Modular CBAM: Epistemological Pluralism in Carbon Accounting

Executive Summary

The European Union's Carbon Border Adjustment Mechanism confronts a dilemma: carbon accounting standards diverge substantially across industrial contexts, yet CBAM assumes methodological convergence toward European ISO frameworks as the sole path to rigor. This paper challenges that assumption through empirical analysis of three non-ISO methodologies—China's GB/T 24067 system boundary approach, India's Perform-Achieve-Trade tiered governance, and Brazil's I-REC distributed verification. These frameworks reflect rational institutional adaptations rather than technical deficiencies. Drawing on Science and Technology Studies and systems theory, we show how allocation method choices embody ontological commitments about industrial organization, generating 15-30% carbon footprint variance that cannot be resolved through "better measurement" but requires translation across distinct epistemic frameworks. CBAM's current design commits epistemic violence by dismissing contextually valid practices. We propose modular architecture organized around interface standardization, operationalized through a Carbon Accounting Methodological Working Group with nested governance structures. Critical safeguards include conservative default penalties for lower-precision modules, preventing methodological arbitrage while preserving diversity. This framework maintains the 1.5°C carbon budget constraint while recognizing multiple epistemic pathways to environmental effectiveness.

Keywords: Carbon accounting, CBAM, epistemological pluralism, ontological politics, modular governance, methodological diversity

1. Introduction

Carbon accounting has a comparability problem. The same steel product can receive ratings differing by 15-30% depending on allocation methodology. Manufacturing facilities face compliance costs varying by an order of magnitude across verification systems. Energy efficiency gains documented through one framework remain invisible to trade mechanisms designed around another.

These are not technical inconsistencies. They reflect deeper questions about whose knowledge counts in climate governance—and whether global coordination requires epistemological monoculture or can accommodate contextual rationality. CBAM, the EU's Carbon Border Adjustment Mechanism, currently assumes the former. This paper argues for the latter.

CBAM emerged from defensible premises: without addressing embedded carbon in traded goods, domestic climate policies risk leakage as production shifts to jurisdictions with weaker regulations. The mechanism applies carbon pricing to imports of emission-intensive products—steel, cement, aluminum, fertilizers—requiring importers to purchase certificates corresponding to emissions that would have been charged under EU ETS. During the transitional phase (October 2023-December 2025), reporting obligations revealed methodological friction. Data from early 2024 showed approximately 70-85% of declarations relied on default emission factors rather than installation-specific measurements, with electricity-related emissions approaching 95-100% default usage in the first quarter. For countries like India, where steel and aluminum exports to the EU totaled approximately $8-10 billion in 2022 (UN Comtrade data), the implications of methodological recognition extend beyond compliance costs to development pathways.

Policy responses have largely framed this as implementation challenge amenable to technical harmonization. The implicit assumption: with sufficient capacity building, infrastructure investment, and regulatory alignment, all producing nations can converge toward measurement and verification standards embedded in European frameworks such as ISO 14067. Yet what if observed methodological divergences reflect not temporary developmental gaps but rational responses to genuinely different industrial ecologies, governance structures, and geopolitical positions?

Three patterns suggest this alternative interpretation. China's GB/T 24067 standard mandates physical allocation for industrial byproducts and permits system-boundary expansion capturing industrial symbiosis benefits—generating calculations that differ 15-30% from ISO 14067 system expansion approaches. This represents an ontological commitment to viewing steel production as embedded within integrated ecosystems where 60-80% of energy comes from captive plants and waste streams systematically become feedstocks for adjacent production. India's PAT mechanism implements three-tier governance (mandatory monitoring above 30,000 TOE annual consumption, simplified reporting for 5,000-30,000 TOE, voluntary cluster programs below 5,000 TOE) reducing total compliance costs from an estimated €216 million under uniform high-standard verification to €36 million—an 83% reduction preserving viability for thousands of small enterprises while delivering 14.08 million tonnes oil equivalent savings in Cycle II alone, exceeding the 8.869 MTOE target by 58%. Brazil's I-REC system leverages 83% hydropower generation (renewables exceeding 85% per EPE 2024 data) to create renewable certificates with industry-estimated fraud rates below 1%, compared to CDM's documented credibility challenges where studies suggested up to 85% of claimed reductions faced questionable additionality.

This paper builds on previous applications of the Theoretical Climate Action Framework to CBAM design challenges. Earlier analyses diagnosed how CBAM's current architecture creates coordination failures (misalignment between European verification procedures and non-EU industrial organization) and legitimacy deficits (epistemic injustice through dismissal of contextually valid methodologies). This paper moves from diagnosis to institutional design, demonstrating how modular architecture with nested governance can address these deficits simultaneously while preserving environmental rigor.

Theoretical Framework, Not Operational Blueprint. This paper offers theoretical framework rather than ready-to-implement specifications. While we develop institutional mechanisms like the Conservative Principle and CAMWG governance structure, we do not prescribe exact calibration parameters (penalty factors, threshold values). These require empirical validation through phased pilots that this conceptual work cannot anticipate. Our contribution lies in demonstrating that modular architecture with nested governance can in principle resolve CBAM's epistemological tensions, providing direction for future policy experimentation rather than operational details. The framework focuses on Scope 1 (direct emissions) and Scope 2 (purchased electricity) where methodological variance, while significant (15-30%), remains manageable. Scope 3 supply chain emissions present more fundamental boundary challenges requiring additional theoretical development beyond this paper's scope.

The empirical work examines three methodologies in depth (Sections 3-5), demonstrating how each addresses coordination challenges through contextually rational designs. Section 6 synthesizes these insights into modular architecture organized around nested governance structures. Section 7 develops implementation pathways with critical safeguards against methodological arbitrage—the risk that enterprises strategically select modules yielding lowest reported intensities. Section 8 addresses broader implications, including how Brazil's "green aluminum" at 2.8 tCO₂e/tonne paradoxically exceeds Chinese steel's 2.1-2.3 tCO₂e/tonne absolute intensity despite superior environmental effort, exposing why cross-sector comparisons require contextual evaluation rather than simple ranking.

The contribution operates at two levels. Practically, it offers CBAM redesign that could reduce trade friction while maintaining environmental integrity. Theoretically, it demonstrates how moving from goal-seeking to learning system paradigms enables global institutions to coordinate across heterogeneous knowledge systems without epistemological imperialism—a challenge extending well beyond carbon accounting to multiple domains of contemporary governance.

2. Theoretical Foundations: Epistemological Pluralism and Learning Systems

2.1 Ontological Politics and the Limits of Standardization

Epistemological pluralism differs fundamentally from relativism. Strong relativism holds that truth claims are wholly context-determined with no basis for comparative evaluation—applied to carbon accounting, this would suggest Chinese, Indian, Brazilian, and European methodologies represent incommensurable frameworks. That is not the position advanced here. Epistemological pluralism, as developed in Helen Longino's contextual empiricism and Science and Technology Studies scholarship, makes a different claim: multiple methodological approaches can achieve comparable rigor within their respective domains while remaining constrained by transparency, internal consistency, empirical testability, and critical community scrutiny. The distinction lies in criteria. Pluralism asserts that validity assessment must attend to the specific problem space a methodology addresses, not that all methodologies are equally valid regardless of context.

When steel mills produce byproduct gases (blast furnace gas, coke oven gas) used for onsite electricity generation, allocating carbon emissions between steel and electricity is not answerable through pure measurement. It requires ontological choice about how to conceptualize the production system. ISO 14067 recommends system expansion—treating byproduct gas electricity as displacing grid power and crediting steel production accordingly. GB/T 24067 tends toward physical allocation based on energy content relationships. Studies of identical facilities under both methods find 15-30% carbon footprint variation, yet this divergence reflects differing but defensible commitments about industrial organization's nature rather than one methodology being incorrect.

Annemarie Mol's ethnographic study of atherosclerosis treatment reveals how what things are gets enacted through practices rather than simply discovered. In patient leaflets, atherosclerosis appears as gradual arterial obstruction amenable to lifestyle intervention. In diagnostic imaging, it presents as quantifiable stenosis percentages requiring threshold-based decisions. In surgery, it becomes problematic tissue requiring physical removal. Mol's key insight: these are not different perspectives on a single disease entity but different enactments brought into being through specific clinical practices and technologies.

Carbon accounting methodologies similarly enact distinct realities. When GB/T mandates physical allocation for co-products from integrated processes, it enacts steel production as industrial metabolism where blast furnace gas is inseparable from primary production. When ISO recommends system expansion, it enacts byproducts as separable commodities whose production could occur through alternative pathways. The case of Caofeidian industrial zone in Hebei Province demonstrates how Chinese steel industry organization validates the first framing. The complex links steel production, seawater desalination, salt manufacturing, and power generation through dedicated infrastructure. Blast furnace slag feeds directly into adjacent cement production. Steel mill waste heat drives desalination equipment producing 50,000-100,000 tonnes daily freshwater. Concentrated brine becomes feedstock for salt production, cutting evaporation cycle time roughly in half. Solid waste utilization reaches 97.5% at Shougang Jingtang operations.

