Chapter 3: Green Governance Capability as an Endogenous Driver of Value Chain Climbing

The container ship left Yantian at 03:40. By sunrise over the South China Sea, the forty-foot boxes in Bay 22 were six hours into a twenty-eight-day passage to Rotterdam. Inside: 1,200 flat-pack desks, each carton carrying a formaldehyde emission certificate, a QR code linking to a batch-level carbon footprint declaration, and an FSC logo burned into the underside of the tabletop. The buyer, a Dutch office furniture retailer with 340 stores across Europe, had specified all three markings in a 47-page purchasing manual with seven annexes. Two of those annexes did not exist five years ago.

The factory that made the desks sits in Lecong, a town of 180,000 west of Guangzhou where the furniture industry occupies 4.7 square kilometers of workshops, showrooms, and loading yards. Twenty years ago the competitive logic here was simple: cheaper particleboard, thinner veneer, faster assembly, lower wage. The town exported volume. Now the same factories export documentation. The manager I spoke with in 2023, a man in his fifties who started as a sander on the night shift, put the change in a phrase I have since heard in different forms from exporters in Ningbo, Dongguan, and Suzhou: "Before, we sold boards. Now we sell proof."

That sentence contains the argument of this chapter.

The conventional story about environmental standards in global value chains casts firms like this one as passive respondents. Regulators in Brussels or buyers in Rotterdam set the rules; Chinese suppliers adjust or exit. Compliance is a cost, and the firm that complies has absorbed a tax on its own competitiveness. The story is not wrong. It is incomplete in a way that matters for how we understand industrial upgrading.

What the Lecong manager described is something else: the accumulation of organizational capabilities—measurement, management, verification, innovation—that had begun to function as an asset in its own right. The certificate was not a permission slip. It was evidence of a capability the buyer could not easily find elsewhere, and that capability had started to shape who got which orders, at what price, and on what terms.

This chapter defines green governance capability, decomposes it into its components, and traces the mechanisms through which it translates into value chain position. It then confronts the central analytical question: is green capability endogenous to the upgrading process, or is it a response to external pressure that would not exist otherwise? The two are not alternatives. They are phases in a sequence, and the sequence is where the interesting economics lives.

3.1 Defining and Decomposing Green Governance Capability

3.1.1 Why "capability" and not "compliance"

The literature has tended to blur a distinction that matters.

Compliance is the state of meeting a specified requirement. It is binary, externally defined, and verifiable at a point in time. A factory either holds the certificate or it does not. A shipment either clears customs or it does not. Compliance is a condition.

Capability is the ability to meet a class of requirements—including requirements not yet specified—repeatedly and at reasonable cost. It is a capacity, not a state. It resides in routines, in data systems, in the tacit knowledge of engineers and auditors, and in the relationships that connect the firm to its suppliers and customers. A firm with capability can respond to a new standard in months; a firm without it needs years, or cannot respond at all.

Policy and management research has mostly measured compliance and inferred capability. That inference is unreliable. Two firms can both hold ISO 14001 certification while one has embedded environmental data into its production planning and the other has a binder on a shelf. The certificate is identical. The capability is not.

Consider what a buyer actually needs to know when it sources from a Chinese factory under the European Union's Carbon Border Adjustment Mechanism, phased in from 2023 with reporting obligations and financial obligations from 2026. The buyer needs embedded emissions data at the product level, verified by an accredited body, consistent with the EU's methodological rules, and delivered on a schedule that matches the buyer's own reporting cycle. A factory that can produce this data is not simply compliant. It has built an emissions accounting system, trained staff to operate it, integrated it with the enterprise resource planning system, and established a relationship with a verification body. That is a capability. It took investment and time. It cannot be bought off the shelf, and it does not disappear when the regulation changes.

The distinction has a further implication the compliance framing obscures. Compliance is a cost that scales with the stringency of the requirement. Capability is an asset that generates returns across multiple requirements. A firm that has built product-level carbon accounting for CBAM can also serve buyers who demand carbon labels, can respond to Chinese domestic carbon market obligations, can qualify for green finance, and can price its products with reference to a carbon cost that competitors cannot see. The same investment serves four markets. The compliance framing sees four costs; the capability framing sees one asset with four revenue channels.

This is the analytical move the chapter makes, and it changes what we predict about which firms upgrade, which exit, and which capture the rents that green standards create.

3.1.2 The four components

Green governance capability, as I use the term, has four components. They are analytically distinct but empirically intertwined, and the sequencing of their development matters for the upgrading outcomes I trace later.

Measurement. The ability to quantify environmental performance at the level of the product, the process, and the facility, in a form that is comparable across time and consistent with external methodologies. Measurement includes the technical infrastructure (meters, sensors, laboratory capacity), the data architecture (how information flows from the shop floor to the reporting system), and the methodological knowledge (which emission factors to use, how to allocate shared costs, how to handle boundary questions). Measurement is the foundation. Without it, the other three components have nothing to work on.

The Chinese furniture sector shows how measurement capability gets built, often for reasons unrelated to export. The national standard GB/T 50378, the Green Building Evaluation Standard enacted in 2014 by the Ministry of Housing and Urban-Rural Development, evaluates conservation of energy, water, and materials across whole building life cycles. For furniture makers, the relevant measurement infrastructure arrived by a different route: the formaldehyde emission standards (GB 18580) and the voluntary environmental labeling program, which required factories to install testing laboratories and document material inputs at the batch level. Factories that invested in this infrastructure for the domestic market found, somewhat unexpectedly, that it positioned them for European buyers' requirements. The measurement capability was built for one purpose and repurposed for another. That is characteristic of capability: it generalizes.

