimensions produces a taxonomy that is tidy on paper and misleading in the field.

Measurement capability is the capacity to produce reliable, auditable, decision-useful data on the firm's environmental flows: energy consumption, emissions, water use, waste, and the embodied carbon of purchased inputs. It sounds mundane. It is not. The Houjie factory owner who could not tell me the carbon content of his sofa was describing a measurement failure with downstream consequences for every decision that depends on that number—which supplier to choose, which product line to keep, which process to retrofit. Measurement capability requires metering at the level of the production line or the individual machine, a chart of accounts that assigns energy costs to products, and a data architecture that can aggregate across facilities and reconcile with external reporting frameworks. The CDP's own history illustrates the scale of the challenge: when it sent its first survey in 2002, it received 221 responses from 500 companies; by 2014 it had 5,003 reporting companies, and the questionnaire had grown past 100 questions spanning the full range of emissions-producing activities (CDP, 2015a, 2015b, 2015c). Reporting has grown. Whether decision-useful measurement has grown at the same rate is harder to gauge, and I suspect it lags considerably.

Measurement is also where the boundary between firm and supply chain becomes porous. A manufacturing firm's Scope 3 emissions—those embodied in purchased inputs and sold products—typically dwarf its Scope 1 and 2 emissions, but they are measured by suppliers, not by the firm. Measurement capability therefore includes the capacity to elicit, verify and aggregate data from suppliers with their own, often weaker, systems. This is a governance problem disguised as a data problem. The firm cannot simply demand the number; it has to build the supplier's capacity to produce it, accept an estimate, or exclude the category. Each choice affects the credibility of the firm's own disclosure.

The second dimension, coordination and accountability, is the capacity to align environmental objectives with cost, quality, delivery and innovation across functions and across the supply chain, and to assign responsibility for results with consequences attached. The classic failure mode is the sustainability report that commits to a 2030 target while procurement continues to optimize on unit price alone. The two functions are not talking, or they are talking but their incentives point in different directions. Making the carbon objective bind on the people who make operational decisions means embedding it into procurement criteria, capital budgeting hurdle rates, product design gates and supplier scorecards. It means giving the sustainability function enough authority to say no to a sourcing decision, or enough analytical credibility to change the terms of the decision.

The renewable energy sourcing menu offers a clean illustration. A firm that wants to reduce its Scope 2 emissions can buy unbundled energy certificates, sign a green pricing program, take a green tariff, sign a power purchase agreement (PPA), or own renewable assets outright. Each option carries a different profile of additionality, credibility, capital cost and risk (Table 1). Unbundled certificates are cheap and simple but lack additionality and are vulnerable to double counting. Ownership internalizes cost savings and ensures additionality but exposes the firm to basis risk and requires upfront capital and maintenance capacity. IKEA's decision to invest 2.5 billion euros in on- and off-site renewable projects over a decade was a coordination decision as much as an investment decision: it required building project development and asset management capabilities that a furniture retailer does not naturally possess. A firm that buys certificates has made a different coordination choice, one that outsources the capability to the certificate market. Neither choice is wrong. The choice reveals what the firm is organized to do.

Table 1: Mechanisms for sourcing renewable electricity and their governance implications

Source: synthesized from corporate renewable procurement literature; see also the discussion of IKEA's ownership strategy.

The two dimensions are analytically distinct but empirically correlated. A firm that cannot measure cannot coordinate, because it has no common metric around which to align functions. A firm that cannot coordinate cannot hold anyone accountable, because responsibility is diffuse. The causal ordering runs roughly from measurement to coordination, but the relationship is recursive: accountability mechanisms create demand for better measurement, and coordination failures reveal measurement gaps. In the firms I have studied, the binding constraint is usually measurement, but the binding organizational constraint is usually accountability—the absence of anyone whose job depends on the number being right.

3.1.3 What the Concept Is Not

Three clarifications will save confusion later. Green governance capability is not the same as green performance. A firm can have high capability and poor performance, because its production process is inherently carbon-intensive, or low capability and good performance, because it happens to use a clean process. The capability is the capacity to act, not the outcome of action. This distinction matters for policy: subsidizing outcomes rewards firms that are already clean, while building capability changes what firms can do in the future.

