The environmental Kuznets curve (EKC) hypothesizes an inverted-U relationship between income per capita and environmental degradation. When CO2 emissions are measured on a consumption basis — using input-output analysis to add emissions embodied in imported goods — the EKC turning point shifts to substantially higher income levels, suggesting that part of rich countries' apparent delinking from emissions reflects displacement of carbon-intensive production through international trade rather than genuine global decoupling. In the spatial econometric framework used to estimate this relationship, neighboring countries' income and emissions spill over across borders, but the mechanisms generating these spillovers — trade-induced pollution transfer, technology diffusion, or environmental-policy mimicry — have not been identified or quantified. The problem is to decompose these channels in the cross-country, consumption-based EKC, to test robustness of the spillover estimates to alternative spatial weight specifications (in particular matrices derived from input-output trade flows), and to integrate structural-decomposition-derived driving forces directly into the EKC estimation framework so that the contribution of each driving force to the curve can be measured.

Background

The environmental Kuznets curve (EKC) is the empirical hypothesis, introduced in the early 1990s, that environmental degradation first rises and then falls with per-capita income as economies develop, producing an inverted-U relationship. Grossman and Krueger (1995) framed the income–pollution link through three effects: a scale effect (more output means more pollution), a composition effect (the sectoral mix of production changes), and a technique effect (cleaner technology per unit of output). Antweiler, Copeland and Taylor (2001) embedded these in a general-equilibrium model of trade and environment and showed that trade's effect on a country's pollution depends on the balance between its pollution haven motive (pollution-intensive industries relocate to countries with weak environmental regulation) and its factor endowments. Most EKC studies measure emissions on a production basis, i.e., emissions released within a country's borders; this ignores international trade, because goods consumed in one country may embody emissions produced elsewhere. Consumption-based accounting uses input-output analysis — a framework that traces inter-industry transactions — applied to multi-region input-output (MRIO) tables to reallocate emissions to the final consumer, yielding a consumption-based CO2 indicator that includes emissions embodied in trade. Comparing the two indicators reveals whether rich countries' emission reductions are genuine or partly transferred abroad. Haller (2014) combined both ideas in a spatial econometric setting: spatial econometrics recognizes that a country's emissions may depend on its neighbors' income and emissions, with the strength of this dependence encoded in a spatial weight matrix WW whose entries quantify pairwise proximity or interaction; a spatial panel model (e.g., a spatial Durbin model, which includes both spatially lagged dependent variables and spatially lagged regressors) then estimates direct effects and spatial spillover effects. Using consumption-based CO2 for a panel of countries, Haller found an EKC for both production- and consumption-based indicators, but with the consumption-based curve turning at a much higher income level, the difference being attributable to the shift of emissions via trade. The study also performed a structural decomposition analysis (SDA) — an accounting method that decomposes the change in emissions over time into driving forces such as emission intensity, sectoral composition, and final-demand level — to describe the forces behind the EKC, but did not connect them to the regression. The concluding discussion left the central interpretive question open: the estimated spatial spillovers could arise from trade itself, from international technology diffusion, or from strategic interaction in environmental policy (so-called policy mimicry, whereby governments match neighboring regulation levels, as documented in the environmental federalism literature), and these channels were not separately identified. Subsequent work has advanced the adjacent pieces — consumption-based accounting in the trade–carbon nexus (Liddle 2018; Makarov and Alataş 2024) and spatial-econometric EKC estimations across various regions and weight matrices (Karahasan and Pinar 2022; Jeetoo and Chinyanga 2023; Bolea et al. 2024; Wang and Zhu 2025) — but the question of which mechanism drives the spillovers, and how much of the production–consumption EKC gap each channel explains, remains unanswered in the cross-country setting posed by the source.

Problem Statement

For the cross-country relationship between per-capita income and CO2 emissions measured on both a production basis and a consumption basis (the latter constructed from multi-region input-output analysis so that emissions embodied in trade are attributed to final consumers), identify and quantify the mechanisms that generate the estimated spatial spillover effects in the environmental Kuznets curve. Specifically: (i) decompose the spatial spillovers into the three postulated channels — trade-induced transfer of pollution-intensive production (the pollution haven channel), international diffusion of cleaner technology, and strategic interaction or mimicry in environmental policy — and estimate each channel's contribution, where, because the three channels are observationally similar in reduced form, the decomposition must rest on a separate identification of each channel rather than on relabeling spatial-lag coefficients obtained under different weight matrices; (ii) determine how much of the observed gap between the production-based and consumption-based EKC (in particular the shift of the turning point to higher income under consumption-based accounting) is accounted for by each channel, with each channel's contribution defined through an explicitly stated attribution rule and reported in units that make the contributions comparable; (iii) test the robustness of the spillover estimates to alternative spatial weight matrix specifications, including a matrix derived from input-output trade flows between countries rather than pure geographic distance; and (iv) integrate the driving forces from structural decomposition analysis (emission intensity, sectoral composition, final-demand level and mix) into the EKC framework, so that their importance to the shape of the curve is tested directly rather than only described separately. These four parts are jointly posed. The source's discussion also flags secondary robustness and extension concerns — adequacy of sample length and country heterogeneity (each country may have its own EKC), potential reverse causality from environmental quality to income, the influence of environmental-policy episodes such as the Kyoto Protocol on trade patterns and the curve, chain-linked rather than double-deflated price conversion of the input-output tables, and extension beyond CO2 to other greenhouse gases and local pollutants — which provide context for the analysis rather than additional mandatory deliverables. The identification strategy used in the analysis must be defensible with respect to reverse causality and other endogeneity concerns and must not assume them away; this is a condition on how the four parts above are executed, not an additional deliverable.