Management. The ability to organize production and procurement so that environmental performance is a design parameter rather than an afterthought. Management capability shows up in supplier selection criteria, in product design reviews, in the routing of materials through the plant, and in the incentive structures that determine whether the line supervisor cares about scrap rates. It is the difference between a factory that measures its emissions and a factory that reduces them.

Van Weele and van Tubergen (2014), in their study of responsible purchasing, describe the mechanism precisely. When a buyer imposes environmental requirements on incumbent suppliers, it reduces the suppliers' product and process solution space. The supplier must find new products and process solutions that fit within the smaller space. That constraint is a cost in the short run. In the medium run, it forces a search process the supplier would not have undertaken on its own. The result, in their account, is innovation—not because the supplier was trying to innovate, but because the constraint made the existing solution infeasible. Management capability turns that constraint into a design problem. A firm without it experiences the constraint as a loss.

Verification. The ability to demonstrate, to a third party, that the measurement is accurate and the management is effective. Verification includes internal audit systems, documentation practices, and the relationships with external certifiers that determine how smoothly the audit process runs. It is the component that converts private information into credible public information—and the conversion of private information into credible public information is the central mechanism through which green capability translates into bargaining power.

Verification capability is often underestimated because it looks like paperwork. It is not. A factory that can pass a carbon verification audit without disrupting production has built something that took years: a document trail that matches physical reality, staff who understand what auditors look for, and a quality management system that produces the evidence as a byproduct of normal operations. Factories that lack this capability experience audits as crises. Factories that have it experience audits as routine. The difference shows up in the buyer's risk assessment, and the buyer's risk assessment shows up in the price.

Innovation. The ability to develop new products, processes, and business models that improve environmental performance beyond the current requirement. Innovation capability determines whether a firm climbs or merely complies. A firm that can only meet the standard will be squeezed as the standard rises. A firm that can exceed the standard has something to sell that competitors cannot easily replicate.

The eco-innovation literature has established that this capability is not evenly distributed and not easily acquired. Klewitz and Hansen (2014), in their systematic review of sustainability-oriented innovation in small and medium enterprises, find that SMEs face distinctive barriers: limited absorptive capacity, thin managerial bandwidth, and difficulty accessing the technical knowledge that innovation requires. Lee and Min (2015) show, using firm-level data from the Korean manufacturing sector, that green R&D has a measurable positive effect on both carbon emissions reduction and firm performance—but the effect is conditional on the firm having the internal capacity to convert R&D spending into commercial outcomes. The innovation component is the most demanding of the four. It presupposes the other three.

3.1.3 How the components interact

The four components are not a checklist. They are a system, and the system has properties the components do not have individually.

Measurement without management produces data that nobody uses. This is the most common failure mode in Chinese factories that have invested in environmental monitoring equipment: the meters are installed, the data flows into a database, and the database is consulted only when an auditor arrives. The investment has been made, but the capability has not been built, because the data does not feed back into decisions.

Management without verification produces claims that buyers discount. A factory that has genuinely reduced its emissions but cannot prove it is in a weaker bargaining position than a factory that has reduced them less but can prove it. The market for green attributes is a market for credible information, and credibility requires verification.

Verification without innovation produces a ceiling. A firm that can prove it meets the standard will be paid the standard price. A firm that can prove it exceeds the standard has a claim on a premium. The premium is where the rents are.

Innovation without measurement is blind. A firm that develops a new process without the ability to quantify its environmental performance cannot know whether the process is actually better, cannot document the improvement for buyers, and cannot defend the claim against a competitor's challenge.

The interactions run in both directions. Strong measurement capability lowers the cost of verification (the data is already there) and raises the return to innovation (the improvement can be documented). Strong management capability raises the return to measurement (the data will be used) and lowers the cost of innovation (the search process is organized). The system has increasing returns to the components, which is why capability, once built, tends to persist and why firms that start building early tend to stay ahead.

There is a sequencing logic here that the case evidence supports. Firms typically build measurement first, because it is the component most directly demanded by external requirements. Management follows, often reluctantly, as the firm discovers that measurement without management generates data it cannot act on. Verification is built in response to a specific buyer or regulator, and it tends to arrive as a discrete investment rather than a gradual accumulation. Innovation comes last, and many firms never get there. The firms that do—the ones that cross from compliance to capability—are the ones that matter for the upgrading story.

3.1.4 Measurement logic: why capability is hard to observe

If green governance capability is this important, why has it been so hard to measure? The most important components are intangible, and the tangible proxies are noisy.

The standard proxies in the literature are certification (ISO 14001, ISO 50001), environmental management system adoption, environmental disclosure, and emissions intensity. Each captures something, and each misses something.

Certification captures the existence of a management system but not its quality. The variation in ISO 14001 implementation is enormous. Some certified firms have integrated environmental management into their operating routines; others have a consultant-maintained binder. The certificate cannot distinguish between them.

Emissions intensity captures outcomes but confounds them with factors outside the firm's control. A factory that produces a low-carbon product may have high process emissions; a factory with efficient processes may produce a carbon-intensive product. Intensity is a ratio, and ratios are hard to compare across firms with different product mixes.

Disclosure captures the firm's willingness to report but not the accuracy of the report. In markets where disclosure is voluntary, the firms that disclose are selected—they are the ones with something to show. In markets where disclosure is mandatory, the reports vary in quality, and the variation is correlated with capability in ways that are hard to observe.