Attempting system expansion allocation encounters practical difficulties: there are no market transactions establishing what grid electricity or alternative salt production would entail, infrastructure prevents byproduct diversion to external uses, and efficiency gains arise from integration rather than discrete process optimization. Physical allocation aligned with system boundaries encompassing the industrial park better captures this integrated metabolism. Conversely, for standalone mills purchasing grid electricity and selling byproduct gases on commodity markets, system expansion more accurately reflects operational independence.

Studies of LCA methodological sensitivity (Finnveden et al. 2009 on allocation choices, Reap et al. 2008 on boundary definitions) document that legitimate methodological differences in system boundaries and allocation procedures can produce 15-30% variance in carbon footprints for identical products. The variance between GB/T and ISO methods falls within this range of structurally driven variation rather than representing measurement error. Both methodologies achieve internal rigor while enacting different, contextually appropriate ways of understanding industrial carbon emissions.

This structural multiplicity poses challenges for governance systems requiring cross-system comparability. CBAM's approach—treating European methodologies as reference standards requiring demonstrated equivalence from alternatives—embeds particular politics about what industrial production is. It positions the atomistic, product-centric understanding underlying ISO standards as universal and relegates alternative understandings to derivative status requiring constant justification. The political dimension becomes visible when considering whose industrial practices align with which ontologies. European manufacturing, characterized by relatively lower vertical integration and market-based coordination, developed methodologies fitting this decoupled structure. Chinese industrial policy deliberately promotes integrated parks and vertical integration within large state-owned enterprises.

Recognizing this as politics about fundamental categories rather than technical standardization opens space for different governance approaches. Rather than seeking convergence on what industrial production fundamentally is, a pluralist approach maintains multiple frameworks while developing translation mechanisms preserving meaningful comparability. The modular architecture proposed in Section 6 operationalizes this through interface standardization—diverse methodologies become modules meeting common transparency, verifiability, and effectiveness requirements while allowing internal diversity in how they conceptualize production systems.

2.2 Learning Systems, Co-Production, and Physical Constraints

Peter Checkland's Soft Systems Methodology distinguishes "goal-seeking systems" from "learning systems." Traditional systems engineering treats organized human endeavors as goal-seeking—entities with pre-defined optimal states reachable through systematic means-ends optimization. This paradigm succeeds in technical domains like aircraft design or supply chain logistics with clear objectives and well-bounded solution spaces. Checkland observed that situations involving multiple stakeholders with competing values resist this framework. In organizational change and policy design, there often exists no single "optimal" state all parties would recognize. Different stakeholders perceive situations through different worldviews, prioritize different outcomes, operate under different constraints.

Soft Systems Methodology proposes reconceptualizing such situations as learning systems where primary value lies in structured inquiry processes rather than predetermined endpoints. Through iterative cycles of rich picture development, root definition formulation, and comparison with reality, participants develop shared understanding and identify accommodation areas even when disagreements persist. The process generates adaptation capacity rather than singular solutions.

CBAM's current architecture embodies goal-seeking logic. European ISO standards are positioned as optimal endpoints toward which all systems should converge. The mechanism's graduated free allowance reduction—97.5% in 2026, declining to zero by 2034—functions as optimization pressure. The approach assumes that with sufficient time and incentive, enterprises worldwide can implement measurement systems structurally identical to those developed for European contexts. Observed deviations are interpreted as temporary deficits awaiting correction.

Reconceptualizing carbon accounting as a learning system does not abandon physical constraints. The 1.5°C carbon budget remains non-negotiable. What the learning paradigm rejects is assuming a single methodological pathway is the only route to this target. Different modules can achieve equivalent environmental rigor through different epistemic approaches—the test is demonstrated emission reductions, not procedural conformity to European norms. The modular architecture creates multiple feedback loops: if one module's enterprises show stagnant carbon intensity while others improve, this triggers investigation. Diversity itself becomes diagnostic, revealing where methods succeed or fail in driving actual decarbonization.

Sheila Jasanoff's co-production framework provides complementary insight. Co-production argues that scientific knowledge and social order mutually constitute each other—knowledge both embeds in and is embedded by social identities, institutions, and discourses. GB/T 24067, PAT, and I-REC did not simply emerge as technical solutions but co-evolved with the industrial structures and political economies they serve. GB/T's emphasis on system-level boundaries reflects and reinforces China's industrial policy around integrated development zones. Provincial authorities use GB/T-compliant carbon intensity targets to evaluate park performance, incentivizing deeper integration and closed material loops—creating empirical conditions that further validate GB/T's commitments about system boundaries.

India's PAT mechanism co-produces specific state-industry relations. The three-tier structure emerged from India's federal governance architecture where central agencies like the Bureau of Energy Efficiency have limited enforcement capacity and must work through state institutions and industry associations. The tiered approach made this constraint productive by creating differentiated governance relationships appropriate to enterprise scales. Large enterprises gain technical capacity and international credibility through rigorous monitoring, medium enterprises avoid prohibitive costs while contributing to national targets, and small enterprises access technical assistance through clusters without existential compliance burdens.

Brazil's I-REC system emerged from conditions where overwhelming reliance on hydropower (83% hydro, 85%+ total renewables per 2024 EPE data) created opportunities for renewable certification to serve climate governance functions, yet geopolitical positioning created vulnerabilities for centralized schemes where single registries could become pressure points. The distributed architecture with blockchain-compatible protocols offers redundancy against both technical failures and political interference.

Understanding these systems as co-produced rather than merely developed has implications for CBAM design. Efforts to achieve convergence through capacity building alone will likely prove insufficient, as divergent methodologies reflect not just knowledge gaps but different configurations of industry-state-society relations. What becomes possible is translation rather than convergence—creating interfaces where diverse knowledge systems can exchange information and coordinate action without requiring homogenization.

3. Systems Thinking in Practice: China's GB/T 24067 and Integrated Industrial Ecologies

3.1 Standard Evolution and Technical Architecture

China's national standard GB/T 24067 for product carbon footprinting has evolved through several major revisions since initial 2009 promulgation. The most recent version represents not convergence toward international frameworks but deeper integration with China's industrial organization characteristics and climate policy infrastructure. The standard's functional positioning as analogous to ISO 14067:2018 is evident in overall methodological architecture drawing from ISO's life cycle assessment foundations. Both require systematic greenhouse gas inventory across product life cycles, mandate transparency in documenting data sources and allocation choices, emphasize consistency enabling meaningful temporal and cross-product comparisons.

Closer examination reveals substantive technical differences accumulating into distinct philosophies. Where ISO 14067 provides flexibility for practitioners to choose allocation approaches based on circumstances, GB/T 24067 introduces more directive guidance steering users toward physical allocation methods over economic allocation for co-product systems. The standard specifies explicit quantitative thresholds for cut-off rules—materials contributing less than 1% of total mass may be excluded provided cumulative impact of all excluded inputs remains below 5% of total carbon footprint. These hard numerical thresholds reduce practitioner discretion compared to ISO's more flexible "significance" criterion, trading context-sensitivity for greater consistency and auditability.

The most substantial divergence appears in provisions for system boundary expansion incorporating industrial symbiosis relationships. GB/T 24067 explicitly requires practitioners to consider and document geographical boundaries, acknowledging that China's regional variations in grid emission factors, industrial clustering patterns, and resource endowments create situations where facility-level accounting divorced from regional context may systematically misrepresent actual environmental impacts. For enterprises operating within designated industrial parks, the standard permits and sometimes requires expanding boundaries to capture material and energy flows among proximate facilities that are institutionally coordinated even if legally separate. This reflects systems thinking where emergent properties of industrial ecosystems—efficiency gains from waste heat cascading, byproduct exchange, shared infrastructure—cannot be captured through summation of individual facility assessments.

3.2 Allocation Methods and Structural Commitments

Methodological divergence around allocation produces quantitatively significant impacts. Large integrated steel facilities operating blast furnace-basic oxygen furnace routes generate multiple energy-carrying byproducts: blast furnace gas from iron ore reduction at 800-950 kcal/Nm³, coke oven gas from coking operations at 4,200-4,500 kcal/Nm³, and converter gas from oxygen blowing. These gases are routinely used for onsite power generation through dedicated gas-fired generators or combustion in boilers feeding steam turbines. A modern integrated mill might generate 60-80% of electricity requirements from captive sources.

