Chapter 5: Labor-Intensive Industries: Compliance Pressure, Relocation Risk, and the Limits of Cost-Based Upgrading

The Shunde furniture factory that anchors this book's argument sits at the end of a low-rise industrial road in Foshan, Guangdong, its loading bays facing a canal that carries container barges to the Pearl River delta terminals. On a Tuesday morning in March 2024, the factory manager, a man in his early fifties named Liang, walked me through the assembly hall where workers were fitting hinges to lacquered cabinet doors. The hall smelled of solvent and sawdust. Liang stopped at a pallet of finished panels and pointed to a barcode on the shrink wrap. That label, he said, was the most expensive thing in the building. It carried a carbon footprint declaration his largest European customer now demanded, and producing it had required him to hire a consultant, install sub-metering on three production lines, and renegotiate contracts with two upstream panel suppliers who could not document their own emissions. The label cost him roughly 40,000 yuan a year to maintain. His margin on the cabinet line was 6 percent. "This is not a green factory," he said, gesturing at the hall. "This is a factory that has learned to fill in the forms."

Liang's formulation captures the central problem of this chapter. Labor-intensive industries in China—furniture, textiles and apparel, footwear, toys, leather goods, basic electronics assembly, food processing—confront a carbon regulatory environment designed principally for the sectors where emissions are concentrated and measurable at the point of combustion or chemical reaction. Cement, steel, aluminum, chemicals, and power generation account for the overwhelming share of industrial greenhouse gas emissions, and the policy instruments now reshaping global trade—the European Union's Emissions Trading System (EU ETS), its Carbon Border Adjustment Mechanism (CBAM), China's national ETS and its forthcoming expansion—target those sectors first. Labor-intensive industries sit downstream of the carbon-intensive core. Their direct emissions are small. Their indirect emissions, embedded in purchased electricity and in the materials they process, are larger but diffuse. Their exposure to carbon regulation runs not through the smokestack but through the customer, the brand, the retailer, and the border.

That distinction matters analytically. Chapter 4 argued that green governance capability is an organizational achievement measurable along dimensions of measurement, abatement, supply-chain coordination, and strategic foresight, and that it is endogenous to the firm's position in the value chain rather than a passive response to regulation. For the labor-intensive firm, the capability question is posed in its sharpest form. The firm faces compliance pressure that is real but indirect, mediated by lead firms and importers who pass down requirements they themselves face upstream. It faces relocation pressure that is equally real, as rising labor costs, land costs, and now carbon-related documentation costs push low-value assembly toward Vietnam, Bangladesh, Indonesia, and inland Chinese provinces. And it faces a cost-based upgrading trap: the strategies that kept it competitive for three decades—scale, labor cost suppression, tax incentives, cheap land—do not generate the capabilities that green competitiveness requires, and may actively foreclose them.

This chapter examines these three pressures in turn. Section 5.1 analyzes compliance pressure and its cost impacts on labor-intensive firms, distinguishing direct from indirect exposure and tracing the transmission channels through which carbon rules reach the factory floor. Section 5.2 examines relocation dynamics, asking which firms move, where they go, and what the consequences are for both the sending and receiving regions. Section 5.3 confronts the limits of cost-based upgrading, arguing that the cost advantages that built China's labor-intensive export machine are structurally incapable of producing green competitive advantage, and that the firms which will survive the transition are those that convert compliance into capability. Throughout, the analysis draws on the furniture cluster in Shunde and Foshan as an extended case, supplemented by evidence from textiles, apparel, and electronics assembly.

5.1 Compliance Pressure and Cost Impacts

5.1.1 The Architecture of Indirect Exposure

A cement plant covered by the EU ETS or China's national ETS faces a direct obligation: it must surrender allowances for each ton of CO2 it emits, and the allowance price enters its cost structure as a line item. A furniture factory in Shunde faces no such obligation. Its direct emissions—from the combustion of natural gas in drying ovens, from the solvents in lacquers and adhesives, from the diesel in its forklifts—are small relative to its output value, and in most jurisdictions they fall below the thresholds that trigger direct regulation. The factory's exposure runs through three channels: the requirements its customers impose because they face regulation upstream; the carbon price embedded in its electricity bill; and the border measures that apply to its products when they reach the importing market.

The customer channel is the most consequential and the least visible in aggregate statistics. When a European retailer sources cabinet doors from Shunde, it does so under a supply-chain code of conduct that now includes carbon reporting requirements. The retailer's own customers—consumers, institutional buyers, regulators—increasingly demand product-level carbon footprints. The retailer therefore requires its suppliers to disclose emissions data, and the disclosure requirement cascades down the chain. This is the mechanism that the global value chain literature calls "private governance" (Gereffi, Humphrey, and Sturgeon, 2005; Mayer and Gereffi, 2010), and it has been extended from labor standards and product quality to carbon. The requirement is contractual rather than statutory, but its effect on the supplier is no less binding: a factory that cannot produce the data loses the order.

Electricity carries the second channel, and the carbon price embedded in it. China's national ETS initially covered only the power sector, and the allowance price has been modest—in the range of 60 to 100 yuan per ton of CO2 in the 2021–2024 period, well below the EU ETS price of 60 to 100 euros over the same period. But the power sector ETS does feed through to industrial electricity tariffs, and the direction of policy is toward higher and more broadly applied carbon pricing. For a labor-intensive firm, electricity is a meaningful but not dominant cost—typically 3 to 8 percent of total production cost in furniture, higher in some textile processes such as dyeing and finishing. A carbon price of 100 yuan per ton, applied to an electricity grid with an emissions intensity of roughly 0.55 kg CO2 per kWh, adds about 5.5 fen per kWh, or roughly 5 to 8 percent to the electricity bill. That is a real cost increase but not a transformative one. The transformative pressure comes from the customer channel and from border adjustment.

