Chapter 4: Green Governance Capability: Conceptualization, Measurement, and Its Endogenous Drivers

In October 2023, a sofa manufacturer in Shunde received an email from its largest European customer. The message contained a spreadsheet template and a deadline: product-level carbon footprint data for every order, within eighteen months. The firm's finance manager downloaded the template, asked the production department for energy figures, and discovered that the factory had monthly electricity bills but no sub-metering by line. She filled in the template using industry averages and sent it back. The customer accepted the submission and flagged the reliance on averages as a supply risk.

Two kilometers away, a competitor of similar size received the same email. Its deputy general manager had been tracking EU carbon policy since the CBAM proposal appeared in 2021. He convened a working group drawn from production, procurement, and export sales. The group commissioned sub-metering on three production lines, built a product-level carbon model from actual energy and material data, and identified two hot spots: embodied emissions in the foam supply and electricity consumed by the drying ovens. The firm switched to a foam supplier that could provide verified lower-carbon material and installed heat recovery on the ovens. When the customer's 2025 requirement arrived, the firm had measured data, a documented reduction pathway, and a differentiated story. It moved onto the customer's preferred supplier list.

Same regulation. Same customer. Same cluster. Different outcomes. The difference is what this chapter calls green governance capability.

4.1 Defining Green Governance Capability: Dimensions and Components

4.1.1 Why a Capability Construct Is Needed

The environmental management literature has produced a long list of firm-level constructs: environmental management systems, eco-innovation, green supply chain management, corporate environmental performance, sustainability orientation. Each captures something real. None captures the specific object this book needs to explain: the organizational capacity that allows a firm to convert external environmental pressure into improved value chain position.

ISO 14001 certification illustrates the gap. The certificate measures the presence of a management system, not the firm's ability to use that system strategically. A firm can hold the certificate and still treat environmental management as a compliance function staffed by one overworked engineer. Another firm can lack the certificate and embed carbon accounting into product design and customer negotiations. The certificate is an indicator; the capability is the underlying capacity.

Nelson and Winter (1982) and Teece, Pisano, and Shuen (1997) give the starting point. Firms are bundles of routines and competences, and competitive advantage rests on the difficulty competitors face in replicating those routines. Extending this logic to environmental governance requires three moves. Environmental governance must be treated as a set of routines rather than a set of outcomes. Those routines must be shown to connect to value chain position, not merely to regulatory standing. And the routines must be shown to be buildable through internal investment and learning, not simply purchased or mandated.

The construct I develop here draws on three literatures that rarely speak to each other. The environmental management literature supplies the components: pollution prevention, product stewardship, and clean technology, following the taxonomy Hart (1995) introduced. The global value chain literature supplies the position logic: governance structures determine which actors capture value, and environmental requirements are increasingly written into those governance structures by lead firms (Gereffi, Humphrey, and Sturgeon, 2005). The dynamic capabilities literature supplies the mechanism: sensing, seizing, and reconfiguring (Teece, 2007).

Green governance capability, in the definition I use throughout this book, is the firm's capacity to sense environmental constraints and opportunities across its value chain, to seize them through changes in products, processes, and supply relationships, and to reconfigure its resource base so that environmental performance becomes a source of differentiation rather than a cost of compliance.

Three implications follow. The capability is relational: it is exercised across firm boundaries, in relationships with suppliers, customers, and regulators, not solely within the firm. It is dynamic: it includes the capacity to revise routines as the regulatory and market environment changes. And it is strategic: it is defined by its contribution to value chain position, not by its contribution to emissions reduction per se. A firm that cuts emissions by shutting down production has not demonstrated green governance capability. A firm that cuts emissions per unit of output while improving product performance and customer relationships has.

4.1.2 The Three Dimensions

I organize the construct into three dimensions corresponding to the sensing, seizing, and reconfiguring functions, with content specific to environmental governance. The dimensions are analytical distinctions, not separate capacities; in practice they overlap and reinforce each other.

Environmental sensing and interpretation. The first dimension is the capacity to perceive environmental signals and translate them into decision-relevant information. Signals arrive from multiple sources: regulatory changes (the ETS expansion, CBAM implementing regulations, provincial energy quotas), customer requirements (brand carbon disclosure requests, retailer specifications, green procurement criteria), supplier changes (availability of recycled materials, alternative fuels, low-carbon logistics), and technology shifts (electrification of process heat, hydrogen-based reduction, digital carbon accounting). Sensing is not simply awareness. It is the capacity to distinguish signals that matter from noise, to interpret them in light of the firm's specific position, and to route them to the right decision-makers.

Most Chinese manufacturers have improved their sensing capacity considerably since 2015, largely because the regulatory signal has become too loud to ignore. ETS reporting requirements alone forced thousands of firms to build emissions inventories. But sensing capacity varies widely in depth. A furniture firm in Shunde that receives a carbon data request from an EU buyer and forwards it to its accountant has minimal sensing capacity. A competitor that maintains a rolling inventory of embodied emissions by product line, tracks the CBAM transitional reporting templates, and has assigned a deputy general manager to monitor EU and Chinese carbon policy has substantial sensing capacity. The difference is not information access, which is roughly equal. It is the organizational routine that converts information into strategic awareness.

Environmental seizing through process and product change. The second dimension is the capacity to act on what has been sensed: to redesign processes, reformulate products, renegotiate supply relationships, and reallocate investment in response to environmental constraints. This is where capability becomes visible. Seizing includes the classic pollution prevention investments (closed-loop water systems, waste heat recovery, solvent recovery in coating operations), product stewardship moves (design for recyclability, substitution of hazardous inputs, reduced material intensity), and clean technology moves (electrification of process heat, biomass substitution in cement kilns, digital process control for energy optimization).

The distinction between compliance and capability becomes concrete here. A firm that installs a scrubber because the regulator requires it has complied. A firm that redesigns its coating line to eliminate the solvent entirely, reducing both emissions and input costs, has seized an opportunity. The second firm's action required engineering capacity, supplier relationships, and a willingness to absorb transition costs that the first firm did not need. The capability is not the scrubber or the redesign as such. It is the organizational capacity to identify, evaluate, and execute the redesign under uncertainty.