The concept is also not sector-neutral in its content, even though it is sector-neutral in its definition. Measurement and coordination look different in a labor-intensive furniture cluster than in a capital-intensive cement plant. In furniture, the dominant measurement challenge is Scope 3—the embodied carbon of wood, foam, fabric and hardware sourced from dozens of small suppliers. In cement, the dominant challenge is process emissions from calcination, measured at the kiln and not reducible by switching suppliers. The dimensions are the same; the content differs. This is the analytical basis for the sector-differentiated pathways the later chapters develop.

Finally, green governance capability is not a fixed endowment. It is accumulated through learning and depreciates if not maintained. The learning is partly experiential—firms learn by doing, by making mistakes, by responding to buyer requests and regulatory inspections. It is partly vicarious—firms learn from peers, consultants and industry associations. And it is partly deliberate—firms invest in training, systems and personnel. The accumulation is path-dependent: early investments shape later options, and late starters face higher catch-up costs. This is why the capability is unevenly distributed even within the same cluster and the same product market.

3.1.4 A Note on Measurement

Can green governance capability be measured? Not directly, and I am skeptical of attempts to reduce it to a single index. In my own fieldwork I have used a composite of proxies supplemented by interviews, and the composite is crude. It can distinguish firms at the top and bottom of the distribution. It struggles in the middle, where most firms sit.

The proxies themselves deserve scrutiny, because each measures something slightly different from the capability. Certification—ISO 14001, ISO 50001, industry-specific schemes—is a binary indicator of the existence of a management system, not of its effectiveness, and the ISO 14001 literature has repeatedly found certification decoupled from practice. Disclosure—CDP scores, sustainability report content, third-party assurance—measures what firms say, which correlates with but is not identical to what they do. Target-setting—science-based targets, net-zero commitments, interim milestones—measures ambition, not capability. Organizational structure—a board committee on sustainability, a Chief Sustainability Officer, environmental criteria in compensation—measures the formal allocation of responsibility, a precondition for accountability but not a guarantee of it. A firm that scores well on all four families is probably capable. A firm that scores well on one is probably performing for an audience.

The quantitative results I report in later chapters should be read with that limitation in mind. The concept is analytically useful because it organizes observable differences; it is not a precision instrument.

3.2 Endogenous Drivers and Mechanisms

The central theoretical claim of this book is that green governance capability is endogenous to the upgrading process—a driver of change from within the firm, not merely a reaction to pressure from outside. "Endogenous" can mean everything and nothing, so I will specify three things I mean by it. The capability is generated by the firm's own activities, particularly its engagement with global value chains, rather than supplied exogenously by regulators or donors. Once generated, it changes the firm's production function and its position in the value chain, so causality runs in both directions. And it is subject to increasing returns—the more a firm has, the cheaper it is to acquire more—which is why its distribution is skewed and why policy intervention can have leveraged effects.

Four mechanisms carry the endogenous dynamic: learning by supplying, the modularity of green tasks, the cluster as a learning environment, and the financialization of environmental performance. They interact, and separating them is a matter of analytical convenience rather than ontological claim.

3.2.1 Learning by Supplying

The most familiar mechanism from the GVC literature is that firms learn to meet environmental requirements by supplying buyers who impose them. The learning is not automatic. It depends on the nature of the buyer-supplier relationship, the specificity of the requirement, and the supplier's absorptive capacity. When the conditions are right, the requirement functions as a forcing device that pushes the supplier to build measurement and coordination capacity it would not otherwise have built.

Take the furniture case. A European buyer subject to the EU Timber Regulation and, increasingly, to corporate sustainability reporting requirements under the Corporate Sustainability Reporting Directive must demonstrate that the wood in its products is legally harvested and, in some cases, that the product's embodied carbon falls below a threshold. The buyer passes the requirement to the Chinese supplier as a questionnaire, an audit, or a specification. To keep the order, the supplier must produce the data. Producing the data requires mapping its own supply chain—which sawmill the wood came from, what the species is, what the chain of custody looks like. The Chinese domestic market did not demand this, so the supplier would not have undertaken the measurement on its own. The buyer's requirement creates the demand; the supplier's response builds the capability.