The EU's CBAM entered its transitional reporting phase in October 2023 and will impose a carbon levy on imports of cement, iron and steel, aluminum, fertilizers, electricity, and hydrogen from 2026. Furniture is not currently a covered sector. But the logic of CBAM—that imports should bear the same carbon cost as domestic production—creates a template that can be extended, and the EU has signaled that it will consider extension to downstream products once the administrative machinery is in place. More immediately, the reporting requirements of CBAM's transitional phase already reach furniture indirectly: a European importer of furniture that incorporates steel components, aluminum fittings, or glass must account for the embedded carbon in those inputs, and must therefore obtain data from its suppliers. The furniture factory that sources Chinese steel hinges and aluminum handles is drawn into the CBAM reporting chain even though its own product is not covered.

The cumulative effect of these three channels is a compliance burden that is small in absolute cost but large in administrative and organizational demand. Liang's 40,000 yuan annual cost for the carbon label is not a crippling sum. What matters is what the label requires: sub-metering, supplier documentation, a consultant, and the manager's time. For a firm with 200 workers and annual revenue of 80 million yuan, the compliance apparatus consumes resources that would otherwise go into product development, worker training, or market expansion. The burden is regressive: it falls proportionally harder on smaller firms with thinner management capacity.

5.1.2 Measuring the Cost Impact: Evidence from the Furniture Cluster

Precise measurement of compliance costs in labor-intensive industries is difficult because the costs are dispersed across administrative, energy, material, and certification categories, and because firms have incentives to underreport them. The evidence assembled here comes from a survey of 84 furniture firms in the Shunde and Foshan clusters conducted between 2022 and 2024, supplemented by interviews with 23 firm managers and 9 industry association officials. The survey was not a random sample; it overrepresents firms with export exposure, which are the firms most affected by carbon requirements. The findings should be read as indicative of the upper bound of compliance pressure rather than as representative of the cluster as a whole.

Table 5-1 summarizes the reported cost impacts by firm size and export orientation. The pattern is consistent with the transmission-channel analysis. Firms that export to the EU report compliance costs averaging 1.8 percent of revenue, compared with 0.6 percent for firms exporting only to non-EU markets and 0.2 percent for firms selling domestically. The EU-exporting firms are larger on average (median revenue 120 million yuan versus 45 million yuan for domestic firms), which means the absolute burden is concentrated in the segment of the cluster that is most integrated into global value chains. The smallest firms, which lack the management capacity to produce carbon documentation, are increasingly excluded from EU-facing supply chains altogether—a selection effect that shows up in the survey as a decline in the number of small firms reporting EU customers.

Table 5-1: Reported carbon compliance costs as share of revenue, Shunde/Foshan furniture firms, 2022–2024

Source: Author's survey of 84 firms, Shunde and Foshan, 2022–2024. Compliance cost includes certification, metering, consulting, audit, and documented staff time; excludes energy cost increases and material cost pass-through.

Two features of the table deserve emphasis. Compliance cost as a share of revenue is higher for small EU-exporting firms than for large ones, even though the absolute cost is lower. The fixed component of compliance—certification, consultant engagement, the installation of monitoring equipment—does not scale with firm size, so it weighs more heavily on smaller firms. This is the regressivity noted above, and it has a dynamic consequence: it accelerates consolidation in the cluster, pushing orders toward larger firms that can absorb the fixed cost. The cost categories are also predominantly administrative rather than technical. The firms are not investing in abatement technology; they are investing in the measurement and documentation apparatus that allows them to demonstrate compliance to customers. Resources flow into proving that the firm is not a problem rather than into becoming a solution.

The survey also asked firms to identify the single most burdensome requirement. The most frequent response (cited by 41 of 84 firms) was the product carbon footprint calculation, which requires data on the emissions embodied in purchased materials—the panels, veneers, adhesives, finishes, hardware, and packaging that make up a cabinet or a chair. The firm's own operations are relatively easy to meter. The supply chain is not. A furniture factory may purchase medium-density fiberboard from one supplier, veneer from another, water-based lacquer from a third, and steel hinges from a fourth, each of which has its own upstream emissions that must be estimated or obtained. The data infrastructure for product-level carbon accounting in Chinese furniture supply chains is thin. Firms rely on default emission factors from databases that may not reflect actual supplier performance, and they have limited leverage to demand primary data from upstream suppliers who are themselves small and poorly instrumented.

This is where the compliance burden becomes an organizational capability problem. The firm that can obtain primary emissions data from its suppliers, verify it, and incorporate it into a product footprint has built a supply-chain coordination capability that its competitors lack. The firm that relies on default factors has complied with the letter of the requirement but has not built the capability. The distinction between these two firms is the distinction between complying and climbing, and it is not captured by the compliance cost figures in Table 5-1. The cost figures measure what the firm spends; the capability question is what the firm learns.

5.1.3 The Compliance Trap in Labor-Intensive Sectors

The concept of the compliance trap, introduced in Chapter 1, requires specification for labor-intensive industries. In capital-intensive sectors, the trap takes the form of investing in end-of-pipe abatement that satisfies the regulator without changing the production process—scrubbers on a cement kiln, for example, that reduce SO2 and particulates but leave the CO2 intensity of the clinker unchanged. In labor-intensive sectors, the trap takes a different form: the firm builds an administrative apparatus for measuring and reporting emissions without building the technical and organizational capacity to reduce them. The apparatus becomes a fixed cost, a barrier to entry that protects incumbents, and a substitute for the harder work of process innovation.

Consider the trajectory of a mid-sized Shunde firm that we will call Hengmei, which produces upholstered sofas for the European market. In 2021, Hengmei's largest customer, a German retailer, informed the firm that from 2023 it would require a product carbon footprint declaration for every SKU. Hengmei hired a consultant, installed electricity sub-meters on its cutting, sewing, and upholstery lines, and engaged its three main fabric suppliers to provide emission factors. The declaration was produced on time. The retailer was satisfied. Hengmei's compliance cost was approximately 1.4 percent of revenue.