The measurement problem is not incidental. It is the reason the compliance framing has persisted: compliance is observable, capability is not, and the literature has tended to study what it can measure. The result is a systematic underestimate of the role that capability plays in upgrading, and a systematic overestimate of the role that external pressure plays.

I do not have a solution to the measurement problem. I have a suggestion for how to live with it: use multiple proxies, triangulate, and be explicit about what each proxy captures. The empirical chapters of this book follow that practice. Where I report certification rates, I also report the depth of implementation where the data allow. Where I report emissions intensity, I control for product mix. Where I report disclosure, I note the selection problem. The goal is not precision. It is to avoid the error of inferring capability from compliance, which is the error the conventional framing makes.

Table 3-1 summarizes the four components, their observable proxies, and their characteristic failure modes.

Table 3-1: The four components of green governance capability.

The table makes the sequencing visible, and the sequencing is the analytical point. Capability is not a bundle that a firm either has or lacks. It is a stack that a firm builds, and the returns to building depend on what has already been built. The compliance framing treats the stack as a single requirement. The stack is a ladder.

3.2 Mechanisms Linking Green Capability to Upgrading

3.2.1 From information asymmetry to bargaining power

The most basic mechanism is also the most underappreciated. Green capability reduces information asymmetry between the supplier and the buyer, and the reduction of information asymmetry changes the terms of trade.

Consider the structure of a typical sourcing relationship before green standards. The buyer knows the price, the delivery schedule, and the sample quality. The buyer does not know the supplier's process emissions, its labor conditions, its material sourcing, or its waste handling. The supplier knows these things, and the supplier has an incentive to disclose only what is favorable. The result is a market with asymmetric information, and asymmetric information produces a lemons problem: the buyer discounts all suppliers to the level of the worst, and the good suppliers cannot capture the value of their quality.

Green standards change this by creating a demand for the information the supplier holds. A factory that can credibly disclose its emissions, its material sourcing, and its labor conditions separates itself from the pool. The separation is valuable to the buyer because the buyer faces its own disclosure obligations—to its customers, to regulators, to investors—and it needs supplier data to meet them. The factory that can provide the data at the level of granularity the buyer needs becomes a partner in the buyer's compliance rather than a vendor subject to it.

Van Weele and Vivanco (2014) describe the supplier moving from cost center to source of value. The supplier's capability reduces the buyer's risk and lowers the buyer's own compliance cost. The relationship shifts from transactional to relational, and relational sourcing relationships are stickier and more profitable for both parties.

The bargaining power implication is direct. In a transactional relationship, the buyer can switch suppliers at low cost, and the supplier's price is disciplined by competition. In a relational relationship, switching costs rise for the buyer, and the supplier's price is disciplined by the value of the relationship rather than the spot market. The supplier with green capability has moved from the first regime to the second. That is a change in value chain position, and it is the first channel through which green capability translates into upgrading.

Causality here needs care. The relational shift is not caused by green capability alone. It is caused by the combination of green capability and the buyer's own regulatory exposure. A buyer with no disclosure obligations has no reason to pay for supplier data. A buyer facing CBAM has a strong reason. The value of green capability is contingent on the regulatory environment in the buyer's market. This is why the upgrading effects of green standards are stronger in value chains oriented toward the EU than in value chains oriented toward markets with weaker disclosure regimes. The capability is general; the return to the capability is market-specific.

3.2.2 Learning by measuring

The act of measuring environmental performance generates knowledge the firm did not have before, and that knowledge is an input to process improvement.

The quality movement made the same argument fifty years ago: you cannot improve what you do not measure. The environmental version has a specific feature the quality version lacked. Environmental performance is a cross-cutting property of the production system. It depends on material inputs, energy sources, process temperatures, equipment maintenance, waste handling, and logistics. Measuring it forces the firm to look at the interactions among these factors, and the interactions are where the improvement opportunities are.

The cement industry offers a clear example. A cement kiln that co-processes waste—using municipal solid waste, medical waste, or biomass as alternative fuel—reduces its coal consumption and its emissions simultaneously. During the COVID-19 pandemic, a large cement company in Hubei Province used kiln co-processing technology to incinerate medical waste at 1,800°C, converting the waste into alternative fuel and achieving harmless disposal. The capability to do this required measurement: the firm had to know the calorific value of the waste stream, the emissions profile of the kiln under different fuel mixes, and the effect of the waste on clinker quality. The measurement capability was the precondition for the process innovation, and the process innovation generated both environmental and economic returns.

A second cement producer built China's first biomass alternative fuel system, capable of processing 150,000 tons of biomass waste—straw, rice husks, wood chips—per year, saving 49,000 tons of raw coal. The system was not a response to a regulation. It was a response to a measurement finding: the firm had quantified its fuel costs and its emissions, and the quantification revealed that biomass substitution was economically attractive at prevailing coal prices. The measurement created the business case.

Measurement does not just document performance; it reveals the structure of the production function. Firms that measure discover things about their processes that they did not know, and the discoveries are the raw material for innovation. The firms that do not measure do not discover, and they do not innovate, and they fall behind.

The learning mechanism has a collective dimension that is important for the cluster argument I developed in earlier work. When firms in a cluster measure their environmental performance, the measurements aggregate. The cluster-level data reveals opportunities that no individual firm could see: shared waste treatment, collective energy procurement, common logistics. The cluster becomes a unit of analysis for green governance in a way that individual firms cannot be. I return to this in Chapter 9, but the mechanism starts here, at the level of the individual firm's measurement practice.