Under ISO 14067's recommended system expansion approach, electricity generation from byproduct gases receives credit for displacing grid electricity that would otherwise be consumed. If the regional grid has emission factor 0.58 kgCO₂/kWh (roughly China's national average), each kWh from byproduct gas reduces steel product allocated emissions by 0.58 kg. The underlying logic treats gas as separable commodity whose alternative fate—if not used onsite—would be combustion for grid electricity. For a facility producing 5 million tonnes crude steel annually and generating 2 billion kWh from byproduct gases, this credit would reduce calculated steel carbon intensity by approximately 0.23 tCO₂/t steel.

GB/T 24067's preference for physical allocation takes a different path. The physical relationship between steel production and byproduct gas generation—roughly 1,500-1,700 Nm³ BFG per tonne hot metal, 50-55 Nm³ COG per tonne coke, 50-70 Nm³ converter gas per tonne crude steel—suggests treating them as joint products with emissions allocated based on energy content, mass, or another physical metric. If BFG represents 15% of total energy content output from blast furnace operation (pig iron constituting remaining 85%), then 15% of blast furnace emissions allocate to gas and 85% to pig iron. Because BFG has lower heating value than coal inputs consumed, this allocation assigns proportionally less carbon to the gas stream than system expansion crediting would provide, resulting in higher calculated steel product intensity.

Empirical studies comparing methods for identical facilities find carbon footprint variations of 15-30% depending on allocation choice, with system expansion consistently producing lower steel product intensities. This range aligns with broader research on life cycle assessment boundary sensitivity. The allocation method choice embeds commitments about industrial production's structural nature. System expansion enacts steel mills as modular processors where inputs and outputs are separable, substitutable, amenable to market-based coordination. Physical allocation enacts them as integrated metabolisms where byproducts are inherent features of primary production rather than independent products. Neither enactment is universally appropriate; each fits different empirical realities.

3.3 Industrial Ecology Validation

At Caofeidian industrial zone in Hebei Province, blast furnace slag from steel production feeds directly into adjacent cement clinker production, displacing virgin limestone. Steel mill waste heat drives desalination equipment producing 50,000-100,000 tonnes daily freshwater. This water serves industrial processes while concentrated brine becomes feedstock for salt production at neighboring Nanpu salt fields, cutting solar evaporation cycle time roughly in half and eliminating fresh seawater intake needs at salt operations.

Quantifying this system reveals why integrated accounting matters. The desalination operation's specific energy consumption stands at 3.5-4.0 kWh per tonne for membrane processes, but this excludes thermal energy normally required absent waste heat integration. Independent seawater desalination typically consumes 10-15 kWh/tonne accounting for both electrical and thermal inputs. Waste heat integration at Caofeidian effectively provides 6-11 kWh/tonne free energy that would otherwise require fossil fuel combustion. Across 18-26 million tonnes annual capacity, this represents 108-286 GWh avoided energy consumption and corresponding emissions reductions of 63,000-166,000 tonnes CO₂ annually at grid emission factors. Comprehensive utilization rates for industrial solid waste reach 97.5% at Shougang Jingtang operations—less than 2.5% of material waste streams require off-site disposal, vastly exceeding typical industry performance where 20-30% solid waste generation is common.

Lubei National Eco-Industrial Demonstration Park in Shandong Province demonstrates broader patterns. Lubei's complex centers on phosphate fertilizer production but achieves elemental-level recycling through integrated chains. The phosphorus-sulfur-calcium circuit addresses phosphogypsum waste plaguing conventional phosphate manufacturing. Traditional phosphoric acid production via wet process generates roughly 5 tonnes phosphogypsum waste per tonne phosphoric acid product. Lubei thermally decomposes phosphogypsum to recover sulfur dioxide (returning to sulfuric acid production for reuse in phosphate processing) and calcium oxide (substituting for limestone in cement production). Ecological network analysis shows sulfur recycling efficiency exceeds 95%. Energy integration parallels material integration: waste heat from power generation drives desalination and concentration processes; chlorine gas from chlor-alkali operations supplies titanium dioxide production; sulfuric acid waste streams from titanium processing assist phosphate ore dissolution. System-level energy efficiency improvements relative to standalone operations reach approximately 86%.

These cases pose acute challenges for methodologies predicated on facility-level boundaries and product-centric allocation. Where should system boundaries lie—at legal entity, industrial park perimeter, or extended to downstream users of recycled materials? How should emissions associated with waste heat be allocated between primary processes generating it and secondary processes beneficially reusing it? GB/T 24067's provisions accommodate these questions. ISO 14067 could theoretically handle them through careful system expansion application, but complexity and data requirements grow rapidly, and absence of market prices for many intermediate flows makes economic allocation impractical.

3.4 Energy System Complexity

China's industrial sector energy consumption creates additional complications. Major energy-intensive industries—steel, aluminum, cement, chemicals—rely heavily on captive power generation rather than grid purchases. Roughly 60-80% of electricity consumed by leading steel companies comes from self-generated sources, primarily through recovery of industrial byproduct gases and waste heat. These captive plants predominantly burn coal or coal-derived gases, yielding emission factors often exceeding grid average values, especially in regions where grids have seen rapid renewable penetration.

For carbon accounting, this creates divergence between approaches focused on direct facility-level measurements and those relying on grid average factors. An enterprise purchasing grid electricity in Sichuan Province (where hydropower dominates and factors may fall below 0.1 kgCO₂/kWh during wet seasons) would calculate very low indirect emissions. An enterprise in Shanxi Province with captive coal-fired generation might consume electricity with emission factors exceeding 1.0 kgCO₂/kWh. GB/T 24067's requirement to specify geographical boundaries and its detailed provisions for accounting treatment of self-generated electricity reflect these realities.

This added complexity serves accuracy in Chinese contexts but creates friction when interfacing with frameworks like CBAM that default to grid factors for electricity-related emissions unless detailed facility-specific data is provided. Exporting enterprises would need to document not only total electricity consumption but also percentages from captive sources, fuel mix and efficiency of captive generation equipment, and emission factors for purchased grid power—increasing verification requirements. Yet failing to make these distinctions could systematically overstate or understate actual emissions depending on regional grid characteristics and captive generation prevalence.

4. Bounded Rationality and Tiered Governance: India's PAT Mechanism

4.1 Policy Architecture and Differentiated Obligations

India's Perform, Achieve and Trade mechanism, operating since 2012 under the National Mission for Enhanced Energy Efficiency, provides contrasting institutional model to both China's integrated industrial park approach and European facility-level comprehensive monitoring. The mechanism targets energy-intensive industries through market-based efficiency improvement, with its distinctive feature lying in how it differentiates obligations across enterprise scales through three-tier structure acknowledging and working within governance capacity constraints.

The legal foundation rests on Energy Conservation Act 2001, granting the Bureau of Energy Efficiency authority to designate "Designated Consumers" subject to mandatory energy efficiency requirements. Within the steel sector—the focus given CBAM exposure—tier boundaries are drawn at 30,000 tonnes of oil equivalent annual energy consumption for mandatory participation and 5,000 TOE for simplified frameworks. Enterprises exceeding 30,000 TOE annually must participate in PAT cycles, typically spanning three years. These large facilities face requirements to achieve facility-specific energy intensity reduction targets denominated in specific energy consumption (energy input per unit product output). Facilities exceeding targets earn Energy Saving Certificates tradeable to underperforming facilities, creating economic incentives for over-compliance.

Medium enterprises consuming 5,000-30,000 TOE annually occupy middle ground. During initial cycles, these remained largely outside mandatory frameworks. More recent policy documents explore incorporating them through simplified reporting or voluntary programs with government technical assistance. The rationale recognizes these mid-sized operations possess neither capital base nor management capacity of large enterprises, making full PAT compliance requirements economically burdensome relative to potential efficiency gains.

Small and micro enterprises below 5,000 TOE—numbering in thousands across India's dispersed steel and metal fabrication sectors—remain exempt from mandatory reporting. They access voluntary technical assistance through BEE-SME initiative and sector-specific cluster programs providing group training, shared auditing services, and access to subsidized energy-efficient equipment procurement. The underlying logic treats very small enterprises as collectively significant but individually unable to bear compliance costs.

4.2 Effectiveness and Cost Economics

Official reporting from BEE documents substantial energy savings across completed cycles. Cycle I (2012-2015) targeted 6.686 million tonnes oil equivalent annual savings and achieved 8.67 MTOE, a 30% over-achievement. Cycle II (2016-2019) increased ambition with 8.869 MTOE target and delivered 14.08 MTOE, representing 58% over-achievement. These figures suggest target-setting has been conservative relative to actual abatement potential, but they confirm the mechanism drives measurable efficiency improvements among covered enterprises. Cycle II's success translated into approximately 66 million tonnes CO₂ emissions avoided over the three-year period.