What Hengmei did not do was change its production process. The foam it used was still petroleum-based. The fabric was still conventionally dyed. The factory's electricity was still drawn from the grid at the regional average emission intensity. The carbon footprint declaration documented a footprint that was, by the standards of the European market, high. The retailer accepted it because the requirement at that stage was disclosure, not performance. But the retailer also began, in 2023, to rank suppliers by footprint and to shift volume toward the lower-carbon suppliers on its list. Hengmei's ranking was middling. Its volume held for one season and then began to slip. By 2024, the firm was investing in a photovoltaic array on its roof and switching to water-based adhesives—not because the regulation required it, but because the customer's sourcing algorithm rewarded it. The compliance apparatus had bought Hengmei time, but it had not bought competitive advantage. The firm was now spending on abatement under duress, with a two-year lag on the competitors that had moved earlier.

Hengmei's trajectory illustrates the trap's mechanism. Compliance investment is necessary but not sufficient. It satisfies the immediate requirement and preserves the customer relationship. But it does not change the firm's cost position relative to competitors, does not differentiate its product, and does not build the capabilities that will matter when the requirement tightens from disclosure to performance. The firm that treats compliance as the endpoint finds itself repeatedly catching up, investing under pressure, and competing on the same cost dimensions as before. The firm that treats compliance as the entry ticket to a longer capability-building process—using the measurement infrastructure to identify abatement opportunities, using the supplier engagement to renegotiate material specifications, using the customer relationship to co-develop lower-carbon products—converts the compliance cost into an investment.

The trap is not inevitable. It is a function of how the firm frames the requirement and how much organizational slack it has to invest beyond the minimum. Firms with strong management, access to finance, and a customer relationship that supports joint problem-solving are more likely to escape it. Firms that are marginal, financially stressed, and dependent on a single customer for survival are more likely to be caught. This suggests that the compliance burden, by accelerating consolidation, may paradoxically improve the cluster's average capability: the firms that survive are the ones with the resources to invest. But consolidation also destroys employment and disrupts the local industrial ecosystem, and the workers who lose their jobs in the consolidating firms do not automatically find them in the survivors.

5.1.4 Heterogeneity Within the Labor-Intensive Category

The category "labor-intensive" covers a wide range of industries with different emissions profiles, different regulatory exposures, and different upgrading trajectories. Compliance pressure and capability response differ systematically across sub-sectors.

Textiles and apparel present the most developed compliance infrastructure, largely because the sector has been the target of private governance for two decades. The apparel industry's sustainability codes—the Higg Index, the Zero Discharge of Hazardous Chemicals (ZDHC) program, the Science Based Targets initiative (SBTi) commitments adopted by major brands—have created a data architecture that extends from brand to supplier. Chinese apparel firms that supply global brands are accustomed to audit and disclosure. The marginal cost of adding carbon reporting to the existing apparatus is lower than in furniture, where the apparatus had to be built from scratch. But the apparel sector also faces the most intense relocation pressure, because the labor cost differential with Bangladesh and Vietnam is large and the production process is relatively easy to transfer. The compliance capability that Chinese apparel firms have built is therefore at risk of being stranded as production moves.

Footwear occupies an intermediate position. The sector is dominated by a small number of brand lead firms (Nike, Adidas, Puma) that have made carbon reduction a central element of their sourcing strategy. Chinese footwear suppliers are under pressure to reduce emissions, and the larger ones have invested in energy efficiency and renewable electricity. But footwear production is also highly mobile, and the brands have actively diversified their sourcing toward Vietnam and Indonesia. The compliance pressure and the relocation pressure are pushing in the same direction: toward a smaller, more capable Chinese supplier base serving the higher-value segments of the market, and away from the mass-production segments that have moved offshore.

Toys and basic electronics assembly present a different profile. The compliance pressure is lower because the products are less exposed to EU carbon regulation and because the brand lead firms in these sectors have been slower to adopt carbon requirements. The relocation pressure is high for toys, which are labor-intensive and price-sensitive, and moderate for electronics assembly, which is more capital-intensive than it appears and is embedded in complex supply chains that are difficult to move. The capability response in these sectors is correspondingly weaker: firms have less incentive to invest in green governance because the market does not reward it and the regulatory pressure is distant.

Furniture, the focus of this book's extended case, sits at the intersection of these profiles. It is labor-intensive in its assembly operations but capital-intensive in its panel production and finishing. It is exposed to EU regulation indirectly through its customers and directly through the CBAM's potential extension. It is relocating, but more slowly than apparel, because the production process is more automated and the supply chain is more localized. And it has a domestic market that is large enough to sustain firms that do not export. These features make furniture a useful lens for observing the full range of compliance and capability dynamics, from the small firm that complies minimally to the large firm that is converting compliance into a differentiation strategy.

5.2 Relocation Dynamics and Risks

5.2.1 The Relocation Calculus

The decision to relocate production is never a single-variable calculation. Firms weigh labor costs, land costs, energy costs, logistics, supplier proximity, tariff exposure, regulatory burden, and political risk. Carbon regulation enters this calculus as one factor among many, and in most labor-intensive sectors it is not yet the decisive one. But it is becoming more salient, and its interaction with the other factors is what makes relocation dynamics in the dual carbon era distinctive.

The standard relocation narrative for Chinese labor-intensive industries runs as follows. Rising wages in the Pearl River Delta and Yangtze River Delta have eroded the cost advantage that built China's export machine. Minimum wages in Guangdong rose from roughly 1,500 yuan per month in 2010 to over 2,300 yuan in 2023, and actual manufacturing wages rose faster. Land costs in the coastal clusters have risen sharply as industrial land is converted to urban uses. Environmental regulation has tightened, raising compliance costs for firms that were previously able to externalize pollution. The combination has pushed low-value, labor-intensive production toward inland provinces (Henan, Sichuan, Anhui) and toward Southeast Asia (Vietnam, Indonesia, Cambodia, Bangladesh). Carbon regulation adds a new layer to this story: firms that cannot document their emissions lose access to premium markets, and firms that can document them face costs that further erode their margin.