Environmental reconfiguring of the resource base. The third dimension is the capacity to change the firm's underlying resource base so that environmental performance becomes durable: investing in people (hiring environmental engineers, training production managers in carbon accounting), building data systems (emissions inventories, product carbon footprints, supplier databases), establishing external relationships (with technology providers, certification bodies, research institutes, lead firm sustainability teams), and modifying organizational structures (creating sustainability committees with real authority, integrating environmental metrics into performance evaluation).

Reconfiguring is the slowest dimension to build and the hardest to imitate. A competitor can copy a scrubber. It cannot quickly copy a decade of accumulated process knowledge, a supplier network trained in carbon reporting, or a management team that has learned to weigh environmental variables in investment decisions. It is also the dimension that Chinese firms in traditional sectors most often neglect, because it requires sustained investment with returns that are difficult to attribute to specific projects.

4.1.3 The Components Within Dimensions

Each dimension contains components that can be assessed separately. Table 4-1 lists the components I use in the measurement framework developed in the next section, organized by dimension.

Table 4-1: Components of green governance capability

The components are not equally important in all sectors. A cement firm's seizing capacity depends heavily on process optimization and alternative fuel substitution; a furniture firm's depends more on product redesign and supply relationship change. A technology-intensive electronics firm's sensing capacity depends on technology scanning; a labor-intensive textile firm's depends more on customer requirement detection. The measurement framework must accommodate this variation, which is why I propose sector-relative assessment rather than a single absolute scale.

4.1.4 What the Construct Is Not

Three boundary clarifications are necessary before proceeding to measurement.

Green governance capability is not environmental performance. A firm can have high capability and mediocre performance if its starting point was very poor, or low capability and good performance if its processes happened to be clean. Performance is an outcome; capability is the capacity that produces outcomes over time. The two correlate, but the correlation is imperfect, and the capability construct is the better predictor of future trajectory.

The construct is also not compliance. Compliance is a subset of seizing capacity: the ability to execute projects that meet regulatory requirements. A firm with only compliance capacity can meet the floor but cannot build above it. The distinction is analytical, not pejorative. Compliance capacity is necessary; it is not sufficient for value chain climbing.

Nor is the construct a measure of intent or commitment. Many firms publish sustainability reports and sign environmental pledges. These are signals, and they may or may not reflect underlying capability. The measurement framework I develop focuses on routines and observable organizational features, not on stated commitments. A firm that has an environmental committee that meets quarterly and makes binding decisions has different capability from a firm that has a sustainability report and no decision-making structure. The framework should distinguish them.

4.1.5 The Shunde Vignette: Capability in a Furniture Cluster

The two firms introduced at the start of this chapter clarify the construct in operation. Both make upholstered sofas for export, both employ roughly 400 workers, both have annual revenue in the 300–500 million yuan range, and both sell to European and North American buyers.

The first firm's response to the carbon data request was to assign the task to its finance manager, who downloaded the customer's spreadsheet template and asked production for energy consumption data. Production had monthly electricity bills but no sub-metering by product line. The finance manager estimated the footprint using industry averages and submitted the template. The customer accepted the submission but flagged the use of averages as a risk. The firm has since done nothing further. Its sensing capacity is minimal: it received the signal but could not interpret it in a way that generated action. Its seizing capacity is absent: it has no routine for product-level carbon accounting. Its reconfiguring capacity is undeveloped: no one in the firm has environmental engineering training, and the data infrastructure does not exist.

The second firm's deputy general manager, who had been tracking EU carbon policy since the CBAM proposal in 2021, convened a working group that included production, procurement, and export sales. The group commissioned sub-metering on the three main production lines, built a product-level carbon model using actual energy and material data, and identified two hot spots: the foam supplier's embodied emissions and the electricity used in the drying ovens. The firm renegotiated with two foam suppliers, selecting one that could provide verified lower-carbon foam, and invested in heat recovery on the drying ovens. By the time the customer's 2025 requirement arrived, the firm had product-level data, a documented reduction pathway, and a supplier relationship that gave it a differentiated story. The customer moved the firm to a preferred supplier list.

Both firms faced the same external signal. The second firm had sensing capacity (the deputy GM's policy tracking), seizing capacity (the working group's ability to commission metering and renegotiate supply), and reconfiguring capacity (the decision authority to commit investment and the relationships to execute). The first firm had none of these in operational form. The vignette also shows why the construct matters for value chain position: the second firm converted an environmental requirement into a preferred supplier status, which is a position improvement. The first firm converted it into a compliance cost.

4.2 Measurement Framework: Indicators and Methods

4.2.1 The Measurement Problem

Measuring green governance capability is difficult for the same reason measuring any organizational capability is difficult: the construct is latent. We observe indicators, not the capability itself. The indicators are imperfect, and different indicators capture different aspects. The measurement framework must therefore be explicit about what it is measuring, how the indicators relate to the construct, and what the limitations are.

Three approaches dominate existing practice. The first uses certification and management system adoption as proxies: ISO 14001, ISO 50001, green factory designations, and similar. These are easy to observe and compare, but they measure the presence of a system, not its strategic use. A firm can hold ISO 14001 and use it as a paperwork exercise. The second uses environmental performance metrics: emissions intensity, energy intensity, water intensity, waste recycling rates. These are meaningful outcomes, but they conflate capability with starting position and with the stringency of local regulation. A firm in a province with strict energy quotas may have low energy intensity because it was forced to invest, not because it has high capability. The third uses disclosure scores: CDP ratings, sustainability report quality, ESG ratings. These measure transparency and, to some extent, commitment, but they are vulnerable to greenwashing and are biased toward large firms with dedicated sustainability teams.

The framework I propose combines elements of all three but weights them differently. It treats certification and disclosure as weak indicators, performance metrics as contextual indicators, and organizational routines as strong indicators. The logic is that routines are closer to the construct: they are what the firm actually does, not what it says or what it has achieved. The framework is designed for application at the firm level, with aggregation to cluster and sector levels, and it is intended to be feasible with data that a research team can collect through structured interviews, site visits, and document review.