Compliance rates across participating facilities have remained consistently high, exceeding 90% in most cycles. Academic analyses note many facilities achieved targets through incremental operational improvements and energy management practices rather than major capital investments in advanced technology. This suggests PAT captures accessible efficiency opportunities rather than driving frontier technological transformation. However, for a mechanism operating across sectors characterized by substantial technical inefficiency relative to global best practice, capturing accessible gains may represent appropriate prioritization.

The cost economics of tiered governance become apparent through comparative analysis. If India imposed uniform high-standard measurement, reporting, and verification requirements comparable to those envisioned under CBAM across all 2,700 steel and ferroalloy enterprises (including those currently outside PAT due to small scale), estimated compliance costs would reach approximately €216 million annually. This calculation assumes average costs of €80,000 per enterprise per year for comprehensive continuous emissions monitoring systems, accredited third-party verification, and detailed reporting aligned with EU ETS standards—a figure derived from European verifier fee structures and monitoring equipment amortization. For large enterprises (Tier 1), such costs represent a small fraction of annual revenues and operating budgets—likely below 0.1% of turnover for billion-dollar facilities.

The economics shift dramatically for smaller scales. Medium enterprises in the 5,000-30,000 TOE range typically show annual revenues of €10-100 million with operating margins of 3-8% depending on market conditions and product mix. MRV costs of €80,000 annually represent 0.1-0.8% of revenues but potentially 2-10% of operating profits. During cyclical downturns when margins compress, these compliance costs could turn marginally profitable operations into loss-makers. Small and micro enterprises below 5,000 TOE face starker choices. With annual revenues often below €10 million and operating profits in the €500,000-€2 million range for successful operations, €80,000 annual compliance costs would consume 4-16% of profits.

India's actual PAT architecture avoids this cost distribution through differentiated requirements. Tier 1 enterprises face rigorous but affordable obligations (estimated €50,000-€100,000 annually for comprehensive energy audits, SEC measurement, reporting, and verification based on BEE cost studies). Tier 2 enterprises under simplified frameworks incur reduced costs (€15,000-€25,000 annually). Tier 3 enterprises outside mandatory coverage participate voluntarily through cluster programs with government-subsidized technical assistance, incurring minimal direct costs. Aggregate annual MRV costs across the entire sector approximate €36 million, representing 83% cost reduction relative to uniform high-standard application while still covering facilities responsible for roughly 70-75% of sectoral energy consumption and emissions.

4.3 Bounded Rationality and Legal Legitimacy

The three-tier structure finds theoretical justification in Herbert Simon's concept of bounded rationality and Lon Fuller's legal philosophy regarding preconditions for legitimate law. Simon's insight—that decision-makers facing complexity and resource constraints "satisfice" rather than optimize—applies to governance design as readily as to individual choice. A perfectly optimal carbon accounting system might mandate comprehensive real-time monitoring of all emission sources with redundant verification and complete supply chain transparency. But implementing such a system requires technical capacity, financial resources, administrative infrastructure, and verification personnel that exist in limited supply. A governance approach demanding the optimal while ignoring capacity constraints will fail not because regulated entities are unwilling but because compliance is literally impossible.

Fuller's analysis of the "morality of law" articulates eight principles valid legal rules must satisfy to generate legitimate obligation. Most relevant here is the principle that law must not demand the impossible—rules requiring actions beyond practical capacity of those subject to them cannot function as law in any meaningful sense. They become expressions of aspiration or exercises of coercive power rather than legitimate normative frameworks. Applied to carbon accounting requirements, uniform MRV standards demanding €80,000 annual compliance costs from enterprises with €500,000 annual profits violate this principle. The requirement is formally universal but practically selective, creating de facto barriers to market participation for capacity-constrained actors while appearing neutral.

The tiered PAT structure institutionalizes satisficing and respects the impossibility principle. Rather than demanding optimal information from all enterprises regardless of scale, it establishes satisficing thresholds: enterprises large enough to generate reliable SEC data cost-effectively must do so; enterprises for whom rigorous measurement imposes substantial burden but remains feasible with assistance face simplified requirements; enterprises for whom even simplified requirements would be prohibitive remain outside mandatory frameworks but access voluntary support. This approach sacrifices information completeness—Tier 2 and 3 data will be less granular and verified to lower standards than Tier 1—but achieves system-wide participation generating aggregate environmental benefits while preserving industrial diversity.

4.4 CBAM Exposure and Export Requirements

India's steel and aluminum exports to the EU totaled approximately $8-10 billion in 2022 (UN Comtrade data), with iron and steel products representing roughly 70% and aluminum contributing 29%. CBAM coverage creates acute tensions with India's tiered governance model. The mechanism's transitional phase reporting revealed that approximately 70-85% of importers relied on default emission factors, rising to 95-100% for electricity-related emissions in early 2024. For Indian exporters, this reliance on defaults likely yields unfavorable carbon intensity estimates, as default factors incorporate mark-ups intended to incentivize actual measurement.

Moving beyond defaults requires documentation aligned with EU verification standards. Indian Tier 1 facilities participating in PAT have developed internal energy and emissions accounting systems potentially adaptable to CBAM requirements with incremental investment. These large exporters—likely fewer than 20 enterprises given concentration in export-oriented segments—possess both technical capacity and financial resources to implement necessary upgrades. However, their current PAT compliance data may not map cleanly to CBAM formats, requiring parallel reporting systems.

Tier 2 facilities face more severe challenges. Many participate in export supply chains as intermediate producers or niche product manufacturers, but their simplified PAT reporting likely falls short of CBAM verification standards. Bringing these enterprises (estimated 50-100 companies with meaningful export exposure) into compliance could require investments of €20,000-€40,000 per facility to upgrade monitoring and engage accredited verifiers.

While Tier 3 enterprises remain exempt from mandatory reporting for domestic purposes, those engaging in exports to CBAM-covered markets must upgrade to Tier 2 simplified reporting standards as condition of export participation. This requirement preserves domestic flexibility for purely local producers while ensuring export streams meet minimum verification thresholds. The upgrade obligation applies only to export-oriented production volumes, allowing enterprises to maintain dual-track reporting: simplified verification for exports, cluster-based voluntary programs for domestic sales. This approach balances development policy objectives (preserving small enterprise viability domestically) with trade obligations (preventing CBAM exposure from becoming backdoor exclusion mechanism).

From an epistemic justice perspective, CBAM's uniform verification requirements commit what Miranda Fricker terms "testimonial injustice"—systematically dismissing credibility of testimony (here, carbon accounting data) from particular sources not because information is unreliable but because it fails to conform to dominant epistemic frameworks. India's PAT mechanism generates real energy savings and emissions reductions verified through government oversight, but this verification operates differently from EU third-party accreditation models. BEE certification carries less weight in CBAM frameworks not because it is less rigorous in its own context but because it represents an alternative epistemological approach—state-led verification calibrated to domestic capacity constraints rather than market-based verification assuming deep capital markets and professional services infrastructure.

The 83% cost differential between tiered Indian governance and uniform CBAM-style requirements represents not inefficiency but different tradeoffs. India's approach optimizes for coverage and participation across heterogeneous capacity levels while accepting some reduction in data precision. CBAM optimizes for precision and third-party verifiability while accepting that high compliance costs exclude smaller actors. Neither optimization is inherently superior; each reflects different constraints and priorities.

5. Distributed Trust and Geopolitical Resilience: Brazil's I-REC System

5.1 Renewable Energy Context

Brazil's energy sector contrasts sharply with fossil fuel-intensive industrial systems examined in China and India. Hydroelectric generation contributes 83% of total electricity production according to 2024 EPE (Empresa de Pesquisa Energética) data, with total renewable energy exceeding 85% when including wind and solar power. This renewable abundance creates unique carbon accounting context where electricity consumption by industrial facilities carries minimal embedded emissions compared to global averages. For energy-intensive industries like aluminum smelting—where electricity represents 40-50% of production costs and 70-80% of carbon footprint—access to low-carbon power shapes competitive positioning.

The International Renewable Energy Certificate (I-REC) Standard provides institutional infrastructure through which Brazilian renewable energy generators can document and monetize environmental attributes of generation separately from physical electricity sales. Under this framework, each megawatt-hour of certified renewable generation creates one I-REC certificate carrying metadata about generation source, facility characteristics, and timing.

While I-REC operates globally across multiple countries, Brazil has emerged as one of the largest and fastest-growing markets. First-half 2025 issuance exceeded 54 million certificates according to I-Track Foundation reports, surpassing entire 2024 annual issuance and suggesting market doubling within 12-18 months. For aluminum production specifically, Brazil hosts major operations producing approximately 1.2-1.5 million tonnes annually using predominantly hydropower, yielding production carbon intensities around 2.8 tCO₂e per tonne—substantially below coal-powered smelters' 12-16 tCO₂e/tonne.