The evidence on relocation is mixed and the aggregate statistics are misleading. China's outward foreign direct investment in labor-intensive manufacturing has grown, but much of it is market-seeking rather than efficiency-seeking, and much of the production that has moved to Southeast Asia is final assembly that sources components from China. The "relocation" is often a regional division of labor within a China-centered production network rather than a wholesale transfer. A furniture firm in Shunde may open a factory in Vietnam to serve the US market while keeping its component production and product development in Guangdong. The carbon accounting for such a firm becomes complicated: the emissions embodied in the Chinese components are attributed to the Vietnamese final product, and the Vietnamese factory's compliance with EU or US requirements depends on data from its Chinese supplier.

This pattern—which the global value chain literature calls "factory Asia" or "the China-centered production network" (Baldwin, 2016; Gereffi, 2018)—has implications for how we think about relocation and carbon. The carbon footprint of a product is not determined by the location of final assembly. It is determined by the emissions embodied in the entire production chain, and if the chain remains anchored in China, the relocation of assembly does not reduce the footprint. It may increase it, because the additional transport adds emissions and because the Vietnamese factory may be less energy-efficient than the Chinese one it replaced. The relocation that is motivated by labor cost and tariff avoidance can therefore be carbon-intensive, even as it is presented as diversification.

5.2.2 Who Relocates and Who Stays

The relocation decision is heterogeneous across firms, and the heterogeneity is patterned. Firms with low margins, standardized products, and arm's-length customer relationships are the most likely to relocate. Firms with higher margins, differentiated products, and relational customer relationships are the most likely to stay. Carbon regulation reinforces this pattern: firms that have invested in compliance capability have a reason to stay, because their investment is embedded in the cluster's infrastructure and supply chain, and moving would strand it. Firms that have not invested have less to lose and more incentive to seek lower-cost locations.

The survey of Shunde furniture firms provides some evidence on this pattern. Among the 84 firms surveyed, 19 reported that they had established or were planning to establish production outside Guangdong, and 11 reported production outside China. The relocating firms were smaller on average than the non-relocating firms (median revenue 38 million yuan versus 72 million yuan) and had lower export exposure to the EU (12 percent of revenue versus 31 percent). The non-relocating firms cited supplier proximity, skilled labor, and customer relationships as reasons to stay; the relocating firms cited labor cost, land cost, and environmental compliance cost as reasons to leave. None of the relocating firms cited carbon regulation as the primary driver, but several mentioned it as a contributing factor—particularly the cost of meeting customer documentation requirements, which they expected to be lower in Vietnam or inland China.

This finding should be interpreted cautiously. The survey captures stated intentions, not realized outcomes, and firms may rationalize relocation decisions that are driven by other factors. But it is consistent with the broader evidence that carbon regulation is a second-order driver of relocation in labor-intensive sectors, behind labor cost and market access. The first-order driver remains the cost differential, and that differential is large: manufacturing wages in Vietnam are roughly one-third of those in Guangdong, and in Bangladesh they are lower still. Carbon compliance costs of 1 to 2 percent of revenue do not close that gap. They may tip the decision at the margin for firms that are already considering relocation, but they do not create the incentive on their own.

The firms that stay are not necessarily the firms that are climbing. Some stay because they are locked into the cluster—their suppliers, their workers, their customers are all there, and the cost of moving exceeds the benefit. Others stay because they have upgraded into segments where labor cost is a smaller share of total cost and where proximity to suppliers and customers matters more. The distinction between these two groups is the distinction between firms that are stuck and firms that are rooted. The stuck firms will eventually move or exit. The rooted firms are the ones that can build green capability, because they have the margin and the stability to invest.

5.2.3 Consequences for Sending Regions: The Cluster Hollowing Problem

The relocation of labor-intensive production has consequences for the clusters that lose it. The most visible consequence is employment loss. The Shunde furniture cluster employed roughly 200,000 workers at its peak in the mid-2010s; by 2023, the figure had fallen to an estimated 120,000, with the decline concentrated in assembly operations that had moved inland or offshore. The workers who lost these jobs were predominantly older, less educated, and less mobile than the average manufacturing worker. Many were migrants from Guangxi, Hunan, and Sichuan who had built their lives around the cluster and who faced the choice of following the work to an inland province or returning to their home villages.

The less visible but more consequential effect is the erosion of the cluster's industrial ecosystem. The furniture cluster in Shunde is not just a collection of assembly factories. It is a dense network of panel producers, veneer suppliers, hardware manufacturers, lacquer and adhesive formulators, logistics providers, design studios, trade fairs, and industry associations. This ecosystem generates knowledge spillovers, reduces transaction costs, and supports the incremental innovation that keeps the cluster competitive. When assembly operations move out, the ecosystem loses volume and the suppliers that serve them lose scale. The panel producers, hardware manufacturers, and formulators that remain face a shrinking domestic market and may themselves relocate or consolidate. The cluster's knowledge base—the tacit skills of its workers, the relationships among its firms, the institutional memory of its associations—is depleted.

This is the cluster hollowing problem, and it has a specific implication for green capability. The capabilities that support green governance—supply-chain coordination, process innovation, product redesign—are embedded in the cluster's ecosystem. A firm that is isolated in a low-cost location, surrounded by suppliers that lack the capacity to document their emissions or innovate their processes, cannot build these capabilities. It can comply, by relying on default emission factors and third-party certification, but it cannot climb. The cluster, for all its cost disadvantages, is the organizational form that makes capability building possible. Relocation that hollows the cluster therefore undermines the very basis for green upgrading.

The policy response to hollowing has been uneven. Guangdong's provincial government has promoted "industrial upgrading" (产业升级) through subsidies for automation, support for design and branding, and the relocation of low-end production to designated industrial parks in the province's less developed regions. The "dual transfer" (双转移) policy, launched in 2008, aimed to move labor-intensive production from the Pearl River Delta to northern and western Guangdong while upgrading the delta's remaining industry. The policy had mixed results: some production moved as intended, but much of it moved to other provinces or countries, and the upgrading of the delta's industry was slower than hoped. The carbon dimension was absent from the policy design, and the relocation that occurred was not evaluated for its emissions consequences.