Learning runs deeper when the requirement is an ongoing relationship rather than a one-off questionnaire. A supplier audited annually for five years has built routines—a designated compliance officer, a document management system, a relationship with a certification body—that persist after the audit. Those routines are the capability. They can be redeployed to meet a different buyer's requirement, or a domestic regulatory requirement, at lower marginal cost. This is the sense in which the capability is endogenous: it is a byproduct of the supplier relationship that becomes an asset in its own right.

The mechanism has limits worth naming. Learning by supplying is weakest when the buyer's requirement is transactional—a questionnaire filled out once, an audit passed and forgotten. It is strongest when the requirement is embedded in a long-term contract with consequences for failure. It is also weakest when the supplier is small, with no slack resources and no absorptive capacity; the requirement then becomes a barrier to entry rather than a learning opportunity, and the supplier exits the export market instead of upgrading. Chapter 2 described this double-edged sword at the industry level; here it reappears at the level of the individual firm.

A further limit deserves more attention than the GVC literature has given it. Learning by supplying builds the capability to meet the buyer's requirement, which is not necessarily the capability to reduce emissions. A supplier can become expert at producing documentation that satisfies an audit without changing its production process. The literature on decoupling—the gap between formal compliance and substantive practice—is relevant, and I take it seriously. The capability built may be a compliance capability rather than an environmental one. Whether the two converge depends on whether the buyer's requirement rewards substantive performance or formal documentation. That is a design question for buyers and regulators, and one of the places where public policy can shape private governance.

3.2.2 The Modularity of Green Tasks

Less discussed, and more important for explaining why the capability accumulates rather than dissipates, is the modularity of the tasks involved in environmental improvement. A modular task can be decomposed into sub-tasks, each standardized, taught and reassembled. Modularity is what allows a capability to scale and transfer rather than remaining tacit and firm-specific.

Green tasks are unusually modular, for a reason rooted in the physics of emissions. An emission is a mass flow: so many kilograms of CO2 per unit of output. It can be measured at a boundary, attributed to a process, and compared across firms and products. That makes it possible to standardize the measurement task (the GHG Protocol), the reporting task (CDP, GRI), the target-setting task (science-based targets) and the verification task (third-party assurance). Each standard is a module. A firm that adopts the modules builds capability faster than a firm that reinvents the measurement architecture for itself.

The modularity has a downside. Because the modules are standardized, they can be adopted superficially. A firm can hire a consultant to fill out the CDP questionnaire without building any internal measurement capacity. The module becomes a black box. This is the "consultant gives me a document" problem the Houjie owner described. The capability is not internalized; it is outsourced. The firm can report but cannot act.

Firms that internalize the modules—that build the measurement system in-house, train their own staff to run it, use the data for decisions—are the ones that accumulate capability. The difference between the two types is not the module adopted but the depth of the organizational change accompanying adoption. Cohen and Levinthal's (1990) concept of absorptive capacity applies directly: the firm's ability to recognize the value of new information, assimilate it, and apply it commercially. A firm with high absorptive capacity can take a standardized module and make it its own. A firm with low absorptive capacity can only adopt it formally.

Modularity also explains why the capability spreads faster inside a cluster than across a whole sector. Within a cluster, firms share suppliers, customers, labor and information. A module that works in one firm is visible to its neighbors, and adoption is cheaper because the complementary investments—training, recruitment, supplier relationships—already exist in the local labor market.

3.2.3 The Cluster as a Learning Environment

My earlier work on the internationalization of industrial clusters argued that clusters are not simply agglomerations of firms; they are environments in which knowledge about production, markets and organization circulates through labor mobility, supplier relationships and informal exchange. The same logic applies to green governance capability. A cluster that has begun to build the capability generates positive externalities for its members: a pool of workers who know how to run an emissions inventory, local consultants who understand the sector's measurement problems, a norm of disclosure that makes non-disclosure conspicuous, and shared suppliers who have themselves been pushed to measure.

The evidence for this mechanism is mostly qualitative. In the furniture clusters I have studied, diffusion is uneven and often mediated by the largest firms. A large exporter pushed by a European buyer to build a chain-of-custody system will require its local suppliers to adopt compatible systems, which diffuses the capability down the local supply chain. The diffusion is not altruistic; a supplier that can produce the data is cheaper to audit than one that cannot. The effect is to raise the capability floor of the cluster.