4.2.2 Indicator Structure

The framework uses twelve indicators, one for each component in Table 4-1, scored on a four-point scale. The scale is ordinal and sector-relative: a score of 3 indicates that the component is fully developed relative to the best practice observed in the sector, a score of 2 indicates partial development, a score of 1 indicates minimal development, and a score of 0 indicates absence. Sector-relative scoring is necessary because the content of best practice differs across sectors. A cement firm's process optimization routine looks different from a furniture firm's, and the scoring rubric must reflect that.

Table 4-2 presents the indicators, the evidence sources, and the scoring anchors for each.

Table 4-2: Indicator structure and scoring anchors

The scoring is deliberately coarse. A finer scale would create false precision. The purpose is to distinguish firms along the capability dimension, not to rank them precisely. In practice, the twelve indicators cluster into the three dimensions, and the dimension scores can be computed as averages or as minimums. I recommend reporting both: the average shows overall development, and the minimum shows the weakest link, which often constrains the capability more than the average.

4.2.3 Data Collection Protocol

The framework is designed for a research team of two to three people conducting a two-day site visit per firm, supplemented by document review and follow-up interviews. The protocol has four stages.

Document review. Before the visit, the team collects whatever is publicly available: annual reports, sustainability reports, certification records, environmental permits, and any customer correspondence the firm is willing to share. The review produces a preliminary score for the indicators that can be assessed from documents alone (S1, S9, S12 in part) and a list of questions for the visit.

Structured interview with management. A two-hour interview with the general manager, the production manager, and whoever has environmental responsibility covers the sensing and reconfiguring indicators. The interview uses a fixed question set but allows follow-up. Key questions include: How did you first learn about the carbon requirements from your main customer? Who in the firm is responsible for tracking environmental regulation? What environmental investments have you made in the last three years, and how were they decided? The interview is recorded with consent and transcribed.

Site walk-through. Half a day on the production floor lets the team see process optimization, data infrastructure, and product redesign directly. They look for sub-metering, waste segregation, heat recovery equipment, coating line configuration, and any visible environmental controls. This is not an audit. It is a chance to verify claims made in the interview and to ask about specific equipment and routines.

Follow-up and scoring. After the visit, the team scores each indicator independently, then reconciles discrepancies through discussion. Discrepancies of more than one point trigger a follow-up call or email to the firm. The final scores are recorded with brief justification for each, and the firm receives a summary report. The report is useful for the firm and builds trust for repeat visits.

The protocol has limitations. It relies on self-reported information for some indicators, and firms have incentives to overstate capability. The site walk-through mitigates this for the seizing and data indicators, but the sensing and relationship indicators are harder to verify. The team should triangulate where possible: customer correspondence, supplier contracts, and interviews with the firm's customers or suppliers, if accessible. In practice, complete triangulation is rarely possible, and the scores should be treated as estimates with known error.

4.2.4 Aggregation and Comparison

Firm-level scores can be aggregated to cluster and sector levels for comparative analysis. The aggregation method matters. A simple average across firms in a cluster weights large and small firms equally, which may not reflect the cluster's overall capability if large firms dominate output. A weighted average, using employment or output as weights, gives a different picture. I recommend reporting both and using the unweighted average for capability comparisons, because the capability construct is about organizational routines, not about output share.

Comparison across sectors requires adjustment for sector-specific best practice. A score of 3 on process optimization means different things in cement and furniture. The sector-relative scoring handles this at the indicator level, but cross-sector comparison still requires caution. The framework is best used for within-sector comparison and for tracking change over time within a firm or cluster. Cross-sector comparison should be limited to dimension-level patterns, not precise scores.

The framework also supports longitudinal analysis. Repeat visits at two- or three-year intervals allow the team to observe capability building (or decay) over time. The longitudinal dimension is where the framework is most valuable, because the construct is dynamic. A single cross-section shows where firms are; repeated observation shows whether they are building capability or stagnating.

4.2.5 Validation

The framework was developed and tested in a pilot study of 24 firms across four sectors (furniture, textiles, electronics, and cement) in Guangdong, Zhejiang, and Hubei provinces between 2022 and 2024. The pilot had three purposes: to test whether the indicators could be reliably scored, to check whether the scores correlated with independent measures of environmental performance and value chain position, and to identify indicators that were redundant or uninformative.

Inter-rater reliability was acceptable for most indicators. Two coders scoring the same firm independently agreed within one point on 10 of 12 indicators in 80 percent of cases. The two indicators with lower reliability were S4 (technology scanning) and S11 (external relationships), both of which rely heavily on interview self-report. The protocol was revised to require documentary evidence for scores of 2 or 3 on these indicators.

Construct validity was assessed by correlating capability scores with three independent measures: energy intensity relative to sector median, export share to markets with carbon pricing, and supplier status with lead firms (preferred versus approved versus transactional). Capability scores correlated positively with all three. The correlation with supplier status was strongest (Spearman's rho = 0.54), which is consistent with the theoretical claim that capability connects to value chain position. The correlation with energy intensity was weaker (rho = 0.31), which is consistent with the argument that performance is an imperfect proxy for capability.

The pilot also revealed that two indicators were doing little work. S8 (compliance project execution) correlated highly with S5 (process optimization) and added little independent information. S4 (technology scanning) was difficult to score reliably and correlated weakly with the other sensing indicators. I have retained both in the framework because they capture distinct components in some sectors, but researchers applying the framework may choose to drop them in sectors where they are uninformative. The framework is a tool, not a dogma.

4.2.6 Interpreting the Scores

A firm's capability profile is more informative than its total score. Two firms can have the same average and very different profiles. A firm with high sensing and low reconfiguring can see what is coming but cannot respond. A firm with high seizing and low sensing can execute projects but may execute the wrong ones. The profile reveals the binding constraint.