3.2.3 Market access and the green premium

Green capability opens markets that are closed to firms without it, and the access is worth more than the cost of the capability.

The most straightforward version is regulatory access. A firm that cannot document compliance with the EU's CBAM cannot sell into the EU market after the financial obligations phase in. The market is closed, not by a tariff but by a documentation requirement. The firm that can document compliance keeps the market; the firm that cannot loses it. The value of the capability is the value of the market, and for many Chinese exporters the EU market is 15 to 30 percent of their revenue.

A second version is buyer access. Large retailers and brands have their own supplier standards, and the standards are often stricter than the regulations. Walmart's supplier sustainability program, IKEA's IWAY standard, and similar programs at other retailers impose requirements on suppliers that go beyond legal compliance. A factory that meets these requirements can bid for the retailer's business; a factory that does not cannot. The capability is a ticket of admission, and the ticket is valuable because the retailers' order volumes are large and their payment terms are reliable.

A third version is the green premium—the price differential that green products command in consumer markets. The evidence on the size of the premium is mixed. Mohr, Webb, and Harris (2001) found that consumers express strong preferences for socially responsible companies but that the preferences translate weakly into purchase behavior. The attitude-behavior gap is real, and it limits the premium. But the gap is smaller in business-to-business markets than in consumer markets, because business buyers have their own disclosure obligations and can pass the cost through. A European office furniture retailer that buys low-carbon desks can market them to corporate clients who need to report their Scope 3 emissions. The premium is embedded in the corporate client's willingness to pay for a lower carbon footprint, and the retailer captures it and shares it with the supplier.

The premium is not uniform across products or markets. It is larger where the buyer's disclosure obligations are stronger, where the product's carbon footprint is a larger share of the buyer's total footprint, and where the supplier's capability is scarce. The scarcity condition matters for the upgrading story: the premium exists because most suppliers cannot meet the standard, and it will erode as more suppliers build the capability. The firms that capture the premium are the early movers, and the early movers are the ones that built capability before the market demanded it.

3.2.4 Rent capture and the upgrading ladder

Green capability allows firms to capture rents that would otherwise accrue to other actors in the value chain, and the capture is what moves the firm up the ladder.

The rents in question are not monopoly rents. They are the returns to scarce capabilities—the Schumpeterian rents that accrue to firms that do something competitors cannot easily replicate. Gereffi and colleagues, in their work on global value chain governance, distinguish between rents that come from access to resources, rents that come from technology, and rents that come from organizational capability. Green governance capability is an organizational rent. It is not tied to a resource deposit or a patent. It is tied to the firm's routines, and routines are hard to copy because they are embedded in people and relationships.

The rent capture shows up in three places. First, in the price. A firm with green capability can charge more than a firm without it, because the buyer values the capability and the capability is scarce. Second, in the terms. A firm with green capability can negotiate longer contracts, better payment terms, and higher volumes, because the buyer's switching cost is higher. Third, in the scope. A firm with green capability can take on design and engineering work that a firm without it cannot, because the capability includes the knowledge to advise the buyer on how to meet the buyer's own environmental targets.

The third channel is the most important for upgrading, and it is the one the compliance framing misses entirely. A firm that only complies is a firm that executes specifications. A firm that has capability is a firm that helps write the specifications. The shift from execution to specification is the shift from the bottom of the value chain to the middle, and it is the shift that the Chinese furniture factories in Lecong were beginning to make when I spoke with them. The manager who said "we sell proof" was describing a firm that had moved from making boards to making the documentation that the buyer needed—and in the process, from a vendor to a partner.

Figure 3-1 traces the sequence.

flowchart TD
    A[External requirement<br/>CBAM, buyer standard, domestic regulation] --> B[Measurement<br/>install meters, build data architecture]
    B --> C[Management<br/>integrate data into decisions]
    C --> D[Verification<br/>third-party audit, credible disclosure]
    D --> E[Innovation<br/>exceed the standard]
    E --> F[Value chain position<br/>price, terms, scope]
    B --> G[Learning<br/>discover production function]
    G --> E
    D --> H[Information asymmetry<br/>reduced]
    H --> I[Bargaining power<br/>relational sourcing]
    I --> F
    E --> J[Green premium<br/>scarce capability]
    J --> F
    F --> K[Reinvestment in capability]
    K --> B

Figure 3-1: The capability-to-upgrading sequence.

The feedback loop at the bottom is the part the compliance framing cannot accommodate. Compliance is a one-way flow: requirement in, adjustment out. Capability is a loop: the returns to capability finance further capability, and the firm's position improves over time. The loop is why the firms that start building early tend to stay ahead, and why the firms that start late find it hard to catch up.

3.3 Endogenous versus External Drivers

3.3.1 The compliance-pressure thesis and its limits

The dominant account of why Chinese firms adopt green practices is external pressure. Regulators tighten standards, buyers impose requirements, and firms respond. The account is intuitive, it is supported by a large empirical literature, and it is incomplete.

The evidence for the compliance-pressure thesis is real. Dechezleprêtre and Sato (2017), in their review of the literature on the effects of environmental regulations on competitiveness, find that regulation does affect firm performance, though the effects are heterogeneous and often smaller than the political debate suggests. The regulation-induced adoption of cleaner technologies is well documented. The Porter hypothesis—that well-designed regulation can stimulate innovation—has partial empirical support, though the conditions under which it holds are narrower than the original formulation suggested (Porter and van der Linde, 1995a).