5.2.4 Consequences for Receiving Regions: The Race to the Bottom

The receiving regions—inland Chinese provinces, Vietnam, Indonesia, Bangladesh—face their own version of the relocation dynamic. They gain employment and industrial activity, but they also gain the environmental burden of labor-intensive production, and they face pressure to compete on cost, which creates a race to the bottom on environmental standards.

Vietnam is the most studied case. The country has become a major destination for Chinese textile, apparel, footwear, and furniture production, and its industrial zones have expanded rapidly. Vietnam's environmental regulation has tightened in response to public concern and international pressure, but enforcement remains uneven, and many industrial zones lack adequate wastewater treatment and emissions monitoring. The carbon intensity of Vietnam's electricity grid is higher than China's in some regions and lower in others, depending on the hydropower share. The net effect of relocating production from Guangdong to Vietnam on the product carbon footprint is therefore ambiguous: labor costs fall, but energy-related emissions may rise, and the additional transport adds emissions.

The race-to-the-bottom dynamic is not inevitable. Some receiving regions have used the relocation as an opportunity to leapfrog, attracting investment in cleaner production and renewable energy. Vietnam's solar and wind capacity has grown rapidly, and some industrial zones have invested in shared utility infrastructure that reduces emissions. But the leapfrogging requires governance capacity that many receiving regions lack, and the competitive pressure to attract investment on cost grounds works against it. The result is a patchwork: some receiving regions are cleaner than the Chinese clusters they are replacing, and some are dirtier.

For the sending regions, the environmental consequence of relocation is a form of leakage. The emissions embodied in the relocated production do not disappear; they move to a jurisdiction with weaker regulation and are attributed to the receiving country. This is the carbon leakage problem that CBAM was designed to address, and it applies to labor-intensive production as much as to steel and cement. The difference is that the leakage in labor-intensive sectors is harder to measure, because the emissions are diffuse and the supply chains are complex. A furniture factory that moves from Shunde to Vietnam may reduce its direct emissions (if the Vietnamese grid is cleaner) or increase them (if it is dirtier), and the net effect depends on the specifics of the production process and the energy source. The policy instruments that address leakage in capital-intensive sectors—border adjustment, free allocation, output-based rebates—are poorly suited to labor-intensive sectors, where the emissions are small and the administrative cost of measuring them is high.

5.2.5 Relocation as a Capability Question

Economic geography and global value chain studies have long treated relocation as a capability question rather than a pure cost calculation. Firms relocate when the capabilities they need are available in the destination and the capabilities they have built in the origin are not worth preserving. The classic example is the apparel industry, where production has moved repeatedly from high-wage to low-wage countries as the industry's capability requirements have remained relatively stable (Gereffi, 1999). The firms that survive relocation are those that can transfer their capabilities—their production knowledge, their customer relationships, their quality systems—to the new location. The firms that cannot transfer them are the ones that lose their customers and exit.

Carbon regulation changes the capability calculus in two ways. It raises the value of the capabilities embedded in the origin cluster. A firm that has built supply-chain coordination, measurement infrastructure, and customer trust in Shunde has assets that are not easily replicated in Vietnam. The firm that relocates must rebuild these assets, and the rebuilding takes time and investment. Carbon regulation also raises the cost of relocating to a jurisdiction with weaker environmental governance. A firm that moves to a country with a high-carbon grid and lax enforcement may find that its product footprint rises, which undermines its access to premium markets. The relocation that reduces labor cost may increase carbon cost, and the net effect may be negative.

Carbon regulation can therefore be a force for keeping production in China. The Chinese clusters have built compliance infrastructure—the consultants, the metering systems, the supplier networks, the certification bodies—that the receiving regions lack. The firm that stays in Shunde can use this infrastructure to demonstrate compliance and to build capability. The firm that moves to Vietnam must build the infrastructure from scratch, and it may find that the cost of doing so exceeds the labor cost savings. This is not an argument for complacency: the labor cost gap is large, and many firms will still find relocation profitable. But it is an argument that the relocation decision is more contingent than the standard cost narrative suggests, and that carbon regulation is one of the factors that makes it so.

5.3 Limits of Cost-Based Upgrading

5.3.1 The Cost-Based Model and Its Exhaustion

The model that built China's labor-intensive export machine was cost-based. It relied on low wages, cheap land, favorable tax treatment, and a large pool of workers willing to migrate from rural areas to coastal factories. The model was extraordinarily successful: China's share of global exports of labor-intensive goods rose from roughly 5 percent in 1990 to over 30 percent in 2015, and the clusters that produced these goods—Shunde in furniture, Dongguan in electronics, Shaoxing in textiles, Jinjiang in footwear—became some of the most productive industrial districts in the world.

The model is now exhausted. Wages have risen, land costs have risen, the pool of rural migrant labor has shrunk as the population ages, and the environmental externalities of the model have become politically unsustainable. The firms that relied on cost advantages face a choice: upgrade into higher-value segments where cost is a smaller share of the competitive equation, or relocate to lower-cost locations, or exit. The upgrading path is the one that policy has promoted, and it is the one that Humphrey and Schmitz (2002) theorized when they distinguished process upgrading, product upgrading, functional upgrading, and inter-sectoral upgrading. Gereffi (2005) added the governance dimension, showing that the upgrading path available to a firm depends on its position in the value chain and on the lead firm's sourcing strategy. But the upgrading path is harder than the policy rhetoric suggests, and the cost-based model leaves a legacy that constrains it.

The low margin is the first constraint. Firms that competed on cost had thin margins—3 to 8 percent in furniture, 2 to 5 percent in apparel—and thin margins leave little room for investment in the capabilities that upgrading requires. The firm that wants to move from assembly to design, from OEM to ODM, from commodity products to branded products, must invest in design staff, market research, brand building, and customer relationships. These investments have long payback periods and uncertain returns, and they are difficult to finance from retained earnings when the margin is 5 percent. The firm can borrow, but the collateral it can offer is its physical assets—land, buildings, machinery—which are themselves undervalued in a cluster that is perceived to be in decline. The financial system that supported the cost-based model, with its reliance on land-backed lending and its preference for short-term trade finance, is poorly suited to financing capability building.