The failure mode is worth naming. Clusters can lock into a low-capability equilibrium when leading firms have no incentive to diffuse and followers have no capacity to adopt. That is more likely when the cluster's export market does not demand environmental performance, or when the leading firms are foreign-invested and keep their environmental systems in-house. The cluster then reproduces its existing capability level and the gap between leaders and laggards widens. Cluster-level interventions—shared measurement infrastructure, collective certification, training programs—can shift that equilibrium in a way firm-level interventions cannot.

The cluster literature has been criticized, fairly, for romanticizing local economies and underplaying extra-local connections. The clusters I study are not self-contained; they are nodes in global value chains, and the most important learning often comes from the buyer relationship rather than from the neighbor. The cluster matters because it determines how quickly learning from the buyer relationship diffuses to other firms. Strong internal linkages amplify the learning; weak linkages confine it to the firms directly connected to the buyer.

3.2.4 The Financialization of Environmental Performance

The most recent mechanism, and the least studied, concerns the growing role of financial actors—investors, lenders, insurers—in defining what counts as good environmental performance and allocating capital accordingly. The CDP's original design is instructive: it contacts companies to report on behalf of investors, and the annual disclosure gives investors a resource for long-term analysis. The reporting is not primarily for regulators or consumers. It is for capital markets. This changes the incentive structure. A firm that reports well and performs well may get cheaper capital; a firm that reports poorly may be excluded from an investment universe or charged a higher risk premium.

The financialization mechanism is endogenous in a specific sense: it creates a feedback loop between environmental performance and the cost of capital that did not exist when environmental performance was purely a regulatory or reputational matter. The loop is weak in China relative to Europe, because the domestic investor base is less sensitive to environmental disclosure and the green bond market, while large, has been criticized for weak additionality. But the loop is strengthening, and it is one channel through which the dual carbon goals reach firms. A Chinese exporter that wants to list on a European exchange, borrow from a European bank, or sell to a European corporate buyer with a science-based target faces the financialized version of the requirement whether or not it faces the Chinese regulatory version.

The distributional consequence will occupy later chapters. Firms with access to capital markets can use the loop to their advantage; firms without access cannot. The former build capability faster because the return on capability is higher for them. The latter are pushed toward the exit. Carbon constraints therefore accelerate consolidation within sectors, which is a reason to be cautious about the claim that green governance is uniformly good for Chinese manufacturing. It is good for the firms that can build it, and a barrier for the firms that cannot.

3.2.5 The Mechanisms in Combination

The four mechanisms do not operate independently. A firm that learns by supplying a demanding buyer also gains access to the modular tools the buyer requires; if it sits in a strong cluster, the learning diffuses to its neighbors; if it has access to capital markets, the financial loop rewards it. The mechanisms reinforce each other, which is why the distribution of capability is skewed and why the returns to early investment are high.

The combination also explains why the capability is path-dependent in a way that is not simply a matter of firm age or size. A small firm in a strong cluster with a demanding buyer can build capability faster than a large firm in a weak cluster with a domestic-oriented customer base. The determinant is not the firm's resources in isolation but its position in the network of relationships through which the capability is generated. This is the analytical payoff of treating the capability as governance rather than management: it is relational, not atomistic.

A caveat about causality closes this section. The mechanisms are plausible and consistent with the case evidence, but the case evidence cannot establish that they are the only mechanisms, or that their effects are large relative to confounders. Firms that build green governance capability may be firms that were already more capable in other dimensions—better managed, more export-oriented, more likely to invest in training. The observed correlation between capability and upgrading may reflect selection rather than causation. The empirical chapters use variation across sectors and clusters to narrow the range of plausible interpretations. For now, the claim is that the mechanisms are real and identifiable, not that they are dominant.

3.3 From Compliance to Competitiveness

The competitiveness debate has a canonical form. The Porter Hypothesis (Porter and van der Linde, 1995) argues that well-designed environmental regulation can spur innovation that offsets compliance costs and enhances competitiveness. The neoclassical counterargument is that regulation imposes costs firms would not voluntarily incur, and that the innovation offset is at best partial. The empirical literature supports a weak version of the Porter Hypothesis—regulation can induce innovation—and much weaker support for the strong version, in which innovation offsets the costs (Ambec et al., 2013). Chinese evidence fits the pattern: environmental regulation has induced innovation in some sectors and imposed net costs in others, with the balance depending on regulatory design and firm characteristics.