In the pilot, three profiles recurred. The first is the compliance profile: moderate scores on S5 and S8, low scores on S1, S2, S3, S9, S10, and S12. These firms do what the regulator requires and little else. The second is the truncated profile: high scores on S1 and S2, moderate on S5 and S6, low on S9, S10, and S12. These firms see the signals and can execute discrete projects, but lack the human capital and data infrastructure to build durable capability. The third is the integrated profile: high scores across most indicators, with particular strength in S9, S10, and S12. These firms have built the organizational foundations that allow capability to accumulate.

The profiles predict trajectories. Compliance-profile firms are vulnerable to relocation or exit as the floor rises. Truncated-profile firms can survive but struggle to climb. Integrated-profile firms are the ones that convert environmental pressure into position improvement. The chapters that follow use these profiles to organize the sectoral analysis.

4.3 Endogenous Drivers and Mechanisms: Internal Factors Fostering Capability Development

4.3.1 The Question of Drivers

If green governance capability were simply a function of regulatory pressure, all firms facing the same regulation would develop similar capability. They do not. The variation within sectors and clusters is large, and it is not explained by firm size alone. Some large firms have compliance profiles; some medium firms have integrated profiles. The drivers must therefore be internal to the firm and its immediate environment, or at least mediated by internal factors.

Five endogenous drivers emerged from the pilot as most consequential: managerial interpretation and attention, existing technological and organizational resources, position in the value chain, cluster-level resources and peer effects, and organizational slack and absorptive capacity. The list is not exhaustive. Each driver is a condition, not a cause in the strict sense; the mechanisms describe how the condition translates into capability building.

4.3.2 Managerial Interpretation and Attention

The first driver is managerial interpretation: how the firm's leadership understands the environmental signal and whether they allocate attention to it. This was the most consistently important factor in the pilot. Firms with integrated profiles almost always had a senior manager who had personally engaged with the environmental agenda, often after exposure to a customer requirement, a trade fair, or a peer firm's experience. Firms with compliance profiles almost always had leadership that treated environmental issues as a technical matter to be delegated.

The mechanism is attention allocation. Environmental signals compete with other signals for management attention, and the allocation is not automatic. A general manager who interprets carbon regulation as a threat to be minimized will allocate attention to compliance and cost pass-through. A general manager who interprets it as a market shift will allocate attention to capability building. The interpretation is shaped by prior beliefs, by the firm's customer relationships, and by the information environment the manager inhabits.

The Shunde case illustrates the mechanism. The second firm's deputy general manager had attended a trade fair in Germany in 2019 where carbon-neutral furniture was a prominent theme. That exposure changed his interpretation of the CBAM proposal when it appeared in 2021. He saw it as a continuation of a market shift he had already observed, not as an isolated regulatory event. The first firm's general manager had not had that exposure, and his interpretation was correspondingly narrower.

Managerial interpretation is not purely individual. It is shaped by the firm's governance structure and by the manager's network. Firms with boards that include external directors, or with managers who participate in industry associations and policy dialogues, tend to have broader interpretations. The driver is therefore partly organizational: the structures that expose managers to diverse information shape how they interpret environmental signals.

4.3.3 Existing Technological and Organizational Resources

The second driver is the firm's existing resource base. Capability building is path-dependent: firms build new capabilities on the foundations of existing ones. A firm with strong engineering capacity can absorb low-carbon process technologies more easily than a firm without. A firm with an established quality management system can extend it to environmental management more easily than a firm without. A firm with experience in product innovation can redesign products for environmental performance more easily than a firm that has only done cost reduction.

The mechanism is absorptive capacity (Cohen and Levinthal, 1990). The ability to recognize the value of new information, assimilate it, and apply it depends on prior related knowledge. A furniture firm with in-house design capacity can engage with a customer's design-for-recyclability requirement; a firm that only manufactures to customer specifications cannot. A cement firm with process engineering expertise can evaluate alternative fuel substitution; a firm that operates its kilns by routine cannot.

The pilot found that the resource base mattered most for the seizing and reconfiguring dimensions. Firms with strong engineering and quality systems scored higher on S5, S6, and S8. Firms with weak systems scored lower regardless of managerial interpretation. The implication is that capability building is constrained by the existing resource base, and that efforts to build capability must start from where the firm is, not from a generic best-practice template.

The resource base is not fixed. Firms can invest in building it, and the investment is itself a capability-building act. But the returns are slow and uncertain, which is why firms with weak foundations often struggle to build environmental capability even when they want to. The constraint is not motivation; it is the absorptive capacity to convert motivation into routine.

4.3.4 Position in the Value Chain

The third driver is the firm's position in the value chain. Firms that supply lead firms with explicit environmental requirements face stronger and more specific signals than firms that sell to less demanding customers. The signal strength affects both the motivation to build capability and the information available to build it. A supplier to a European brand that provides detailed carbon reporting templates, technical assistance, and preferred supplier incentives has more resources for capability building than a supplier to a domestic wholesaler that provides none.

The mechanism is what Gereffi, Humphrey, and Sturgeon (2005) call governance: the lead firm's coordination of the value chain transmits requirements and, in some cases, capabilities. The transmission is not automatic. It depends on the governance form. In modular governance, where suppliers make products to the buyer's specifications, the buyer may provide detailed requirements but little assistance. In relational governance, where buyer and supplier have a long-term relationship, the buyer may invest in supplier capability. In captive governance, where the supplier is dependent on a single buyer, the buyer may impose requirements without building capability.

The pilot found that firms in relational governance relationships had higher capability scores than firms in modular or captive relationships, controlling for size and sector. The relational firms received not just requirements but also technical assistance, training, and sometimes co-investment. The capability building was joint, not unilateral. This finding qualifies the assumption that capability is purely endogenous: it is built within the firm, but the resources and incentives for building it often come from the value chain relationship.

The value chain position also affects the returns to capability. A firm that supplies a lead firm with a carbon labeling commitment can capture a price premium or a preferred supplier status. A firm that supplies a price-sensitive domestic market may not. The expected return shapes the investment decision, which is why position and interpretation interact.

4.3.5 Cluster-Level Resources and Peer Effects

The fourth driver is the cluster environment. Firms are embedded in local production systems that provide shared resources: technical service providers, trade associations, training institutions, and infrastructure. The cluster also provides peer effects: firms observe what neighboring firms are doing and adjust their own behavior accordingly.