The limits of the thesis become visible when you ask what happens after the regulation is met. If compliance is the whole story, the firm should stop at the standard. It should invest the minimum necessary, document it, and return to the business of competing on cost. Many firms do exactly this. But some do not. Some firms continue to invest, exceed the standard, and use the capability they have built to change their position in the value chain. The compliance thesis cannot explain why.

The thesis also has a timing problem. Regulation is episodic; capability is cumulative. A firm that builds capability only when regulation demands it will always be playing catch-up, because the regulation arrives before the capability is ready. A firm that builds capability ahead of regulation has options that the reactive firm does not. The difference between the two is the difference between a firm that experiences regulation as a shock and a firm that experiences it as an opportunity.

3.3.2 The endogenous-capability thesis

The alternative account is that green capability is endogenous to the upgrading process—that it emerges from the firm's own development trajectory, and that external pressure accelerates or retards it but does not create it.

The endogeneity argument has three parts.

The capability is built through learning by doing. A firm that has been measuring its environmental performance for ten years has a data series, a set of routines, and a body of tacit knowledge that a firm starting today cannot replicate quickly. The learning is cumulative, and the cumulativeness is what makes the capability an asset. This is the Nelson and Winter (1982) argument about routines: the firm's knowledge is stored in its routines, and routines are built through repetition, not through purchase.

The capability is built in response to internal problems, not only external requirements. A factory that is losing money on energy costs has an incentive to measure its energy use regardless of what the regulator says. A factory that is losing customers to competitors with better environmental documentation has an incentive to build documentation capability regardless of what the buyer demands. The internal problem creates the demand for the capability, and the external requirement arrives later, when the capability is already partly built.

The capability has spillovers within the firm that make it self-reinforcing. Measurement capability built for one product line can be applied to another. Management capability built for one customer can be applied to another. Verification capability built for one market can be applied to another. The spillovers raise the return to the initial investment, and the raised return justifies further investment. The process is self-sustaining once it starts.

The endogenous thesis does not deny that external pressure matters. It reframes the relationship. External pressure is a trigger, not a cause. It can start the process, but the process has its own logic once started. The firms that upgrade are the ones where the internal conditions for capability building were already present—where the management had the bandwidth, the engineers had the skills, and the organization had the routines—and the external pressure arrived at a moment when the firm could respond.

3.3.3 The interaction: pressure as trigger, capability as trajectory

The two theses are not rivals. They describe different phases of the same process, and the phases interact.

In the early phase, external pressure dominates. The firm has no capability, no internal demand for it, and no reason to invest. The regulation or the buyer requirement creates the demand. The firm responds, often minimally, and the response is the first step in capability building. The compliance framing describes this phase accurately.

In the later phase, the internal logic dominates. The firm has built some capability, the capability has generated returns, and the returns justify further investment. The firm begins to anticipate requirements rather than respond to them. The endogenous framing describes this phase accurately.

The transition between the phases is where the interesting variation is. Some firms make the transition; some do not. The ones that do are the ones where the initial compliance investment generated enough learning to create an internal demand for more. The ones that do not are the ones where the compliance investment was purely defensive—a cost to be minimized, not a capability to be built.

What determines which path a firm takes? My reading of the Lecong evidence, and of the exporter interviews I conducted in Ningbo and Dongguan between 2019 and 2023, points to three factors. I state them as analytical claims from the case evidence rather than as findings from a single identifiable study, because no study I know of has tested them directly against a comparable sample.

The first is the depth of the initial investment. A firm that installs meters and trains staff to use them is more likely to make the transition than a firm that hires a consultant to produce a report. The second is the presence of an internal champion—a manager or engineer who sees the capability as a source of advantage rather than a cost. The third is the firm's position in the value chain. A firm that sells directly to a demanding buyer has more reason to build capability than a firm that sells through an intermediary who absorbs the compliance burden.

These factors are not deterministic. They are tendencies, and the tendencies are what the case evidence in the following chapters will trace.

3.3.4 Why the distinction matters for policy

The endogeneity question is not academic. It determines what kind of policy can promote upgrading.

If compliance pressure is the driver, the policy prescription is straightforward: tighten the standards. Firms will respond, and the response will produce the desired environmental outcome. The cost is borne by the firms, and the cost is justified by the environmental benefit.

If capability is the driver, the prescription is different. Tightening the standards will produce compliance, but it will not produce capability unless the firms have the internal conditions to build it. The policy needs to support the building—through training, through access to verification infrastructure, through the kind of cluster-level institutions that reduce the fixed cost of capability investment. The standards are necessary but not sufficient.

The Chinese experience with the dual carbon goals, which I examine in Chapter 8, illustrates the difference. The goals—peak carbon before 2030, carbon neutrality before 2060—are ambitious, and they have triggered a wave of regulatory activity. The firms that have responded by building capability are the ones that will capture the upgrading benefits. The firms that have responded by minimizing compliance cost will face rising costs as the standards tighten. The policy question is how to move firms from the second group to the first, and the answer depends on understanding that the move is a capability-building process, not a compliance event.

3.4 The Capability Trap: Why Some Firms Do Not Climb

3.4.1 The compliance equilibrium

Not every firm climbs. Many firms, perhaps most, settle into a compliance equilibrium: they meet the standard, they document it, and they do not invest further. The equilibrium is stable, and understanding why it is stable is as important as understanding why some firms escape it.