Organizational legacy is the second constraint. Firms that competed on cost developed organizational routines optimized for cost minimization: tight control of labor time, minimal investment in training, adversarial relationships with suppliers, and a focus on volume rather than value. These routines are functional for cost competition but dysfunctional for capability building. The firm that wants to upgrade must change its routines—invest in worker skills, build collaborative supplier relationships, develop the marketing and design functions that it never needed before. Organizational change is slow and difficult, and it is resisted by managers and workers whose skills and status are tied to the old routines. Teece, Pisano, and Shuen (1997) and Nelson and Winter (1982) emphasize that firms are path-dependent: the capabilities they have built constrain the capabilities they can build. The cost-based firm is path-dependent in a way that makes the transition to green capability particularly hard.

Market position is the third constraint. Firms that competed on cost occupied the low-value segments of global value chains, where they were interchangeable suppliers to lead firms that controlled design, branding, and distribution. The lead firms set the terms of trade, captured the rents, and allocated the tasks. The supplier's upgrading path was constrained by the lead firm's sourcing strategy: a furniture factory could move from making components to making finished goods, but it could not easily move into branding or retail, because the lead firm controlled those segments and had no interest in sharing them. Gereffi, Humphrey, and Sturgeon (2005) call this "captive" governance, and it describes the position of many Chinese labor-intensive firms. The captive firm can improve its process and product within the parameters set by the lead firm, but it cannot change the parameters. Its upgrading is bounded by the governance structure of the chain.

5.3.2 Why Cost Advantages Do Not Translate into Green Advantages

Cost-based advantage rests on the ability to produce a given product at lower cost than competitors. Green advantage rests on the ability to produce a product with lower environmental impact, or to produce a product that enables customers to reduce their impact, in a way that commands a price premium or secures preferential access to markets. The two advantages are analytically distinct: a firm can be the low-cost producer and the high-carbon producer at the same time, and many Chinese labor-intensive firms are. The strategies that produce cost advantage—scale, labor intensification, cheap energy, lax environmental enforcement—are often the strategies that produce high emissions. The firm that competes on cost has an incentive to externalize environmental costs, not to internalize them.

The transition from cost advantage to green advantage requires the firm to change its production function, not just its cost position. It must invest in energy efficiency, renewable electricity, low-carbon materials, and product redesign. These investments have upfront costs and uncertain returns, and they are difficult to justify on cost grounds alone when the carbon price is low and the customer is not willing to pay a premium. The firm that invests in green capability is making a bet on the future direction of regulation and consumer demand. The firm that does not invest is making a bet that the future will resemble the past. The first bet is risky; the second is riskier, because the direction of policy is clear and the cost of being unprepared is rising.

The evidence from the Shunde furniture cluster suggests that few firms have made the first bet deliberately. The investments in photovoltaic arrays, water-based adhesives, and energy-efficient drying ovens that appear in the survey are mostly reactive—made in response to customer pressure or the expectation of regulation, not as part of a strategic plan to build green advantage. The firms that have made the bet deliberately are the larger ones, with revenue above 200 million yuan, that have a brand or a design capability and that see green as a way to differentiate. For the majority, green investment is a cost to be minimized, not a capability tobuilt.

This is the limit of cost-based upgrading in its starkest form. The firm that treats green investment as a cost will invest only to the extent that the customer or the regulator forces it, and it will choose the cheapest compliance option. The cheapest option is usually administrative—documentation, certification, default emission factors—rather than technical. The firm therefore complies without building capability, and it remains vulnerable to the next tightening of the requirement. The firm that treats green investment as a capability will invest beyond the minimum, because it expects the investment to pay off in market access, price premium, or regulatory headroom. The two firms face the same regulation, but they respond differently, and the difference compounds over time.

5.3.3 The Sectoral Evidence: Textiles, Apparel, and Footwear

The argument that cost advantages do not translate into green advantages is supported by evidence from the sectors that have gone furthest down the compliance path. Textiles and apparel are the most instructive, because the sector has the longest experience with private environmental governance and the most developed data infrastructure.

The apparel industry's compliance architecture was built in response to labor and chemical safety scandals, not carbon. The collapse of the Rana Plaza factory in Bangladesh in 2013, which killed more than 1,100 garment workers, triggered a wave of brand commitments to factory safety and supply-chain transparency. The Zero Discharge of Hazardous Chemicals (ZDHC) program, launched in 2011, created a common standard for chemical management in textile supply chains. The Higg Index, developed by the Sustainable Apparel Coalition, created a standardized tool for measuring the environmental performance of products and factories. These initiatives built the data infrastructure that carbon reporting now relies on. A Chinese apparel factory that has been audited under ZDHC and Higg has already done much of the work that carbon disclosure requires.

The result is that Chinese apparel firms are, on average, more advanced in their environmental governance than Chinese furniture firms. They have environmental managers, they track energy and water use, they have relationships with auditors and certifiers. But the capability they have built has not protected them from relocation pressure. The apparel firms that have invested most in environmental governance are also the firms most exposed to the loss of orders to Vietnam and Bangladesh, because the brands that demand environmental compliance are the same brands that are diversifying their sourcing. The capability is real, but it is not a moat. It does not prevent the brand from moving the order to a factory in Vietnam that has the same certifications at a lower labor cost.

This is a crucial finding, and it qualifies the capability-based view of green competitiveness that this book advances. Capability matters, but it matters within the governance structure of the value chain. In a captive chain, where the lead firm can switch suppliers at low cost, the supplier's capability does not translate into bargaining power. The supplier can comply, and it can even lead in compliance, but it cannot use compliance to change the terms of trade. The lead firm captures the value of the supplier's investment by demanding it without paying for it. The supplier bears the cost and the lead firm takes the credit.

The footwear sector shows a similar pattern. The brands that dominate the sector—Nike, Adidas, Puma—have made carbon reduction a central element of their sustainability strategies, and they have pushed their suppliers to invest in energy efficiency and renewable electricity. The Chinese suppliers that have invested most are the large ones that have long-term relationships with the brands and that produce high-value products. The smaller suppliers, which produce commodity footwear, have invested less and are more exposed to relocation. The capability divide within the Chinese footwear industry is widening, and it maps onto the divide between firms that are rooted in the brands' supply chains and firms that are marginal.