The mechanism has two parts. The first is resource pooling. A cluster with a specialized environmental testing laboratory, a trade association that runs carbon accounting workshops, or a local university with environmental engineering programs gives firms access to resources they could not afford individually. The Shunde furniture cluster, for example, has a testing center that offers product carbon footprint verification; firms can use the service without building in-house laboratories. The availability of the service lowers the cost of capability building.

The second is observational learning. Firms in clusters watch each other. When one firm invests in a new process and it works, others follow. When one firm wins a preferred supplier status by building carbon reporting, others notice. The diffusion is not automatic, and it is often uneven, but it is a real mechanism. The pilot found that firms in clusters with at least one integrated-profile firm had higher average capability scores than firms in clusters without one, controlling for sector and size. The presence of a visible example lowered the perceived uncertainty of capability building.

Cluster effects are not uniformly positive. Clusters can also diffuse a compliance mentality, if the dominant firms treat environmental issues as a cost to be minimized. The direction of the peer effect depends on which firms are most visible and most imitated. Policy can influence this by recognizing and publicizing integrated-profile firms, but the effect is mediated by the cluster's social structure.

4.3.6 Organizational Slack and Absorptive Capacity

The fifth driver is organizational slack: the resources available for investment in capability building beyond what is required for current operations. Slack comes in several forms: financial (retained earnings, access to credit), human (staff time not fully committed to current production), and managerial (attention not consumed by crises). Firms with more slack can invest in capability building without sacrificing current performance. Firms with less slack are constrained to compliance, even if they understand the strategic case for building.

The mechanism is the allocation of resources to exploratory versus exploitative activities (March, 1991). Capability building is exploratory: it involves uncertain returns and long time horizons. Compliance is exploitative: it involves known requirements and short time horizons. Firms under financial pressure allocate resources to exploitation. Firms with slack can afford exploration.

The pilot found that slack mattered most for the reconfiguring dimension. Firms with slack invested in human capital, data infrastructure, and organizational structures. Firms without slack could execute discrete projects but could not build the foundations. The finding is consistent with the broader literature on innovation and slack, and it has an uncomfortable implication: capability building is easiest for firms that are already successful, which means that the gap between integrated-profile and compliance-profile firms may widen over time.

Slack is not purely financial. A firm with a stable, experienced management team has managerial slack even if its financial resources are modest. A firm with a high turnover of production workers has less human slack, because it is constantly training replacements. The composition of slack matters for which capabilities can be built.

4.3.7 The Mechanisms in Combination

The five drivers do not operate independently. Managerial interpretation shapes how slack is allocated. The resource base constrains what interpretation can achieve. Value chain position provides both the signal and the resources for building. Cluster resources lower the cost of building. The combination explains why capability building is concentrated in some firms and absent in others.

The pilot data suggest a rough ordering of importance. Managerial interpretation was the most consistently associated with high capability scores, followed by value chain position and existing resources. Cluster effects and slack were important but secondary. The ordering is not universal; in sectors where the regulatory signal is very strong, interpretation matters less because the signal cannot be ignored. But in sectors where the signal is ambiguous or emerging, interpretation is decisive.

The ordering also has implications for policy. If interpretation is the binding constraint, then policy should focus on information provision and demonstration: exposing managers to market shifts, publicizing integrated-profile firms, and providing technical assistance that changes how managers understand the opportunity. If resources are the binding constraint, then policy should focus on access to finance and technical services. The two are not mutually exclusive, but the emphasis differs.

4.3.8 A Note on Exogenous Factors

The focus on endogenous drivers does not deny the importance of exogenous factors. Regulation, carbon prices, trade policy, and customer requirements all shape the environment in which capability building occurs. The argument is that these factors operate through internal mechanisms, and that the same exogenous shock produces different responses depending on the internal conditions. This is why the book treats the dual carbon architecture as a floor rather than a determinant. The architecture sets the baseline; the internal drivers determine how far above the baseline firms build.

The distinction has practical consequences. A policy that raises the floor without addressing the internal drivers will produce compliance, not capability. A policy that combines floor-raising with support for interpretation, resources, and cluster-level infrastructure is more likely to produce capability building. The next chapter takes up this theme in the context of labor-intensive industries, where the constraints on capability building are most severe.

4.3.9 The Compliance Trap Revisited

The concept of the compliance trap, introduced in Chapter 1, can now be stated more precisely. The trap is a dynamic in which a firm responds to rising environmental requirements by investing only in compliance, which raises costs without improving position, which reduces slack, which further constrains the firm's ability to invest in capability, which deepens its reliance on compliance. The trap is self-reinforcing, and it explains why some firms fall further behind as regulation tightens, even though they are doing what the regulation requires.

The endogenous drivers are the escape routes. A firm that changes its interpretation, builds slack through better position, or draws on cluster resources can break the cycle. The trap is not a failure of effort. It is a failure of the mechanism that converts effort into capability. Firms in the trap work hard at compliance and get no strategic return. Firms that escape work no harder, but they work through different routines.

The Shunde vignette shows the trap in miniature. The first firm's finance manager did what was asked: downloaded the template, collected what data was available, submitted the estimate. The effort was real. The return was zero, because the effort was organized around meeting a requirement rather than building a routine. The second firm's working group did more, but the additional work was structured to produce a durable asset: sub-metering, a product-level model, a supplier relationship. The difference is not hours worked. It is the organizational form of the work.

4.4 Capability Profiles and Sectoral Variation

4.4.1 Why Profiles Differ Across Sectors

The framework developed above is sector-relative, but the profiles that emerge in different sectors are not random. They reflect the factor intensity of production, the structure of value chain governance, and the regulatory exposure of the sector. Labor-intensive sectors tend to have truncated or compliance profiles because the resource base is thin and the returns to capability are uncertain. Technology-intensive sectors tend toward integrated profiles because the resource base supports absorptive capacity and the customer relationships reward differentiation. Capital-intensive sectors have distinctive profiles shaped by the scale of investment and the long asset lives.These patterns are the subject of Chapters 5 through 7, but a brief preview here clarifies how the capability construct connects to the sectoral analysis. The point is not that sectors determine capability; it is that the drivers operate differently in different factor-intensity contexts, and the measurement framework must be interpreted with that variation in mind.