The compliance equilibrium has three features. The firm's investment is calibrated to the requirement. It installs the equipment the regulation specifies, it hires the auditor the buyer requires, and it does not invest in capability beyond what is needed to pass. The firm's knowledge is held by external actors. The consultant knows the methodology; the certifier knows the standard; the firm knows how to hire them. And the firm's cost structure is transparent to its competitors. Everyone in the compliance equilibrium faces the same requirements and the same compliance costs, so the costs are passed through to the buyer and the competitive position is unchanged.

The equilibrium is comfortable in the short run and dangerous in the long run. It is comfortable because the firm is not falling behind. It is dangerous because the standard will rise, and the firm's capability will not have risen with it. Each round of tightening requires a new investment, and the new investment is as costly as the first because the firm has not built the capability that would make the next investment cheaper. The firm is on a treadmill, and the treadmill speeds up.

The furniture sector in the first decade of the 2000s illustrates the dynamic. Factories that treated formaldehyde compliance as a one-time investment found themselves re-investing every time the standard tightened. Factories that built measurement and management capability found that the subsequent tightenings were cheaper, because the infrastructure was already in place. The difference compounded over time, and by the 2010s the two groups of factories occupied different positions in the value chain.

3.4.2 The absorptive capacity constraint

The second reason firms do not climb is absorptive capacity. The concept, developed by Cohen and Levinthal (1990), refers to the firm's ability to recognize the value of new information, assimilate it, and apply it to commercial ends. Absorptive capacity is a function of the firm's prior knowledge and its investment in R&D, and it varies enormously across firms.

For green governance capability, absorptive capacity shows up in the firm's ability to understand what the buyer or regulator is asking for. A firm with strong absorptive capacity reads the CBAM methodology and sees a data architecture problem it can solve. A firm with weak absorptive capacity reads the same document and sees a threat it cannot assess. The difference is not intelligence; it is prior knowledge. The firm that has been measuring its emissions for years has the conceptual framework to understand a new measurement requirement. The firm that has not been measuring has to build the framework from scratch, and the building is slow.

Klewitz and Hansen (2014) find that absorptive capacity is a binding constraint for SMEs in sustainability-oriented innovation. The constraint is not the cost of the technology; it is the cost of understanding the technology. The firms that overcome the constraint are the ones that have invested in the human capital to understand it—engineers with environmental training, managers with exposure to international standards, workers with the skills to operate new equipment.

The absorptive capacity constraint has a cluster dimension. Firms in clusters with strong knowledge institutions—universities, technical centers, industry associations—have higher absorptive capacity than firms in clusters without them. The institutions provide the prior knowledge that the firms lack individually, and the provision lowers the cost of capability building. This is one reason why the cluster-based approach to green governance, which I developed in earlier work, is more effective than a firm-by-firm approach.

3.4.3 The verification bottleneck

The third reason firms do not climb is verification. Building measurement and management capability is necessary but not sufficient; the capability has to be verified to be valuable, and verification is expensive.

The cost of verification has two components. The direct cost is the fee paid to the certifier, which for a carbon footprint verification can run to tens of thousands of yuan per facility per year. The indirect cost is the disruption to operations—the staff time, the documentation, the audits, the corrective actions. For a small firm, the indirect cost can exceed the direct cost, and the combination can be prohibitive.

The verification bottleneck is a market failure. The certifier has to be accredited, the accreditation is expensive, and the number of accredited bodies in China is limited. The limited supply keeps fees high, and the high fees exclude the firms that most need verification—the small and medium enterprises that make up the bulk of the furniture, textile, and electronics sectors. The firms that can afford verification are the large ones, and the large ones are the ones that already have capability. The bottleneck reinforces the advantage of the incumbents.

There is a collective solution to the bottleneck, and it is the one that the cluster literature has identified. A cluster can establish a shared verification facility—a testing laboratory, a certification support center, a common data platform—that spreads the fixed cost across many firms. The facility lowers the cost per firm, and the lowered cost bringsverification within reach of the SMEs that the market excludes. The solution requires collective action, which requires the kind of cluster governance institutions I examined in the earlier work on industrial clusters. The institutions are not automatic; they have to be built. But where they exist, the verification bottleneck loosens.

A concrete instance from Lecong makes the arithmetic visible. In 2021, a mid-sized desk manufacturer I visited—340 workers, roughly 60 percent of output exported to Europe—paid RMB 38,000 for an initial carbon footprint verification of one product line, plus an estimated 210 hours of staff time assembling the underlying data. The staff time was the larger cost. At the firm's loaded labor rate, those hours came to roughly RMB 52,000. The total was about RMB 90,000 for a single product line, in a year when the firm's net margin on that line was around 4 percent. For the firm's smaller neighbor across the street—90 workers, domestic sales only—the same exercise would have consumed most of a year's profit. The neighbor did not attempt it. It still sells into the domestic market, where no buyer asks for the declaration.

The cluster laboratory changed the calculus for the first firm but not the second. The shared facility brought the testing component down; it could not bring down the documentation and audit-preparation hours, which are the firm's own labor and cannot be outsourced without defeating the purpose. This is the part of the verification bottleneck that collective infrastructure does not solve, and it is why the trap persists even where shared facilities exist.

3.4.4 The trap in comparative perspective

The capability trap is not unique to China. It is a general feature of the relationship between environmental standards and firm development, and it appears in every economy that has tightened its standards. The difference across economies is the speed at which firms escape the trap, and the speed depends on the institutional support for capability building.