5.2 Architectural Differences and Trust Mechanisms

The I-REC Standard's technical architecture differs notably from earlier renewable energy certification schemes in ways reducing vulnerability to systematic fraud and geopolitical pressure. Traditional certification systems, exemplified by the Clean Development Mechanism under Kyoto Protocol, operated through centralized authority structures. CDM project developers seeking carbon credits (Certified Emission Reductions) submitted documentation to limited number of Designated Operational Entities—private sector audit firms accredited by CDM Executive Board. These DOEs conducted validation and verification, with final issuance authority resting with UNFCCC Secretariat.

This architecture created single points of failure and potential capture. Several documented cases revealed DOEs approving projects with inflated baseline scenarios, manipulated measurements, or questionable additionality claims, motivated by client retention pressures and limited oversight capacity. Statistical analysis applying Benford's Law to CDM credit issuance patterns suggested that up to 85% of claimed reductions might be inflated (Cames et al. 2016), though this figure remains contested.

I-REC's architectural response involves distributed verification combined with digital registry infrastructure. Rather than channeling all certification decisions through handful of designated entities, I-REC accredits multiple local registry operators in participating countries. In Brazil, I-REC Brazil registry operates under governance independent of global I-REC Standard Foundation while adhering to standard's technical requirements. Generation data flows from electricity meters at certified facilities to national registry through automated reporting systems, reducing manual data entry vulnerabilities. Certificate issuance occurs through digital platforms with immutable transaction logging, creating audit trails facilitating ex-post verification and fraud detection.

While some centralization persists, operational verification distributes across multiple nodes, reducing single-point-of-failure risk. Industry estimates suggest actual fraud rates below 1%, though this remains empirical speculation rather than rigorously verified fact. Architectural features supporting relatively high integrity include automated meter-to-registry data transmission reducing falsification opportunities, digital certificate tracking preventing double-counting, and distributed registry operations making systematic collusion more difficult than in centralized models.

5.3 Geopolitical Factors

Brazil's positioning in global politics provides often-overlooked motivation for distributed verification infrastructure. As non-aligned middle power, Brazil maintains complex relationships with both traditional Western powers and rising non-Western economies. This positioning creates vulnerabilities for climate-related export sectors dependent on certification by institutions potentially subject to geopolitical pressure.

Distributed registry architecture with blockchain-compatible protocols addresses this concern through redundancy and permissionlessness. If single registry faced political pressure to alter records or reject legitimate certificates, distributed ledger structure would preserve evidence of original issuance across multiple nodes. Certificate holders could invoke dispute resolution mechanisms or shift to alternative registries while maintaining proof of environmental claims. The technical architecture thus serves political function: reducing leverage any single authority can exercise over Brazilian commodity exports competing on environmental credentials.

Elinor Ostrom's design principles for robust commons governance provide theoretical grounding for this architectural choice. Ostrom's eighth principle—nested enterprises for large-scale systems—emphasizes that successful resource governance often requires multiple layers of organization. I-REC's architecture operationalizes this: local facility-level metering and verification, national registry operations, regional coordination, and global standard-setting, with each layer performing functions suited to its scale while maintaining some autonomy from other layers.

6. Synthesis: A Modular Architecture for Epistemologically Plural Carbon Governance

6.1 Core Design Principles

The preceding empirical work documented three methodological frameworks—GB/T 24067, PAT, and I-REC—that diverge from European carbon accounting standards in substantive ways while maintaining internal rigor and delivering environmental benefits. The governance challenge is synthesizing these observations into architecture preserving methodological diversity without sacrificing environmental integrity or meaningful comparability. The solution builds on software engineering and systems architecture concepts: modularity, interface standardization, and protocol compatibility.

Modular architecture in engineering contexts refers to systems composed of discrete, interchangeable components (modules) interacting through standardized interfaces. Each module can differ internally—using different algorithms, data structures, or implementation approaches—while adhering to common interface specifications enabling reliable interaction with other modules. This design philosophy offers advantages: modules can be developed and updated independently without breaking system-wide compatibility, failure of one module doesn't compromise others, users can select modules appropriate to their specific contexts while maintaining system participation, and innovation can occur within modules without requiring global coordination.

Applied to carbon accounting, modular CBAM would recognize multiple methodological frameworks as legitimate "modules" provided they meet common interface requirements. Interface requirements specify what information each methodology must provide and to what standards of transparency and verifiability, but do not prescribe how that information is generated internally. Interface requirements include transparent documentation of system boundaries and allocation choices, third-party verification or government oversight meeting specified independence criteria, quantitative uncertainty assessment for reported values, data retention and audit trail provisions enabling retrospective checking, and demonstrated environmental effectiveness through either absolute emission reductions or intensity improvements over time.

GB/T 24067 constitutes one module, recognized for use by Chinese enterprises whose industrial organization involves integrated production complexes and significant captive power generation. PAT constitutes second module, recognized for Indian enterprises, with internal tiering structure preserved. I-REC constitutes third module, providing renewable electricity documentation for enterprises globally but particularly relevant for Brazilian and other Latin American exporters. ISO 14067/14064 frameworks remain fourth module, maintained as option for any enterprise finding it suitable.

Figure 1: Modular CBAM Architecture.

This modular approach shifts governance challenge from "how do we get everyone to use same methodology?" to "how do we ensure interoperability among diverse methodologies?" The second question is more tractable because it demands convergence only at interface points rather than in internal operations. Precedents exist in other domains. Internet protocols enable radically heterogeneous computer systems to communicate reliably. Electrical power systems interconnect generators using different technologies through standardized frequency and voltage protocols.

This modular architecture raises immediate questions about quality assurance. How do we prevent methodological arbitrage where enterprises game the system by selecting modules yielding lowest reported intensities? Section 7.3 develops safeguards—conservative principle applying precision-adjusted penalties, module lock-in provisions preventing opportunistic switching, and comparative benchmarking detecting systematic under-reporting. These mechanisms ensure that methodological diversity serves contextual fit rather than enabling race to bottom.

Notes: Cost in Euros except I-REC (USD). Precision represents 95% confidence interval (σ).

6.2 Nested Governance Structure

Maintaining modular system requires ongoing coordination to update interface standards, resolve disputes about module compliance with requirements, and incorporate new methodological innovations as they emerge. Ostrom's eighth design principle—nested enterprises for large-scale systems—provides theoretical guidance. Effective governance of complex commons often requires multi-layered institutional structures where different functions occur at appropriate scales.

The proposed Carbon Accounting Methodological Working Group (CAMWG) operationalizes this nested structure. At national level, participating countries maintain sovereign authority over their methodological standards. China's GB/T Technical Committee, India's Bureau of Energy Efficiency, Brazil's I-REC Registry, and EU member state technical committees continue developing and refining respective frameworks based on domestic industrial structures and policy priorities. This preserves methodological sovereignty and enables responsive adaptation to changing conditions without requiring international consensus.

At regional level, methodological forums coordinate among countries with similar industrial structures or shared regional interests. An Asia-Pacific Carbon Accounting Forum might include China, India, Japan, South Korea, and Southeast Asian nations, providing venue for discussing harmonization opportunities (where convergence makes sense) and interface specification needs (where diversity will persist). A Latin American Methodological Coordination body might develop shared approaches for renewable energy certification and sector-specific guidelines relevant to region's industrial profile.

At global level, CAMWG Secretariat functions include maintaining registry of recognized methodological modules, facilitating technical dialogue among national frameworks, operating dispute resolution mechanisms when questions arise about module compliance with interface standards, commissioning independent assessments of environmental effectiveness, and coordinating with trade bodies (particularly WTO) on compliance with international trade law.

CAMWG operates through consensus-based decision-making on core interface standards, with regional veto power for standards affecting more than 30% of a region's exports. This ensures no unilateral changes can be imposed by any single bloc. The Secretariat rotates chairship among EU, Asia-Pacific, Latin America, and Africa on 18-month cycles. Technical committees work through qualified majority voting (requiring 60% approval) for operational decisions like module recognition applications, but consensus remains required for fundamental interface standard modifications. Independent technical secretariat housed in neutral location (Geneva) rather than European institutions reinforces autonomy.

Funding CAMWG operations could draw from multiple sources to ensure independence. CBAM revenue allocation dedicates 5-10% of collections (roughly €75-150 million annually given estimated revenue projections of €1.5-2.1 billion by 2028-2030) to technical assistance and coordination. Multilateral climate finance mechanisms under UNFCCC contribute additional resources, particularly for capacity building in least developed countries. National government contributions and private sector participation fees diversify funding base further.

6.3 WTO Compatibility and Equal Treatment

Trade policy integration is essential given CBAM's character as trade measure with climate objectives. WTO compatibility requires avoiding arbitrary discrimination among trading partners or disguised protectionism. Modular CBAM strengthens WTO defensibility by demonstrating that CBAM does not privilege European methodologies purely because they are European, but rather provides framework where any methodology meeting transparent interface requirements gains recognition.

