Environmental Policy Stringency and the Composition of Carbon and Grey-Water Footprints in Global Supply Chains: Country-Industry Evidence from an MRIO Framework
Published:
Status: Working Paper / Preprint, August 2026.
Authors:
Jiajing Sun, Weimin Jiang, Ruting Wang, and Michael Cole.
Research Question and Contribution:
Climate and air emissions are increasingly visible through policy indicators and supply-chain accounts, whereas water-quality pressures remain harder to trace internationally. This paper examines how environmental policy stringency is associated with the composition of carbon and grey-water footprints embodied in global supply chains.
The multi-country, multi-sector model links environmental governance to expected pollution prices, producers’ choices over abatement and output, and buyers’ choices over sourcing and product mix. Carbon and water-pollutant abatement reduce their own residual loads but compete for shared engineering and treatment capacity. The model separates strict substitution from proportional rebalancing, then connects these production responses to embodied-trade balances through Armington sourcing and an exact accounting identity.
Data and Empirical Design:
The empirical analysis combines Eora26 final-demand footprints for 40 countries and 25 industries over 2010-2018 with the OECD Environmental Policy Stringency index. Carbon and grey-water accounts are constructed on the same unaggregated production system, using a common intermediate matrix, output vector, Leontief inverse, and final-demand matrix. The analysis reports a pooled all-industry benchmark and a complete set of 14 goods-industry comparisons relative to services and other non-manufacturing industries, with country-clustered and family-wise inference.
Main Results:
The pooled grey-water-minus-carbon balance coefficient is close to zero (-0.0078; p = 0.369; country-cluster wild p = 0.394). Across the 14 goods-industry comparisons, Mining and Quarrying has the largest positive differential relative to services under the primary Q-vintage specification (+0.0577; raw p = 0.004; country-cluster wild p = 0.019; Romano-Wolf family-wise p = 0.083).
The mining differential is generated by a more negative carbon net-export balance (-0.0769; p = 0.059) rather than an increase in the grey-water balance (-0.0192; p = 0.540). A direct total-grey-minus-blue comparison is also positive (+0.0278; p = 0.028). Weighting, trimming, and country-omission checks preserve the positive mining differential.
Interpretation and Scope:
The result is specific to the Leontief final-product allocation and captures the full upstream footprint serving final demand classified as Mining and Quarrying, rather than pollution released only at mine sites. Source-based estimates are close to zero, while specifications with separate account-era effects, country-specific mining trends, first differences, and leads produce smaller estimates. The evidence therefore points to a persistent cross-era supply-chain association, not a causal plant-level substitution mechanism or an absolute rise in water pollution.
Keywords:
Carbon and grey-water footprint composition; multi-pollutant environmental policy; grey-water footprint; embodied carbon; mining; MRIO; global supply chains.
JEL codes: Q56, F18, Q53, Q25, Q58.
Availability:
Recommended citation: Sun, J., Jiang, W., Wang, R., & Cole, M. (2026). Environmental Policy Stringency and the Composition of Carbon and Grey-Water Footprints in Global Supply Chains: Country-Industry Evidence from an MRIO Framework. Working paper / preprint, August 2026.