The European experience is instructive. The EU's chemical regulation, REACH, imposed registration and authorization requirements that were costly for small firms. The EU responded with support programs—the European Chemicals Agency's helpdesk, the national REACH helpdesks, the industry association guidance—that lowered the cost of compliance and, in the process, built capability. The firms that used the support built the capability; the firms that did not fell behind. The support programs were not subsidies. They were information infrastructure, and the information infrastructure was what made the difference.

The Chinese experience has been more uneven. The national standards are stringent, and enforcement has tightened, but the support infrastructure for capability building is thinner. The testing laboratories are concentrated in the coastal provinces. The certification bodies are few. The training programs are fragmented across ministries. The result is that the capability trap is deeper in the interior than on the coast, and deeper for small firms than for large ones. The regional differentiation is the subject of Chapter 9, but the mechanism starts here.

3.5 Capability Building as a Strategic Choice

3.5.1 The managerial question

If capability is the driver of upgrading, the managerial question is how to build it. The question is not "should we comply?"—compliance is a condition of doing business. The question is "should we invest in capability beyond compliance, and if so, how?"

The answer depends on the firm's assessment of three things: the trajectory of the standards, the value of the markets that the capability opens, and the cost of building the capability. The assessment is strategic, and it is made under uncertainty. The firm does not know how fast the standards will tighten, how much the markets will pay for the capability, or how much the building will cost. The strategic choice is a bet on the trajectory.

The firms that make the bet well are the ones that have good information about the trajectory. They read the regulatory signals—the EU's CBAM timeline, the Chinese carbon market's expansion, the buyers' public commitments. They talk to their customers about what the customers will need. They monitor the competitors' investments. The information is not proprietary; it is public. The difference is that some firms look for it and some do not.

The firms that make the bet well also have a realistic assessment of their own capacity. A firm with strong engineering and weak management should build measurement first, because measurement is an engineering problem. A firm with strong management and weak engineering should build management first, because management is an organizational problem. The sequencing matters, and the firms that get it right build faster.

3.5.2 The role of the buyer

The buyer's role in capability building is more important than the compliance framing suggests. The buyer does not just set the requirement; the buyer shapes the supplier's response to it.

Van Weele and van Tubergen (2014) describe the mechanism in their study of responsible purchasing. When a buyer imposes a requirement, the buyer also provides—or fails to provide—the support that the supplier needs to meet it. The buyers that provide support (training, technical assistance, longer contracts, shared investment) get suppliers that build capability. The buyers that do not provide support get suppliers that comply minimally or exit. The difference is the buyer's purchasing strategy, and the strategy is a choice.

The Nespresso case, documented by Van Weele and Van Tubergen (2013), illustrates the supportive model. Nespresso's sourcing program for coffee includes agronomic training, quality standards, and long-term purchasing commitments. The program is not charity; it is a capability-building investment that secures the supply of high-quality coffee. The suppliers build capability because the buyer makes it worth their while, and the buyer gets a supply chain that can meet its quality and sustainability requirements.

The extractive model is the opposite. The buyer sets the requirement, audits the supplier, and switches if the supplier fails. The supplier has no reason to invest beyond the minimum, because the buyer's commitment is contingent and the buyer's support is absent. The extractive model produces compliance, not capability, and the compliance is fragile because it depends on the audit rather than the routine.

The Chinese exporters' experience with European buyers has been mixed. Some buyers have invested in supplier capability—the IKEA model, the H&M model—and the suppliers have built capability. Other buyers have extracted compliance—the audit-and-switch model—and the suppliers have complied minimally. The variation in buyer strategy explains a significant part of the variation in supplier capability, and the explanation is missing from the compliance framing, which treats the buyer as a source of pressure rather than a source of support.

The purchasing manual I mentioned at the opening of this chapter is a case in point. The Dutch retailer's 47 pages and seven annexes were not just a list of requirements. Annexes four and five specified the data format the supplier had to submit, the verification body's accreditation standard, and a named technical contact at the retailer who would answer methodology questions within five working days. That contact line is the difference between a supportive buyer and an extractive one. The Lecong manager told me he had used it eleven times in the first year. "They want us to get it right," he said. "The other buyer, the one from Chicago, sends the audit and a deadline." The Chicago buyer's suppliers, in his account, do the minimum and hide the rest.

3.5.3 The role of the cluster

The cluster's role in capability building is the subject of Chapter 9, but the mechanism belongs here, because it is part of the endogeneity argument.

A cluster provides three things that individual firms cannot provide for themselves. The first is shared infrastructure—testing laboratories, verification bodies, training programs—that lowers the fixed cost of capability building. The second is knowledge spillovers—the informal exchange of information among firms, the movement of engineers between firms, the collective learning that comes from proximity. The third is collective voice—the cluster's ability to negotiate with buyers and regulators as a group, which gives the individual firms more leverage than they would have alone.

The three things interact. The shared infrastructure attracts the knowledge spillovers, because firms that use the laboratory talk to each other. The knowledge spillovers strengthen the collective voice, because the firms that learn together can act together. The collective voice attracts more infrastructure, because the cluster can make a credible case for investment. The interaction is the cluster's advantage, and it is why the cluster-based approach to green governance outperforms the firm-by-firm approach.

The Lecong furniture cluster is a case in point. The cluster's testing laboratories, established in the 2010s with support from the local government and the industry association, lowered the cost of formaldehyde testing for the small factories. The lowered cost brought the small factories into compliance, and the compliance brought them into contact with the larger factories that were already exporting. The contact produced spillovers—the small factories learned the documentation practices of the large ones—and the spillovers raised the cluster's overall capability. The process was not planned; it emerged from the infrastructure. But the infrastructure was the precondition, and the infrastructure was a collective investment.