4.4.2 Labor-Intensive Sectors: The Thin Foundation

In labor-intensive sectors such as furniture, textiles, and footwear, the typical firm has limited engineering capacity, a workforce with low formal environmental training, and thin financial slack. The resource base is the binding constraint on seizing and reconfiguring. Firms can execute discrete compliance projects, often with the help of equipment vendors, but they struggle to build routines that accumulate.

The value chain position in these sectors is often modular or captive. Firms manufacture to customer specifications, and the customer controls the design and the environmental requirements. The governance relationship transmits requirements but not always capability. The pilot found that furniture firms supplying European brands with relational relationships had higher capability scores than firms supplying domestic wholesalers, but even the higher-scoring firms had truncated profiles. They could sense and execute, but they could not build the human capital and data infrastructure that would make capability durable.

Managerial interpretation in these sectors is shaped by the immediacy of cost pressure. A furniture firm operating on thin margins has little managerial slack to think about a market shift that may or may not materialize. The interpretation that dominates is defensive: environmental requirements are a cost to be managed, not an opportunity to be seized. The interpretation is rational given the constraints, which is why changing it requires changing the constraints, not just providing information.

4.4.3 Technology-Intensive Sectors: The Absorptive Advantage

In technology-intensive sectors such as electronics and electrical equipment, the resource base is stronger. Firms have engineering teams, quality systems, and experience with product innovation. The absorptive capacity for environmental technologies is correspondingly higher. Firms can evaluate and adopt low-carbon process technologies, and they can redesign products for environmental performance without sacrificing function.

The value chain position in these sectors is often relational or modular with high capability requirements. Lead firms in electronics have been imposing environmental requirements for two decades, starting with restrictions on hazardous substances and extending to carbon footprint and circularity. The requirements have built up a body of supplier capability that is now available for carbon-specific applications. The pilot found that electronics firms had the highest average capability scores of the four sectors studied, with several firms approaching integrated profiles.

Managerial interpretation in these sectors tends to be strategic because the customer relationships reward it. A firm that can offer a lower-carbon product with verified data can win design wins and preferred supplier status. The return to capability is visible, which reinforces the interpretation. The mechanism is self-reinforcing in the opposite direction from the compliance trap: capability building produces position improvement, which produces slack, which supports further capability building.

4.4.4 Capital-Intensive Sectors: The Scale Constraint

In capital-intensive sectors such as cement, steel, and chemicals, the profile is shaped by the scale and longevity of assets. Process optimization and technology substitution require large capital investments with long payback periods. The seizing dimension is dominated by a small number of high-stakes decisions rather than a continuous routine. The reconfiguring dimension is shaped by the need to manage stranded asset risk and to plan for process route changes that may take a decade to execute.

The value chain position in these sectors is often characterized by long-term contracts and regional market structures. The environmental requirements come from regulation more than from customers, although downstream customers in construction and automotive are beginning to impose embodied carbon requirements. The pilot found that cement firms had high scores on S1 (regulatory monitoring) and S5 (process optimization) but lower scores on S6 (product redesign) and S7 (supply relationship change). The profile reflects the sector's technological and market structure.

Managerial interpretation in these sectors is shaped by the capital cycle. A cement firm that has just invested in a new kiln line has a different interpretation of carbon regulation than a firm approaching a reinvestment decision. The interpretation is tied to the asset replacement cycle, which is why capability building in capital-intensive sectors is lumpy and path-dependent.

4.4.5 The Cross-Sector Pattern

The cross-sector pattern is that capability building is easiest where the resource base is strong, the value chain rewards differentiation, and the returns are visible. It is hardest where the resource base is thin, the value chain transmits requirements without resources, and the returns are uncertain. The pattern has an uncomfortable implication: the sectors that most need capability building, because they face the greatest relocation risk under carbon pricing, are the sectors where it is hardest to build. Chapter 5 takes up this implication directly.

4.5 Measuring Capability in Practice: Three Firm Cases

4.5.1 Case Selection

The three cases below are drawn from the pilot study. They are selected to illustrate the three recurring profiles: compliance, truncated, and integrated. The firms are anonymized, but the sector, location, and size are accurate. Each case presents the capability profile, the drivers that explain it, and the trajectory that the profile implies.

4.5.2 Case A: A Textile Firm in Shaoxing (Compliance Profile)

The firm dyes and finishes fabrics for domestic and export customers. It employs about 600 workers and has annual revenue of roughly 400 million yuan. Its main export market is Southeast Asia, with a smaller share to Europe.

The capability profile is compliance. S1 (regulatory monitoring) scores 1: the firm's environmental officer tracks local wastewater and energy regulations but does not monitor carbon policy. S2 (customer requirement detection) scores 1: European customers have begun asking for carbon data, but the requests are forwarded to the quality department and answered with generic statements. S3 (supply chain visibility) scores 0: the firm has no data on the environmental attributes of its dye and chemical inputs. S5 (process optimization) scores 2: the firm has invested in wastewater heat recovery and has an annual energy-saving target, driven by local energy quotas. S8 (compliance project execution) scores 2: projects are completed, but with cost overruns. S9 (human capital) scores 1: one part-time environmental officer, no engineering support. S10 (data infrastructure) scores 1: monthly utility bills, no sub-metering. S12 (organizational authority) scores 1: environmental responsibility is informal, assigned to the environmental officer without budget authority.

The drivers explain the profile. Managerial interpretation is defensive: the general manager describes environmental requirements as "costs imposed from outside" and sees the firm's competitive advantage as cost and delivery speed. The resource base is thin: the firm has no in-house engineering team and relies on equipment vendors for process changes. The value chain position is modular: European customers provide specifications but no technical assistance, and the firm has no relational relationship with a lead firm. Cluster resources are available (Shaoxing has a textile research institute and several environmental service providers), but the firm has not used them. Slack is minimal: margins are thin and the firm is focused on maintaining volume.

