Abstract
Scope 3 emissions—indirect greenhouse gas emissions across a company’s value chain—often constitute the largest share of corporate carbon footprints yet remain inadequately measured due to methodological challenges. Unlike the GHG Protocol’s general guidelines, which leave multi-tier allocation, circular dependencies, and complex ownership to individual practitioners, this paper develops an integrated, algorithmically implementable framework for multi-tier Scope 3 emission calculation. Our three-component methodology comprises: a tier-specific economic allocation model using revenue–transaction ratios, a path-based circular dependency resolution algorithm, and an equity-weighted attribution mechanism for complex ownership structures. Unlike sector-level approaches such as Economic Input–Output Analysis (EIOA) and Multi-Regional Input–Output (MRIO) models, our framework operates at the firm-transaction level, enabling company-specific accountability. We empirically validate this framework using data from Chinese companies across nine industries, tracing carbon flows through supply chains up to five tiers deep. Tier 1 suppliers account for a median of 99.6% (bootstrap 95% CI [98.5%, 99.8%]; mean: 89.4%) of total calculated emissions (mean 37,305.1 tCO 2 ), confirmed by Friedman test ( $$\chi ^2 = 285.40$$ χ 2 = 285.40 , $$p = 1.52 \times 10^{-60}$$ p = 1.52 × 10 - 60 ) and pairwise Wilcoxon signed-rank tests (all $$p_{\textrm{BH}} < 0.001$$ p BH < 0.001 ). Comparison with company-reported Scope 3 reveals industry-specific discrepancies, with calculated-to-reported ratios ranging from near-zero (Real Estate) to 3.157 (Industrials), indicating significant heterogeneity and potential under-reporting. We identify dominant intra-industry carbon flows within Industrials (2,099,959.4 tCO 2 ) and significant cross-sector transfers from Materials to Financials (91,183.8 tCO 2 ). Network analysis reveals that 32.5% of supply chain nodes participate in circular dependencies, with the largest strongly connected component containing 947 nodes; our resolution algorithm eliminates double counting while preserving mathematical consistency. These findings support focusing carbon reduction efforts on Tier 1 and Tier 2 suppliers and carry direct implications for China’s dual carbon goals through mandatory supply chain carbon disclosure frameworks and industry-specific reduction targets.