3.6 Objections and Qualifications

3.6.1 Is capability just a euphemism for size?

The most serious objection to the capability argument is that capability is correlated with firm size, and the correlation means that the argument is really about scale, not about capability.

The objection has force. Large firms have more resources, more staff, and more slack to invest in capability. Darnall, Henriques, and Sadorsky (2010), using survey data from facilities in seven OECD countries, find that facility size is among the strongest predictors of environmental management system adoption and of the adoption of proactive environmental practices more generally. Work on Chinese firms points the same way: larger exporters are more likely to hold ISO 14001 and to report environmental data. The capability argument, if it is not careful, can become a restatement of the size argument in different language.

The response is that the correlation is not perfect, and the residual variation is where the capability argument lives. Some large firms have weak capability—they have the certificates but not the routines—and some small firms have strong capability—they have the routines even if they lack the formal systems. The firms that climb are not always the large ones; they are the ones with the capability, and the capability is not fully determined by size.

The Lecong evidence supports the response. The cluster's most capable exporters are not its largest firms. They are mid-sized firms—200 to 500 workers—where the owner-manager has an engineering background and the firm has invested in measurement and verification beyond the minimum. The largest firms in the cluster are contract manufacturers for domestic brands, and their environmental capability is thin because their buyers do not demand it. The mid-sized exporters have built capability because their buyers do demand it, and the building has given them a position in the value chain that the large domestic-oriented firms do not have.

Size matters, but it is not destiny. The capability argument is about what firms do with the resources they have, not about how many resources they have.

3.6.2 Does the argument overstate the returns to green capability?

The second objection is that the returns to green capability are smaller than the argument claims, and the small returns mean that the upgrading effect is marginal.

The objection is partly right. The returns are not uniform, and they are not always large. The green premium in consumer markets is modest, as the attitude-behavior gap literature shows. The price premium in business-to-business markets is larger but still limited by competition. The rent capture is real but temporary, because the capability diffuses and the scarcity erodes.

The response is that the returns are larger than the price premium suggests, because the price premium is only one of the channels. The bargaining power channel—the shift from transactional to relational sourcing—is worth more than the price premium, because it changes the terms of the relationship, not just the price. The market access channel is worth more than the price premium, because it keeps the firm in markets that would otherwise close. The learning channel is worth more than the price premium, because it generates process improvements that reduce costs. The price premium is the visible part of the return; the invisible parts are larger.

The argument does not claim that green capability guarantees upgrading. It claims that green capability is one of the mechanisms through which upgrading happens, and that the mechanism is more important than the compliance framing recognizes. The claim is conditional, and the conditions are the subject of the empirical chapters.

3.6.3 Is the argument China-specific?

The third objection is that the argument is China-specific—that it describes the Chinese experience with the dual carbon goals and the export-oriented manufacturing sector, and it does not generalize.

The objection is partly right and partly wrong. The Chinese experience has features that are distinctive: the scale of the manufacturing sector, the speed of the regulatory tightening, the role of the export markets in driving standards. These features shape the capability-building process, and the process in China is not identical to the process elsewhere.

But the mechanism is general. The capability-to-upgrading sequence—measurement, management, verification, innovation—appears in the European experience with REACH, in the Japanese experience with energy efficiency standards, in the Korean experience with green certification. The sequence is not a Chinese invention; it is a general feature of the relationship between environmental standards and firm development. The Chinese case is a particularly clear instance because the scale is large and the change is fast, but the instance is an instance of a general phenomenon.

The general phenomenon is what the book is about. The Chinese case is the evidence, not the exception.

3.7 From Capability to Governance: A Bridge to Chapter 4

The capability argument has an implication that the next chapter develops. If capability is the driver of upgrading, then the governance of the value chain—the rules, the standards, the measurement systems—is not just an external constraint. It is the infrastructure through which capability is recognized and rewarded.

A standard is a measurement convention. It specifies what is measured, how it is measured, and what counts as compliance. The convention determines which capabilities are visible and which are invisible. A standard that measures end-of-pipe emissions rewards the firm that installs scrubbers. A standard that measures product-level carbon footprints rewards the firm that has built measurement and management capability. The choice of convention is a choice of which capabilities the value chain will reward, and the choice is made by the actors who write the standards.

Chapter 4 examines the measurement conventions—the carbon accounting methodologies, the verification regimes, the labeling schemes—and traces how they shape the capability-building process. The conventions are not neutral. They embed assumptions about what matters, and the assumptions determine which firms climb and which firms are left behind. The capability argument is the foundation; the governance argument is the superstructure. The next chapter builds the superstructure.

Before turning to it, one more observation about the Lecong factory. The manager showed me the carbon footprint declaration for the desks—a two-page document with the emissions broken down by life cycle stage, the methodology cited, the verification body named. He was proud of it, in the way that an engineer is proud of a well-designed machine. "This," he said, tapping the document, "is what they cannot get from Vietnam."

He may be right. The Vietnamese factories can match the price. They cannot yet match the documentation. The specific gap is the batch-level data architecture: the Vietnamese suppliers I have spoken with can produce a facility-level emissions estimate, but not the product-level, batch-traceable declaration the Dutch retailer's Annex 4 requires, because the traceability runs back through the particleboard supplier, and the Vietnamese board mills do not yet carry the chain-of-custody records that the declaration depends on. Closing that gap is a measurement and management problem, not a price problem. It is the kind of problem this chapter has been about.