The critical test under GATT Article III (National Treatment) and Article I (Most-Favored-Nation) is whether the measure discriminates based on product origin or production method. Current CBAM design faces vulnerability because it de facto requires conformity with European verification procedures, potentially constituting discrimination against "like products." Modular architecture addresses this by establishing that CBAM does not mandate European verification procedures but rather accepts any methodological module meeting environmental effectiveness and transparency criteria applied even-handedly.

Module recognition criteria are transparent and publicly available, scientifically justified based on environmental effectiveness metrics, and open to applications from any country's methodological framework. Independent technical assessment—not political approval—determines whether a methodology qualifies. Rejected applications can invoke dispute resolution, with clear burden of proof requirements preventing arbitrary exclusion. These procedural safeguards shift legal terrain from "is CBAM discriminatory?" to "are CBAM's interface requirements reasonable and applied even-handedly?"—more defensible if requirements focus on transparency, verification integrity, and environmental effectiveness rather than procedural conformity to European norms.

7. Implementation Pathways and Safeguards

7.1 Phased Roll-Out Strategy

Implementing modular CBAM need not require wholesale replacement of existing structures but rather phased integration and expansion of recognized methodologies. Phase 1 (2026-2027) establishes governance and pilot programs. EU authorities maintain current CBAM implementation while CAMWG Secretariat staff develop detailed interface specifications, recruit technical experts representing diverse methodological traditions, design dispute resolution procedures, and launch pilot programs with volunteer enterprises from China, India, and Brazil testing module recognition processes.

Phase 2 (2027-2028) begins limited module recognition for large enterprises. Chinese and Indian Tier 1 exporters demonstrating robust GB/T or PAT compliance receive provisional CBAM recognition, meaning reported carbon intensities are accepted for tariff calculation without requiring re-verification under ISO methodologies. Brazilian aluminum exporters with I-REC certification plus comprehensive production data similarly receive recognition. European regulators conduct intensive monitoring of recognized modules, comparing claimed carbon intensities with default values and investigating anomalies exceeding 20% variance.

Phase 3 (2028-2029) expands to medium enterprises and additional sectors if Phase 2 demonstrates module reliability. Recognition expands to Tier 2 Indian enterprises using simplified reporting and to additional product categories beyond initial CBAM coverage. Technical assistance programs help medium enterprises upgrade monitoring and reporting to meet interface standards while maintaining methodological diversity.

Phase 4 (2029-2030) achieves full operational status with most exporting enterprises able to participate through preferred methodological module. Default values increasingly function as penalties for non-participation rather than common practice. Periodic reviews assess environmental effectiveness—are recognized modules delivering actual emission reductions comparable to what ISO-based approaches would achieve?

Phase 5 (2030-2032) focuses on continuous improvement and expansion. CAMWG reviews effectiveness data, updates interface requirements based on experience, considers applications for new methodological modules, develops guidelines for emerging issues, and potentially extends modular framework to other environmental accounting domains beyond carbon.

7.2 Resource Requirements

Financial requirements for CAMWG operations depend on scope and ambition. Minimal viable operation might run on €20-30 million annually, covering secretariat staff, technical committees, dispute resolution panels, and basic IT infrastructure. More ambitious operations supporting extensive capacity building in developing countries, funding independent effectiveness assessments, and developing sophisticated digital infrastructure could require €100-150 million annually.

Technical expertise requirements span several domains: carbon accounting methodologists familiar with multiple frameworks, industrial process engineers understanding sector-specific production technologies, verification and audit professionals, trade policy and WTO law specialists, blockchain and digital infrastructure developers, and organizational governance experts experienced with multi-stakeholder coordination bodies.

Capacity building programs focus on three priority areas. First, verifier training and accreditation in developing countries, particularly expanding pools of qualified auditors familiar with both local industrial contexts and international verification protocols. Second, digital infrastructure deployment enabling automated meter-to-registry data transmission, blockchain-compatible certificate issuance, and secure data exchange protocols. Third, enterprise-level monitoring equipment and management systems, particularly for Tier 2 and 3 enterprises requiring subsidized access to continuous monitoring equipment and energy management software.

7.3 Safeguards Against Gaming

Methodological diversity raises legitimate concerns about environmental integrity erosion. Without adequate safeguards, modular CBAM could become race to bottom where countries compete to offer laxest verification standards.

How do we prevent methodological arbitrage—enterprises gaming the system by choosing whichever module reports lowest intensity? The answer lies in precision-adjusted penalties.

Conservative Principle: Lower-precision modules must accept higher default penalty factors proportional to their uncertainty bands. For any module with reported carbon intensity μ and uncertainty σ (95% confidence interval), CBAM tariff calculation applies:

Tariff Basis = μ + k·σ

where k represents precision penalty factor calibrated to uncertainty level. High-precision methods (uncertainty <5%) face minimal penalties; medium-precision methods (5-15% uncertainty) face moderate penalties; low-precision methods (15-20% uncertainty) face substantial penalties. Methods exceeding 20% uncertainty prove unsuitable for module recognition.

The exact penalty schedule requires pilot testing across representative enterprises to balance two objectives: sufficient disincentive against strategic selection of low-precision methods, while avoiding prohibitive costs that exclude legitimate participants. Initial estimates suggest penalty factors in range of 1.0-2.0 for uncertainty bands of 5-20%, but empirical validation through phased implementation is essential. The critical design principle is that precision differential translates into tariff differential, eliminating arbitrage incentive while preserving methodological choice for contextually appropriate reasons.

Module Lock-In: Precision adjustments alone aren't sufficient. Enterprises might still cycle between modules exploiting temporal fluctuations. Module lock-in provisions require three-year commitments, with switching allowed only for major operational changes (facility restructuring, mergers, regulatory shifts). During transition periods, higher of old or new method results applies—eliminating switching arbitrage.

Comparative Benchmarking: Annual cross-module benchmarking provides reality check. If one module's steel producers systematically report intensities diverging significantly from others using different methods for comparable technology, CAMWG investigates. The threshold for triggering investigation—systematic divergence exceeding roughly 25%—derives from two considerations. First, studies of LCA methodological sensitivity (Finnveden et al. 2009, Reap et al. 2008) document that legitimate methodological choices in system boundaries and allocation procedures can produce variance approaching 30% in identical systems. Setting threshold below this maximum allows genuine methodological differences while flagging outliers likely reflecting systematic bias. Second, analysis of existing carbon intensity distributions across production systems shows within-technology-category variance typically remaining below 20% for comparable scales. A threshold capturing variation beyond normal range indicates genuine anomaly requiring investigation rather than legitimate diversity. The burden of proof lies with module to demonstrate divergence has legitimate technical basis.

7.4 Cross-Module Blind Spot Detection

Conservative Principle addresses divergence between modules but cannot detect blind spots shared across modules. If all recognized frameworks systematically undercount certain emission sources—such as upstream supply chain emissions (Scope 3) or fugitive emissions from aging infrastructure—precision-adjusted penalties won't reveal the gap.

External benchmarking against physical constraints provides first safeguard. CAMWG annually compares aggregate reported emissions across all modules against top-down estimates from atmospheric monitoring, material flow analysis, and energy balance calculations. Systematic divergence >15% triggers module-wide investigation into potential shared blind spots. This approach leverages physical reality as ultimate check—the atmosphere doesn't care about methodological choices, only actual molecule flows.

Mandatory periodic scope expansion reviews occur every 5 years. CAMWG commissions independent assessment of whether boundary definitions across all modules adequately capture evolving emission sources. Recent examples include refrigerant leakage from cooling systems and embodied emissions in capital equipment—both historically under-reported across multiple frameworks. These reviews identify emerging sources requiring incorporation into all module interfaces.

Civil society challenge pathway enables environmental NGOs to formally petition CAMWG to investigate potential shared blind spots, providing external pressure to expand accounting boundaries when new emission sources emerge or existing ones prove larger than previously estimated. This maintains accountability to broader climate objectives beyond what industry self-regulation might achieve.

7.5 Scope Limitations

This framework focuses primarily on Scope 1 (direct emissions) and Scope 2 (purchased electricity) where methodological differences, while significant (15-30% variance), remain manageable through conservative penalties and comparative benchmarking. Scope 3 (supply chain emissions) presents more fundamental challenges: boundary definitions, allocation of shared infrastructure emissions, and temporal accounting create variance potentially exceeding 100%. Extending modular architecture to Scope 3 requires additional theoretical development beyond this paper's scope, likely demanding sector-specific protocols rather than universal interfaces. Current CBAM implementation similarly limits coverage to Scope 1+2, making this boundary appropriate for present purposes.