References

  1. Bastien Haller (2014), 'The Implications of International Trade on the Environmental Kuznets Curve', SSRN preprint, DOI 10.2139/ssrn.2527302, https://doi.org/10.2139/ssrn.2527302.
  2. G. M. Grossman and A. B. Krueger (1995), 'Economic Growth and the Environment', Quarterly Journal of Economics 110(2), 353-377, DOI 10.2307/2118443, https://doi.org/10.2307/2118443.
  3. Werner Antweiler, Brian R. Copeland and M. Scott Taylor (2001), 'Is Free Trade Good for the Environment?', American Economic Review 91(4), 877-908, DOI 10.1257/aer.91.4.877, https://doi.org/10.1257/aer.91.4.877.
  4. Brantley Liddle (2018), 'Consumption-based accounting and the trade-carbon emissions nexus', Energy Economics 69, 71-78, DOI 10.1016/j.eneco.2017.11.004, https://doi.org/10.1016/j.eneco.2017.11.004.
  5. Igor Makarov and Sedat Alataş (2024), 'Production- and consumption-based emissions in carbon exporters and importers: A large panel data analysis for the EKC hypothesis', Applied Energy 363, 123063, DOI 10.1016/j.apenergy.2024.123063, https://doi.org/10.1016/j.apenergy.2024.123063.
  6. Lei Jiang, Shixiong He, Zhangqi Zhong, Haifeng Zhou and Lingyun He (2019), 'Revisiting environmental Kuznets curve for carbon dioxide emissions: The role of trade', Structural Change and Economic Dynamics 50, 245-257, DOI 10.1016/j.strueco.2019.07.004, https://doi.org/10.1016/j.strueco.2019.07.004.
  7. Zhangqi Zhong, Lei Jiang and Peng Zhou (2018), 'Transnational transfer of carbon emissions embodied in trade: Characteristics and determinants from a spatial perspective', Energy 147, 858-875, DOI 10.1016/j.energy.2018.01.008, https://doi.org/10.1016/j.energy.2018.01.008.
  8. Shun Jia Liu and Xin Long Xu (2021), 'The pollution halo effect of technology spillover and pollution haven effect of economic growth in agricultural foreign trade: two sides of the same coin?', Environmental Science and Pollution Research 28(16), 20157-20173, DOI 10.1007/s11356-020-11786-w, https://doi.org/10.1007/s11356-020-11786-w.
  9. Libo Wu, Meimin Yang and Chujun Wang (2021), 'Strategic interaction of environmental regulation and its influencing mechanism: Evidence of spatial effects among Chinese cities', Journal of Cleaner Production 312, 127668, DOI 10.1016/j.jclepro.2021.127668, https://doi.org/10.1016/j.jclepro.2021.127668.
  10. Burhan Can Karahasan and Mehmet Pinar (2022), 'The environmental Kuznets curve for Turkish provinces: a spatial panel data approach', Environmental Science and Pollution Research 29(17), 25519-25531, DOI 10.1007/s11356-021-17706-w, https://doi.org/10.1007/s11356-021-17706-w.
  11. Jamiil Jeetoo and Earnest Rungano Chinyanga (2023), 'A spatial econometric analysis of the environment Kuznets curve and pollution haven hypothesis in Sub-Saharan Africa', Environmental Science and Pollution Research 30(20), 58169-58188, DOI 10.1007/s11356-023-26306-9, https://doi.org/10.1007/s11356-023-26306-9.
  12. Lucía Bolea, Adrián Espinosa-Gracia and Sofía Jimenez (2024), 'So close, no matter how far: A spatial analysis of CO2 emissions considering geographic and economic distances', The World Economy 47(2), 544-566, DOI 10.1111/twec.13424, https://doi.org/10.1111/twec.13424.
  13. Zining Wang and Pengyu Zhu (2025), 'Trade of eco-friendly products and CO2 emissions in China: A spatial spillover analysis within the EKC framework', Journal of Cleaner Production 516, 145734, DOI 10.1016/j.jclepro.2025.145734, https://doi.org/10.1016/j.jclepro.2025.145734.