The trajectory is toward relocation risk. The firm's export share to Europe is small but growing, and the CBAM transitional reporting requirements will eventually reach its products through downstream customers. The firm has no capability to respond beyond compliance, and its compliance investments are raising costs without improving position. The escape route would require a change in interpretation, which in turn would require exposure to a customer or peer that demonstrates the return to capability. Absent that, the firm is likely to retreat to domestic and Southeast Asian markets as European requirements tighten.

4.5.3 Case B: An Electronics Components Firm in Dongguan (Truncated Profile)

The firm makes power supply components for consumer electronics and industrial equipment. It employs about 1,200 workers and has annual revenue of roughly 1.5 billion yuan. Its customers are mainly original equipment manufacturers in China and Taiwan, with some direct exports to Europe.

The capability profile is truncated. S1 scores 3: the firm has a dedicated compliance team that tracks environmental regulation in all its markets, including carbon policy. S2 scores 3: customer requirements are routed to a cross-functional team and integrated into product planning. S3 scores 2: the firm has environmental data for its main suppliers, mostly estimated. S4 scores 2: the firm scans technology at trade fairs and through supplier relationships, but has no formal roadmap. S5 scores 2: annual efficiency projects with targets, but no continuous improvement system. S6 scores 2: design for environment is applied to some product lines. S7 scores 1: environmental criteria are in some contracts but not used to drive supplier change. S9 scores 2: one dedicated environmental engineer, no team. S10 scores 2: sub-metered energy data, product-level estimates. S11 scores 2: regular relationships with certifiers and main suppliers, no research partnerships. S12 scores 2: a formal environmental committee that meets quarterly, but without budget authority.

The drivers explain the profile. Managerial interpretation is strategic but constrained: the general manager understands the market shift and has invested in sensing and compliance, but has not committed to the reconfiguring investments that would make capability durable. The resource base is moderate: the firm has engineering capacity but it is fully committed to current production. The value chain position is relational with some customers, which provides requirements and some technical assistance, but the customers have not co-invested in capability. Cluster resources are available (Dongguan has a strong electronics supply base and several environmental service providers), and the firm uses them for compliance but not for capability building. Slack is moderate: the firm is profitable but reinvesting in capacity expansion.

The trajectory is uncertain. The firm can survive current requirements and may be able to climb if it commits to reconfiguring. The binding constraint is managerial attention: the general manager sees the opportunity but has not allocated the resources to build the human capital and data infrastructure that would make capability durable. The escape route would require a decision to invest in reconfiguring before the market forces it, which is a classic exploration-exploitation trade-off. The firm's profile suggests it will wait until the requirement is unavoidable, which may be too late to capture the position improvement.

4.5.4 Case C: A Cement Firm in Hubei (Integrated Profile)

The firm operates two kiln lines with a combined capacity of about 5 million tonnes per year. It employs about 2,000 workers and has annual revenue of roughly 3 billion yuan. Its market is regional, with some sales to infrastructure projects that have begun to specify low-carbon cement.

The capability profile is integrated. S1 scores 3: a dedicated team tracks carbon policy at national and provincial levels, including ETS allocation rules and the emerging embodied carbon standards for construction. S2 scores 3: customer requirements are anticipated and integrated into product planning, including the development of a lower-carbon cement blend. S3 scores 2: the firm has data on the embodied carbon of its main inputs, including alternative fuels and supplementary cementitious materials. S4 scores 3: the firm has a formal technology roadmap that includes alternative fuel substitution, energy efficiency, and long-term process route options. S5 scores 3: continuous improvement system with sub-metering on all major processes. S6 scores 2: product redesign is limited by the sector's technology, but the firm has developed lower-carbon blends. S7 scores 2: environmental criteria in main supplier contracts, including alternative fuel suppliers. S8 scores 3: projects completed on time with documented learning. S9 scores 3: an environmental engineering team with management access. S10 scores 3: integrated data system with measured process data. S11 scores 3: research partnerships with a local university and engagement with downstream customers on embodied carbon. S12 scores 3: a sustainability committee with budget authority and environmental metrics in performance evaluation.

The drivers explain the profile. Managerial interpretation is strategic and long-term: the general manager has been engaged with carbon policy since the ETS pilot phase and sees the firm's future as depending on its ability to produce lower-carbon cement at competitive cost. The resource base is strong: the firm has process engineering expertise and a history of investment in efficiency. The value chain position is regional and relational: downstream customers are beginning to demand embodied carbon data, and the firm has engaged with them to shape the standards. Cluster resources are strong (Hubei has a cement research institute and a developing alternative fuel supply chain), and the firm uses them. Slack is adequate: the firm is profitable and has a multi-year investment plan.

The trajectory is toward position improvement. The firm is positioned to capture the emerging market for lower-carbon cement and to manage the transition to the ETS expansion. The integrated profile does not guarantee success, because the technology and policy environment is uncertain, but it gives the firm the capacity to adapt. The case also shows that integrated profiles are possible in capital-intensive sectors, where the scale of investment supports the reconfiguring dimension.

4.5.5 Cross-Case Observations

The profile is not determined by size. The textile firm and the electronics firm are both medium-sized, but their profiles differ because their drivers differ. Size provides slack, but slack is only one driver. The binding constraint is often organizational rather than financial. The electronics firm has the resources to build reconfiguring capacity but has not allocated them. The textile firm lacks both resources and interpretation. The cement firm has both. And the trajectory is path-dependent. The cement firm's integrated profile is the product of a decade of investment and interpretation. It cannot be replicated quickly by a firm starting from a compliance profile.

4.6 Operationalizing the Framework: A Diagnostic Protocol

4.6.1 Purpose and Audience

The framework is intended for use by researchers, but it can also be used by firms and policymakers as a diagnostic. The diagnostic protocol below is a simplified version that a firm's management team can apply internally, or that a trade association or local government can apply to a group of firms. It is not a substitute for the full research protocol, but it produces a profile that is useful for identifying binding constraints and prioritizing investments.