Each recognized module must demonstrate environmental effectiveness on ongoing basis. Interface requirements include minimum effectiveness criteria assessed through three-year review cycles. Modules must show that covered enterprises achieve either absolute emission reductions or carbon intensity improvements exceeding 2% annually on average. Every three years, CAMWG commissions independent technical assessment evaluating whether different modules yield comparable environmental outcomes when applied to similar industrial contexts. Initial recognition lasts 5 years, after which modules must reapply demonstrating continued compliance with interface standards and environmental effectiveness criteria.

8. Discussion: Implications, Limitations, and Broader Horizons

8.1 The Aluminum-Steel Paradox and Sector-Specific Benchmarking

Brazilian aluminum versus Chinese steel illuminates both potential and limitations of modular approaches. Brazilian aluminum production at 2.8 tCO₂e/tonne using 83%+ hydropower represents near-optimal performance given current technology—further decarbonization would require replacing remaining fossil fuel use in auxiliary processes or pursuing speculative technologies like inert anode electrolysis still far from commercial deployment. Chinese steel at 2.1-2.3 tCO₂e/tonne using blast furnace routes powered by coal and captive generation represents mid-range performance with substantial decarbonization potential through electric arc furnace adoption, hydrogen-based direct reduction, and grid decarbonization.

Absolute carbon intensity metrics cannot distinguish "already optimized" from "high improvement potential" contexts. CBAM relying solely on tonnes CO₂ per tonne product comparison would potentially penalize Brazilian aluminum despite superior environmental effort while providing inadequate incentives for Chinese steel transformation. The modular approach enables more nuanced assessment through sector-specific interfaces. Aluminum module specifications incorporate not just absolute intensity but also metrics like percentage renewable electricity, efficiency relative to theoretical minimum energy, and temporal trends showing continuous improvement. Steel module specifications similarly incorporate technology-specific benchmarks—BF-BOF facilities compared against best-available BF-BOF performance, EAF facilities against best EAF, DRI facilities against best DRI.

This sector-specific approach does not abandon physical constraints. The 1.5°C carbon budget imposes absolute limits on cumulative emissions, not intensity targets. Modular CBAM addresses this through dual mechanisms. First, each module must demonstrate temporal improvement—aluminum producers using 90% renewable electricity must still reduce that remaining 10% over time. Second, aggregate monitoring compares total reported emissions across all CBAM-covered imports against top-down estimates from energy statistics and material flows. If sectoral differentiation systematically understates absolute emissions, the divergence becomes visible and triggers interface standard revision.

The aluminum-steel paradox thus reveals not relativist acceptance of any methodology but rather recognition that context-blind absolute intensity rankings can penalize leaders while rewarding laggards. Brazilian aluminum's 2.8 tCO₂/t represents near-theoretical minimum for electrolytic reduction using current technology and 90% clean electricity. Imposing higher tariffs on this than on Chinese coal-fired steel at 2.1 tCO₂/t—which could readily halve emissions through available technologies—inverts climate policy logic. The modular approach preserves the 1.5°C constraint while aligning incentives with actual decarbonization potential.

The limitation is complexity. More differentiated interface requirements demand more sophisticated governance structures to develop, update, and enforce them. Tradeoff exists between precision (recognizing contextual differences) and simplicity (maintaining manageable administrative structures). Modular CBAM accepts higher governance complexity in service of fairness and effectiveness, but this choice is not obviously superior in all circumstances.

8.2 Learning Systems and Continuous Adaptation

Characterizing CBAM redesign through learning systems rather than goal-seeking framework has practical implications for how we evaluate success and adapt over time. Understanding modular CBAM as ongoing inquiry into how diverse industrial systems can coordinate toward climate objectives while preserving contextual rationality, we expect continuous evolution.

The process of maintaining module recognition—periodic reassessment, incorporation of new methodologies, refinement of interface standards—becomes valuable itself rather than mere administrative burden. Through this process, participants surface emerging best practices, identify gaming opportunities before they become widespread, and negotiate mutually acceptable evolution of requirements. This iterative learning arguably generates more robust long-term outcomes than would emerge from one-time design of perfect rules, because climate governance operates in non-stationary environment where technologies, policies, and industrial structures continually shift.

The limitation is that learning systems can drift or degrade if attention wanes. Maintaining productive dialogue requires sustained commitment and resources. Economic pressures or political changes could lead to under-investment in CAMWG operations, reducing capacity for sophisticated evaluation and creating opportunities for low-quality modules to gain recognition through regulatory capture or simple neglect. The governance architecture must build resilience against such decay through redundant oversight mechanisms and periodic external review.

8.3 Geopolitical Economy and Procedural Fairness

Methodological sovereignty carries geopolitical significance extending beyond technical optimization. Control over carbon accounting standards influences which industrial configurations remain viable, which countries can export profitably to climate-conscious markets, and which technical paradigms become globally dominant.

European leadership in establishing CBAM and defining carbon accounting standards embeds assumptions about how industries should organize (market-based coordination, arm's-length transactions, facility-level optimization) that may systematically disadvantage non-European industrial models (state-coordinated industrial parks, long-term relational contracting, system-level optimization). Modular CBAM can reduce this bias but cannot eliminate it entirely if interface standards themselves embed European ontological commitments. Requiring third-party verification privileges countries with deep professional services sectors and may disadvantage countries relying more on state verification capacity. Defining system boundaries at legal-entity level privileges countries where legal and operational boundaries align.

This is not self-contradiction but honest acknowledgment of power asymmetries that modular architecture cannot eliminate. CAMWG would not be forum among equals—European regulatory capacity, market size, and institutional sophistication provide structural advantages in shaping outcomes. What modular architecture achieves is procedural fairness rather than equality of power.

Consensus-based decision-making on interface standards ensures no unilateral changes imposed by any bloc. Regional veto power for standards affecting more than 30% of region's exports provides defensive shield against disadvantageous modifications. Rotating chairship among EU, Asia-Pacific, Latin America, Africa ensures no permanent control by any single region. Independent technical secretariat housed in neutral location (Geneva, not Brussels) reinforces operational autonomy from European institutions. Public proceedings and challenge mechanisms enable civil society monitoring preventing backroom deals.

The goal is not equality of power but equality of methodological legitimacy—ensuring GB/T, PAT, I-REC receive same evidentiary treatment as ISO in CBAM recognition processes. Chinese enterprises using GB/T should face same burden of proof demonstrating environmental effectiveness as European enterprises using ISO, no more and no less. This procedural evenhandedness cannot eliminate European structural advantages but can prevent their arbitrary exercise.

The proposed framework attempts navigating these tensions by making ontological commitments explicit and subject to negotiation rather than treating them as technical necessities. But this navigation occurs within constraints of geopolitical power. Whether modular CBAM serves as genuine accommodation of diversity or sophisticated incorporation of peripheral economies into core regulatory frameworks remains question empirical implementation would test.

8.4 Limitations and Alternative Approaches

This analysis carries limitations worth explicit acknowledgment. First, empirical evidence base for some claims—particularly regarding I-REC fraud rates, PAT's environmental effectiveness in lower tiers, and GB/T's accuracy for capturing industrial symbiosis benefits—remains thin. More extensive independent assessment would strengthen or potentially challenge characterization of these methodologies as rigorous alternatives to ISO frameworks.

Second, the paper has focused on methodological diversity without deeply examining carbon pricing levels or allocation of costs between producing and consuming countries—arguably more important dimensions of CBAM design from distributive justice perspectives. Comprehensive climate trade policy would need to address these alongside methodological questions.

Third, modular approach assumes interface standardization can meaningfully separate methodology from outcomes, enabling assessment of environmental effectiveness independent of methodological details. This assumption may not hold if methodological choices systematically bias results in ways interface requirements fail to detect.

Fourth, governance architecture proposed here might prove too complex for practical implementation given limited institutional capacity and competing priorities among potential participants. Simpler alternatives—either stricter convergence toward single standard or looser recognition of national certifications with minimal interface requirements—might achieve better results if governance capacity constraints bind more tightly than analysis suggests.

Alternative approaches merit consideration. One option would maintain methodological convergence pressure but adjust CBAM implementation timelines to allow longer transition periods for developing countries and smaller enterprises to build compliance capacity. Another option would replace carbon border adjustments with consumption-based carbon accounting and climate clubs offering preferential trade access to members meeting shared standards. A third option would fully decouple carbon accounting from trade measures, instead using development finance to subsidize monitoring infrastructure buildout in developing countries.

Conclusion

The challenge of measuring carbon emissions across heterogeneous industrial systems, governance capacities, and development contexts is epistemological in nature. The empirical evidence presented here—from China's industrial ecology practices to India's differentiated governance to Brazil's renewable energy verification—suggests observed methodological diversity often reflects rational adaptation to context rather than developmental lag behind universal best practice.