One detail from the footwear interviews is worth recording because it shows how the compliance apparatus can outrun the abatement it is supposed to measure. A supplier in Jinjiang that makes running shoes for two European brands installed energy meters on every production line in 2022, at the request of one of the brands. The meters showed that the factory's largest energy use was not the injection molding machines, as management had assumed, but the compressed air system, which leaked at roughly 30 percent of output. The firm fixed the leaks for less than 200,000 yuan and cut its electricity bill by 11 percent. The same firm then commissioned a carbon footprint study and discovered that the second-largest emissions source was the solvent-based adhesive used in sole bonding, for which a water-based substitute existed but required reformulating the bonding process and retraining the line workers. The firm has not made that change. The payback period was estimated at four years, the brand did not require it, and the line managers resisted the disruption. The cheap abatement was captured; the expensive abatement was deferred. The pattern is exactly what the marginal abatement logic predicts, and it left the firm with a footprint that was lower than its competitors' but still far above what a tightening standard would require.

5.3.4 The Furniture Case: Why the Cluster Matters

Furniture differs from apparel and footwear in a way that is analytically important. The furniture value chain is less concentrated at the brand end and more concentrated at the production end. The lead firms—retailers like IKEA, Ashley, and the large European buying groups—are powerful, but they do not dominate the chain to the same extent as the apparel brands. The production process is more capital-intensive, more automated, and more dependent on local supplier networks. The result is that the furniture cluster in Shunde has more bargaining power vis-à-vis its customers than the apparel cluster in Dongguan, and more of the value of the final product is captured in the cluster.

This difference matters for the capability argument. The furniture firm that invests in green capability has a better chance of capturing the return on its investment than the apparel firm, because the customer cannot easily switch to a supplier in another country that has the same capability. The furniture supply chain is localized: the panel producers, the hardware manufacturers, the lacquer formulators, and the assembly factories are clustered within a few dozen kilometers of each other, and the relationships among them are dense and long-standing. A customer that wants to source a low-carbon cabinet cannot simply move the order to Vietnam, because the Vietnamese supply chain does not have the same depth. It can move the assembly, but it must still source the components from China, and the carbon footprint of the final product will reflect the Chinese components.

The cluster is a capability asset in this sense. The density of the supplier network, the tacit knowledge of the workers, the institutional infrastructure of the trade fairs and associations, and the proximity of the design and engineering functions create a capability that is not easily replicated. The firm that is embedded in the cluster can draw on this capability to build green advantage: it can work with its panel supplier to develop a lower-carbon board, with its lacquer supplier to switch to water-based finishes, with its hardware supplier to reduce the weight of the fittings. These collaborations are possible because the suppliers are nearby, they know each other, and they have a shared interest in the cluster's survival. The firm that is isolated in a low-cost location cannot do this, because it does not have the supplier network to collaborate with.

The cluster is not a sufficient condition for green capability. The Shunde cluster also contains many firms that are stuck in the compliance trap, and the cluster's institutions—the associations, the technical centers, the trade fairs—have been slow to incorporate carbon into their programs. But the cluster is a necessary condition for the kind of capability building that goes beyond compliance. The policy implication is that the cluster, not the individual firm, is the unit at which green upgrading should be supported. The investments that matter—shared testing facilities, common carbon accounting platforms, supplier development programs, design centers for low-carbon products—are collective goods that no single firm can provide on its own.

5.3.5 The Limits of the Cost-Based Response to Carbon Pricing

The final limit of cost-based upgrading concerns the firm's response to carbon pricing itself. When a firm faces a carbon price, its first instinct is to minimize the cost by reducing emissions at the lowest marginal cost. This is the logic of the marginal abatement cost curve, and it is economically rational. The firm will invest in the cheap abatement options first—energy-efficient lighting, waste heat recovery, improved insulation—and it will defer the expensive options—process substitution, renewable electricity, material innovation—until the carbon price rises enough to justify them.

The problem is that the cheap abatement options are quickly exhausted in labor-intensive firms. The firm that has installed LED lighting, fixed its compressed air leaks, and insulated its drying ovens has captured most of the low-cost abatement potential. The next tier of options—rooftop solar, electrification of process heat, switching to low-carbon materials—requires capital investment that is an order of magnitude larger, and it requires organizational capabilities that the firm may not have. The firm that has optimized for cost has, by definition, already captured the cheap abatement, and it now faces a discontinuous jump in the cost of further reductions.

This discontinuity is the source of the compliance trap's persistence. The firm that has exhausted the cheap options cannot meet a tightening carbon requirement through incremental investment. It must either make a large, risky investment in transformative change or accept that it will fall behind competitors that have made the investment. The cost-based firm is structurally biased toward the second option, because its organizational routines, its financial structure, and its management mindset are all oriented toward incremental cost reduction rather than transformative investment. The firm that escapes the trap is the one that recognizes the discontinuity early and makes the transformative investment before it is forced to. That recognition is a strategic capability, and it is precisely the capability that cost-based firms lack.

The evidence from the Shunde survey supports this interpretation. The firms that had invested in rooftop solar—the largest single abatement investment observed in the cluster—were not the firms with the highest energy costs or the highest carbon exposure. They were the firms with the strongest management, the best access to finance, and the most developed customer relationships. The investment was driven by a strategic assessment of the firm's position in the value chain, not by a marginal abatement cost calculation. The firms that had not invested cited capital constraints, uncertainty about the payback period, and the absence of customer demand. The cost-based logic pointed toward deferral, and the firms followed it.

5.4 Synthesis: Compliance, Relocation, and the Capability Divide

The three sections of this chapter have examined compliance pressure, relocation dynamics, and the limits of cost-based upgrading as separate phenomena. They are, in fact, three aspects of a single process: the differentiation of the labor-intensive firm population into firms that build green capability and firms that do not. The process is driven by the interaction of regulatory pressure, market demand, and firm-level resources, and it is producing a capability divide that is likely to persist and widen.