Relitigating the Porter Hypothesis is not my purpose. The narrower question for this book is: under what conditions does green governance capability translate into competitive advantage for Chinese manufacturing firms in global value chains? Four conditions, each grounded in the mechanisms above, define the answer. They are not a checklist. They interact, and a firm that satisfies three of the four may capture nothing.

3.3.1 Rarity and Depth

A capability generates competitive advantage only if it is rare and difficult to imitate. Measurement capability, once standardized, is neither. Any firm can hire a consultant to run the GHG Protocol calculation. The advantage comes from the depth of the capability—the ability to measure at a granularity competitors cannot, to use the data in decisions competitors do not, to integrate measurement with product design and procurement in ways that are organizationally demanding. The Houjie owner who outsourced measurement to a consultant had a capability that was neither rare nor hard to imitate; it was a purchased service. The firm that built the measurement system in-house, trained its engineers to use it, and redesigned its product line around the results had something a competitor could not buy off the shelf.

This condition has an uncomfortable implication for policy. Subsidies for certification—paying firms to obtain ISO 14001, for example—may increase the quantity of certification without increasing the depth of capability, because certification is imitable by design. The advantage comes from what the firm does with the certification, not from the certificate. This is consistent with the finding in the ISO 14001 literature that certification alone is a weak predictor of environmental performance.

3.3.2 Buyer Valuation

A firm-specific capability no customer values is a hobby, not a competitive advantage. The value of green governance capability depends on the structure of the market the firm sells into. In the European market, where corporate buyers face CSRD reporting requirements and consumer-facing brands face reputational risk, the capability is valued, and the firm can capture some of the value through price, contract length, or preferred-supplier status. In the domestic Chinese market, where environmental attributes are less visible to consumers and procurement is more price-driven, the capability is valued less, and the firm may bear the cost without capturing the benefit.

This is the core of the sector-differentiated argument the book develops. The same capability has different returns in different markets. A furniture firm selling to IKEA or a European importer faces a buyer who values the capability; a furniture firm selling to a domestic wholesaler does not. The export-oriented segment of the industry therefore has a stronger incentive to build capability than the domestic-oriented segment, and the capability gap between the two widens over time. This is not a story about Chinese firms versus foreign firms. It is a story about market segments within China, and it has implications for the distributional effects of the dual carbon goals.

3.3.3 Complementary Assets

Green governance capability is rarely valuable on its own. It is valuable when combined with other assets—design capability, brand, supply chain relationships, manufacturing scale—that allow the firm to convert the capability into a product or a contract. A firm that can measure the carbon content of its sofa but cannot design a low-carbon sofa has a measurement capability without a product. A firm that can design a low-carbon sofa but cannot measure it cannot prove the claim to the buyer. The advantage comes from the combination.

This condition explains why the returns to green governance capability are higher for larger, more sophisticated firms. They have the complementary assets—R&D, design, marketing, supplier management—that make the capability productive. Smaller firms may have the capability but lack the complements, and it sits idle. Capability-building programs targeted at small firms may therefore fail unless bundled with complementary support—design assistance, market access, supplier development—that lets the capability be used.

3.3.4 Organizational Durability

The final condition is organizational. A capability that depends on a single champion—a sustainability manager who leaves, a CEO who retires—is fragile. The capability is sustained when embedded in routines, incentives and authority structures that survive personnel changes. This is the accountability dimension of the concept, and it is the one Chinese firms most often lack. Many firms have a sustainability report and a target; few have a board committee that reviews progress, a compensation system that rewards it, or an internal audit function that verifies the data.

The fragility is not unique to Chinese firms. It is a general problem of organizational change, more acute in firms that are growing fast, family-owned, or not yet professionally managed. The Houjie factory owner who relied on a consultant had a governance structure in which the capability was located outside the firm. The firm that builds the capability inside—through hiring, training and the redesign of decision rights—has a structure that can sustain it.