A critical question remains: does methodological diversity compromise global carbon accounting integrity? The opposite may be true. Current CBAM architecture achieves apparent precision through uniform methodological requirements that most enterprises cannot meet, resulting in 70-85% reliance on default values that systematically overestimate or underestimate actual emissions. Modular CBAM trades methodological uniformity for verified diversity—multiple frameworks achieving comparable rigor through different epistemic approaches, with digital infrastructure (blockchain-based certificates, automated meter-to-registry transmission) providing unprecedented traceability. When Conservative Principle adjusts for precision differentials and Comparative Benchmarking flags systematic divergence, aggregate emissions accounting becomes more reliable than under nominally uniform system that cannot be implemented. The 1.5°C carbon budget is best protected not by epistemological monoculture but by methodological ecosystem that maintains rigor while respecting contextual rationality.

The modular architecture proposed here offers one pathway through this challenge. By shifting focus from methodological convergence to interface standardization, it creates space for epistemological pluralism while maintaining environmental rigor and cross-system comparability. The nested governance structure embodied in CAMWG operationalizes Ostrom's principle that robust commons management requires multi-level institutional coordination. Integration with emerging digital infrastructure—particularly blockchain-based verification and product passport systems—provides technical means to implement distributed trust mechanisms reducing fraud risk without requiring centralized authority. Critical safeguards including Conservative Principle prevent methodological arbitrage while preserving diversity.

Yet this proposal cannot escape its own situatedness. It emerges from analytical tradition valuing participatory governance, acknowledging path dependency in institutional development, and prioritizing long-term resilience over short-term optimization. Alternative traditions emphasizing standardization's benefits, questioning coherence of non-Western methodological frameworks, or prioritizing rapid decarbonization over procedural inclusiveness might reach different conclusions.

For CBAM specifically, this implies that current implementation trajectories treating non-ISO methodologies as temporary deviations pending harmonization may inadvertently undermine the policy's environmental effectiveness and political durability. If diverse methodologies represent co-produced knowledge aligned with different industrial ecologies rather than inferior approximations of European standards, then forcing convergence destroys valuable contextual fit while generating political resistance threatening broader climate trade policy project.

For climate governance generally, the analysis highlights tension between two approaches to global coordination. One path seeks universal standards, treating heterogeneity as coordination problem to be solved through negotiation or power asymmetries. Another path accepts heterogeneity as inherent, treating global coordination as interoperability problem requiring interface standardization but not internal homogenization. The former aligns with mid-twentieth-century system optimization thinking and has achieved notable successes in domains amenable to technical standardization. The latter reflects late-twentieth-century lessons from complex adaptive systems research and distributed computing architecture, suggesting some coordination challenges require preserving rather than eliminating diversity.

Climate change's planetary scope and generational timescale creates pressure toward universal frameworks. But irreducible diversity of industrial systems, governance capacities, and development contexts creates countervailing pressure toward localized approaches. Threading this needle—achieving sufficient coordination for effective climate response while preserving sufficient diversity for contextual fit and political legitimacy—represents one of central governance challenges of current century.

The 1.5°C carbon budget constraint is non-negotiable. Multiple epistemic pathways can reach this physical target—the question is whether we construct governance systems enabling diverse knowledge traditions to contribute their contextual wisdom or whether we insist on epistemological monoculture potentially sacrificing both effectiveness and equity. This paper has argued for the former while acknowledging its complexities and limitations. The stakes extend beyond carbon accounting to fundamental questions about how global institutions can function across heterogeneous knowledge systems in an era of persistent civilizational plurality.

References

Baldwin, C. Y., & Clark, K. B. (2000). Design Rules: The Power of Modularity. MIT Press.

Beer, S. (1972). Brain of the Firm. John Wiley & Sons.

Cames, M., Harthan, R. O., Füssler, J., Lazarus, M., Lee, C. M., Erickson, P., & Spalding-Fecher, R. (2016). How Additional Is the Clean Development Mechanism? Analysis of the Application of Current Tools and Proposed Alternatives. Öko-Institut e.V.

Checkland, P. (1981). Systems Thinking, Systems Practice. John Wiley & Sons.

Checkland, P., & Poulter, J. (2006). Learning for Action: A Short Definitive Account of Soft Systems Methodology and Its Use for Practitioners, Teachers, and Students. John Wiley & Sons.

Coase, R. H. (1937). The Nature of the Firm. Economica, 4(16), 386-405.

EPE (Empresa de Pesquisa Energética). (2024). Brazilian Energy Balance 2024. Ministry of Mines and Energy, Brazil.

European Commission. (2024). Report on CBAM Implementation (COM(2024) 783). Brussels.

Finnveden, G., Hauschild, M. Z., Ekvall, T., Guinée, J., Heijungs, R., Hellweg, S., Koehler, A., Pennington, D., & Suh, S. (2009). Recent developments in Life Cycle Assessment. Journal of Environmental Management, 91(1), 1-21.

Fricker, M. (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford University Press.

Frosch, R. A., & Gallopoulos, N. E. (1989). Strategies for Manufacturing. Scientific American, 261(3), 144-152.

Fuller, L. L. (1969). The Morality of Law (Revised ed.). Yale University Press.

Graedel, T. E., & Allenby, B. R. (1995). Industrial Ecology. Prentice Hall.

Hertwich, E. G., & Peters, G. P. (2009). Carbon Footprint of Nations: A Global, Trade-Linked Analysis. Environmental Science & Technology, 43(16), 6414-6420.

Holling, C. S. (1973). Resilience and Stability of Ecological Systems. Annual Review of Ecology and Systematics, 4, 1-23.

I-Track Foundation. (2025). I-REC Market Report Q2 2025. https://www.itrackfoundation.org

India Bureau of Energy Efficiency. (2020). PAT Cycle II Completion Report. Ministry of Power, Government of India.

Jasanoff, S. (Ed.). (2004). States of Knowledge: The Co-production of Science and Social Order. Routledge.

Longino, H. E. (2002). The Fate of Knowledge. Princeton University Press.

Mol, A. (2002). The Body Multiple: Ontology in Medical Practice. Duke University Press.

North, D. C. (1990). Institutions, Institutional Change and Economic Performance. Cambridge University Press.

Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.

Pierson, P. (2000). Increasing Returns, Path Dependence, and the Study of Politics. American Political Science Review, 94(2), 251-267.

Reap, J., Roman, F., Duncan, S., & Bras, B. (2008). A survey of unresolved problems in life cycle assessment. International Journal of Life Cycle Assessment, 13(5), 374-388.

Simon, H. A. (1956). Rational Choice and the Structure of the Environment. Psychological Review, 63(2), 129-138.

Simon, H. A. (1962). The Architecture of Complexity. Proceedings of the American Philosophical Society, 106(6), 467-482.

Tapscott, D., & Tapscott, A. (2016). Blockchain Revolution: How the Technology Behind Bitcoin Is Changing Money, Business, and the World. Portfolio.

UN Comtrade. (2023). International Trade Statistics Database. United Nations. https://comtrade.un.org

Williamson, O. E. (1985). The Economic Institutions of Capitalism. Free Press.

Publication & Licensing

Title: Modular CBAM: Epistemological Pluralism in Carbon Accounting
Version: 1.0 | January 2026
Author: Alex Yang Liu
Publisher: Terawatt Times Institute | ISSN 3070-0108
Document ID: MCBAM-V1.0
Citation Format: Liu, A. Y. (2026). Modular CBAM: Epistemological Pluralism in Carbon Accounting. Terawatt Times (ISSN 3070-0108), v1.0. DOI: [To be assigned]

Copyright & Use

Copyright © 2026 Terawatt Times Institute. All rights reserved.

This work presents a theoretical framework for modular carbon accounting governance, introducing the Conservative Principle mechanism for precision-adjusted tariffs, the CAMWG nested governance architecture, and epistemological pluralism as foundation for equitable climate trade policy.

You are free to:

▷ Read, cite, and reference this work
▷ Use it for academic research, policy analysis, and education
▷ Share the document in full or in part, with proper attribution
▷ Discuss, critique, and apply the theoretical framework for governance analysis
▷ Reference the modular CBAM architecture and Conservative Principle concepts in scholarly discourse

Commercial or Engineering Use Requires Licensing

Any form of implementation, replication, or derivative deployment of the modular CBAM framework, Conservative Principle mechanism, or CAMWG governance architecture requires explicit written permission from the Terawatt Times Institute. This includes, but is not limited to:

– Operationalizing the Conservative Principle formula (Tariff Basis = μ + k·σ) in carbon pricing or trade systems
– Implementing the modular architecture design or CAMWG governance structure in climate institutions
– Developing commercial carbon accounting assessment products derived from this framework
– Use in professional consulting, advisory services, or policy implementation products
– Engineering pilot programs or operational systems that implement the framework beyond citation or illustrative use

Permissions & Licensing

For permissions, licensing inquiries, or authorized derivative use: alex.liu@terawatttimes.org Terawatt Times Institute