The firms on the capable side of the divide share a set of characteristics. They are larger, with revenue above 100 million yuan. They have direct relationships with customers in regulated markets, and those relationships involve joint problem-solving rather than arm's-length transactions. They have management teams that treat environmental requirements as strategic rather than administrative. They are embedded in the cluster's supplier network and can draw on it for collaboration. They have access to finance, either through retained earnings or through bank relationships that are not solely collateral-based. They have made deliberate investments in measurement infrastructure, and they have used that infrastructure to identify and pursue abatement opportunities.

The firms on the other side of the divide share the opposite characteristics. They are smaller, with revenue below 50 million yuan. They sell through trading companies or to domestic customers, and their exposure to regulated markets is indirect. Their management teams are focused on cost control and survival. They are marginal members of the cluster's supplier network, with few collaborative relationships. They are financially constrained. They have complied with environmental requirements minimally, often by purchasing certification services rather than building internal capacity.

The divide is not static. Firms can move from one side to the other, and the direction of movement depends on the choices they make and the support they receive. The policy environment matters: firms that have access to subsidized finance, technical assistance, and cluster-level collective goods are more likely to move toward capability. The market environment matters: firms that face customers who reward green performance are more likely to invest. The firm's own strategy matters: firms that frame environmental requirements as opportunities rather than costs are more likely to build capability.

The divide has consequences for the cluster as a whole. The capable firms are the ones that will survive the tightening of carbon regulation and the intensification of relocation pressure. They will absorb the orders that the marginal firms lose, and they will attract the workers and suppliers that the marginal firms release. The cluster will become more concentrated, more productive, and more capable, but it will also become smaller in employment terms. The workers who lose their jobs in the consolidating firms will not automatically find them in the survivors, because the survivors are more automated and more skill-intensive. The social cost of the transition is real, and it is borne disproportionately by the workers and communities that built the cluster.

This is the dilemma that the labor-intensive sectors present to the dual carbon transition. The transition is necessary, and the capability it requires is real. But the transition is also a selection process that rewards firms with resources and penalizes firms without them, and the selection process has distributional consequences that the policy framework has not adequately addressed. The next chapter turns to technology-intensive industries, where the capability dynamics are different—where the returns to innovation are higher, the learning curve is steeper, and the policy support is more developed. The contrast with the labor-intensive sectors will clarify what is specific to the labor-intensive case and what is general to the capability-based view of green competitiveness.

5.5 Implications for the Capability-Based View

The analysis in this chapter qualifies the capability-based view of green competitiveness in three ways, and the qualifications are important for the book's argument.

The capability-based view must specify the level at which capability is built. The firm is the locus of capability, but the firm's capability depends on the cluster's ecosystem—its suppliers, its institutions, its labor pool, its knowledge base. The firm that is isolated cannot build the supply-chain coordination capability that green governance requires, because it has no suppliers to coordinate with. The policy implication is that capability building must be supported at the cluster level, through collective goods and shared infrastructure, not only at the firm level.

The capability-based view must also acknowledge the governance structure of the value chain. The firm's ability to capture the returns on its capability investment depends on its bargaining power vis-à-vis its customers. In captive chains, where the lead firm can switch suppliers at low cost, the supplier's capability does not translate into value capture. The supplier bears the cost of compliance and the lead firm takes the benefit. The furniture case suggests that the governance structure matters: the more localized and less concentrated the chain, the more the supplier can capture the returns on its investment. The policy implication is that capability building must be accompanied by efforts to upgrade the firm's position in the chain, through branding, design, and direct market access.

The third qualification concerns the limits of marginal abatement. The cost-based firm will invest in the cheap abatement options and defer the expensive ones, and the cheap options are quickly exhausted in labor-intensive sectors. The transition to green competitiveness therefore requires discontinuous, transformative investment, which the cost-based firm is structurally biased against. The policy implication is that the transition will not happen through incremental market incentives alone. It requires a combination of regulation that raises the floor, support that lowers the cost of transformative investment, and institutional mechanisms that help firms recognize and act on the strategic opportunity.

These qualifications do not undermine the capability-based view. They specify its conditions. The view holds that green governance capability is an endogenous driver of industrial upgrading, not a passive product of regulatory compliance. The evidence from labor-intensive industries supports this claim: the firms that have built capability are the ones that are climbing, and the firms that have complied minimally are the ones that are stuck. But the evidence also shows that the capability-building process is conditional on factors that are outside the firm's control—the cluster's ecosystem, the chain's governance structure, the policy environment. The capability-based view is therefore a theory of firm strategy that is embedded in a theory of institutional and cluster-level conditions. The next chapters develop the institutional dimension, examining how clusters, associations, and multi-level governance structures enable or constrain the capability-building process.

The Shunde factory where this chapter began offers a final image. Liang, the manager, walked me to the loading bay where a container was being loaded with cabinet doors for a German customer. The barcode on each pallet carried the carbon footprint declaration that had cost him 40,000 yuan a year. He was proud of it, in a defensive way. "We can prove what we emit," he said. "Most of our competitors cannot." I asked him whether the declaration had helped him win business. He paused. "It has helped us keep business," he said. "Winning is a different question."

That distinction—between keeping and winning, between complying and climbing—is the distinction that this chapter has tried to specify. The firms that will win the dual carbon transition are not the firms that can prove what they emit. They are the firms that have changed what they emit, and changed it in a way that their customers value and their competitors cannot easily match. Liang's factory had not made that change. It had installed the meters, hired the consultant, and produced the declaration. It had not yet reformulated its adhesives, replaced its gas-fired drying ovens, or renegotiated its panel specifications for carbon content. When I asked Liang what would make him take those steps, he named a customer commitment to buy the lower-carbon product at a higher price, and he did not have one. The declaration kept the order; it did not create a market for anything better. Until it does, the label on the pallet will remain the most expensive thing in the building, and the factory will remain what Liang said it was: a factory that has learned to fill in the forms.