3.3.5 The Competitiveness Payoff: What the Evidence Shows

What do we actually know about whether green governance capability pays off for Chinese firms? The evidence is thinner than the policy rhetoric suggests.

The first body of evidence concerns the effect of environmental regulation on firm performance. The Chinese literature has grown rapidly since the 2015 Environmental Protection Law and the 2017 launch of the national emissions trading scheme pilot. Findings are mixed. Some studies find that stricter regulation reduces profitability in the short run and improves it in the long run, consistent with a dynamic Porter effect. Others find the effect negative and persistent, particularly for small firms in pollution-intensive sectors. The heterogeneity is the story: the average effect masks a distribution in which some firms gain and others lose, and the losers are concentrated among firms with low capability.

The second body concerns the effect of green capability on export performance. The literature on environmental standards and trade has long debated whether standards are barriers or catalysts. Chinese evidence suggests the answer depends on the firm's position in the value chain. Exporters supplying global buyers with environmental requirements have invested in capability and maintained or increased market share; exporters supplying price-sensitive markets have not, and have lost share as compliance costs rise. This is consistent with the sector-differentiated argument, and it is one of the empirical patterns the later chapters examine in detail.

The third body concerns industrial relocation. The policy concern is that carbon constraints will push carbon-intensive industries out of China to countries with weaker regulation, a form of carbon leakage. The evidence so far is mixed. Some energy-intensive industries have relocated, but the relocation is driven by a combination of factors—labor costs, land costs, trade tensions—of which carbon is one. The firms that relocate tend to be the ones with low capability, consistent with the view that carbon constraints accelerate the exit of marginal firms rather than driving entire industries offshore.

The evidence does not show that green governance capability is a reliable path to competitiveness for all firms. It shows that the capability is associated with better outcomes for firms that have the complementary assets and market access to use it. For firms lacking those complements, the capability is a cost without a return, and the rational response may be to exit the export market rather than upgrade. The policy debate often obscures this distributional reality.

3.3.6 The Limits of the Argument

Three limits bound the argument, and the credibility of the book depends on stating them plainly.

The first is causal. The evidence reviewed here is consistent with the claim that green governance capability drives upgrading, but it is also consistent with the claim that upgrading drives capability, or that both are driven by a third factor such as management quality or export orientation. Case evidence cannot distinguish these hypotheses. The empirical chapters use variation across sectors and clusters to narrow the range of plausible interpretations, but they do not resolve the identification problem. The causal claim is a working hypothesis supported by mechanism evidence, not a settled finding.

The second is scope. The book focuses on manufacturing firms in global value chains. The mechanisms may not apply to firms that sell only domestically, or to sectors where environmental performance is not valued by buyers. The furniture case is instructive precisely because it sits at the boundary: the export segment faces the requirement, the domestic segment does not, and the divergence between the two is the object of study. Generalizing from the export segment to the whole industry would be a mistake.

The third is normative. The book argues that green governance capability is a source of competitive advantage under certain conditions. It does not argue that the conditions are just, or that the distribution of capability is socially optimal. The firms that build capability are the firms that already have resources, market access and management depth. The firms that do not are pushed toward exit. Whether that is desirable depends on values outside the scope of the analysis—whether one prioritizes aggregate productivity growth or the survival of small firms, whether one accepts consolidation as the price of decarbonization or regards it as a cost to be mitigated. I have views on these questions, but they are not derived from the evidence, and I keep them separate from the analysis.

Chapter 4 takes up the dual carbon goals as a competitive landscape, using the concept developed here—green governance capability as an endogenous, relational, path-dependent asset—as the lens. It examines three sectors in sequence: furniture, where the export segment's measurement capability is being tested by European buyer requirements; textiles, where the buyer pressure is weaker and the capability gap between export and domestic segments is wider; and photovoltaics, where the capability is embedded in the product itself and the competitive dynamics run in the opposite direction. The chapter draws on firm-level survey data from 214 manufacturers in Guangdong, Zhejiang and Jiangsu, collected between 2021 and 2023, and on 38 interviews with factory managers, buyers and certification bodies. The question it asks is not whether the dual carbon goals will impose costs on Chinese manufacturing. They will. The question is which sectors have the organizational foundations to convert those costs into capabilities, and which do not.