4.6.2 The Diagnostic Steps

Assemble the team. The diagnostic should involve the general manager, the production manager, the person responsible for environmental compliance, and the person responsible for customer relationships. A team of four to six is sufficient. The team should meet for a half-day workshop.

Score the indicators. Using the scoring anchors in Table 4-2, the team scores each of the twelve indicators. The scoring should be evidence-based: the team should identify the specific evidence that supports each score. Where the team disagrees, the disagreement is itself informative and should be recorded.

Identify the profile. The team plots the scores by dimension and identifies the profile: compliance, truncated, or integrated. The profile is a starting point for discussion, not a label.

Identify the binding constraint. The team asks: which dimension is weakest, and which driver explains the weakness? If sensing is weak, the constraint may be managerial attention or information access. If seizing is weak, the constraint may be engineering capacity or supplier relationships. If reconfiguring is weak, the constraint may be slack or organizational authority.

Prioritize actions. The team identifies two or three actions that address the binding constraint. The actions should be specific and feasible within a twelve-month horizon. Examples: commissioning sub-metering on the main production lines (reconfiguring), assigning a manager to track carbon policy (sensing), or negotiating environmental criteria into the next supplier contract (seizing).

Set a review date. The team sets a date six or twelve months out to re-score the indicators and assess progress. The review is the mechanism that converts the diagnostic into a routine.

4.6.3 Limitations of the Diagnostic

The diagnostic is a self-assessment, and self-assessments are vulnerable to optimism and to strategic misrepresentation. The team should be encouraged to be candid, and the review should focus on evidence rather than on stated intentions. The diagnostic is not a substitute for external verification, but it is a low-cost way to start the capability-building process.

The diagnostic also cannot resolve the deeper constraints that the research framework identifies. A firm with a thin resource base cannot build an integrated profile through a workshop. The diagnostic can identify the constraint, but addressing it may require external resources: cluster services, customer assistance, or policy support. The diagnostic is a first step, not a solution.

4.7 Capability, Position, and the Logic of the Book

4.7.1 Connecting Capability to Value Chain Climbing

The connecting thread of this chapter is that capability is the mechanism through which environmental pressure translates into position change. Firms with capability convert pressure into differentiation; firms without capability convert pressure into cost.

The value chain literature has long argued that upgrading is a process of moving to higher-value activities (Gereffi, Humphrey, and Sturgeon, 2005; Humphrey and Schmitz, 2002). The literature has focused on the activities: moving from assembly to design, from manufacturing to branding. The argument of this book is that environmental capability is a route to upgrading that is increasingly available and increasingly necessary. It is available because lead firms are demanding environmental performance and are willing to reward suppliers who provide it. It is necessary because the regulatory floor is rising and firms that cannot build above it will lose position.

The route is not automatic. The sectoral chapters show that the route is easier in some sectors than others, and that the drivers operate differently across factor intensities. The capability construct provides the common language for comparing the sectors and for identifying the conditions under which the route is open.

4.7.2 The Endogenous Turn

The emphasis on endogenous drivers is a deliberate analytical choice. The alternative would be to treat capability as a function of external pressure: the stronger the regulation, the more capability firms build. That alternative is empirically inadequate, because firms facing the same pressure build different capabilities. It is also policy-naive, because it implies that raising the floor is sufficient to produce capability. The endogenous turn says that the floor is necessary but not sufficient, and that the mechanisms inside the firm and its cluster determine the response.

The turn has a cost. It makes the analysis more complex, because it requires attention to organizational and relational factors that are harder to measure than regulation. It also makes the policy implications less tidy, because there is no single lever that produces capability. The benefit is that the analysis is more accurate and more useful: it identifies where the binding constraints are and what would relax them.

4.7.3 Anticipating Objections

Three objections are worth addressing. The first is that the capability construct is too broad to be useful: it includes sensing, seizing, and reconfiguring, which are almost everything a firm does. The response is that the construct is defined by its object, environmental governance, and by its purpose, explaining position change. The dimensions are broad, but the indicators are specific, and the framework distinguishes firms in ways that predict trajectories. The pilot data support the claim that the construct is doing analytical work.

The second objection is that the framework is subjective: the scoring relies on interviewer judgment, and the indicators are not standardized. The response is that all capability measurement is subjective to some degree, and the alternative approaches (certification, disclosure scores) are also imperfect. The framework's advantage is that it is explicit about its judgments and that it can be replicated. The inter-rater reliability in the pilot was acceptable for most indicators, and the protocol can be tightened where it was not.

The third objection is that the endogenous drivers are not really endogenous: managerial interpretation, value chain position, and cluster resources are all shaped by external factors. The response is that the distinction is analytical, not ontological. The drivers are internal in the sense that they operate within the firm and its immediate environment, and they mediate the effect of external factors. The point is not that external factors do not matter; it is that they matter through the drivers. The policy implication is that external factors should be designed to activate the drivers, not to bypass them.

4.7.4 What Comes Next

Chapter 5 applies the framework to labor-intensive industries, where the compliance trap is most acute and the relocation risk is highest. The chapter examines why the thin resource base and modular value chain position constrain capability building, and what conditions would allow firms to escape the trap. Chapter 6 turns to technology-intensive industries, where the absorptive advantage supports integrated profiles. Chapter 7 examines capital-intensive industries, where scale and asset longevity shape the capability profile. Chapter 8 applies the framework to the furniture industry, the book's extended case, and traces the interaction between capability, export competitiveness, and relocation under CBAM.

The common thread across the chapters is the distinction between complying and climbing. The distinction is not moral: firms that comply are doing what the regulation requires, and many of them are doing it well. The distinction is analytical: compliance and capability are different organizational achievements, and they lead to different trajectories under a rising floor.

The first CBAM reporting deadline for the furniture sector's downstream customers falls in 2026, when importers must submit verified embedded emissions data rather than default values. For the Shunde firm that submitted industry averages in 2023, that date marks the end of the transitional tolerance. For its competitor, it marks the first year of a preferred supplier contract. Chapter 5 begins with a textile exporter in Shaoxing that faces the same deadline with neither the data nor the customer relationship to meet it, and asks what, if anything, can be done for firms that enter the transition without the capability to absorb it.