The research tackles a long-standing blind spot in climate economics. When a country exports goods, the emissions generated in producing them are counted in the exporting nation’s territorial inventory, even though the goods are consumed elsewhere. This mismatch, known as the problem of embodied or trade-embedded emissions, has been studied extensively since Davis and Caldeira’s landmark consumption-based accounting work, but most analyses treat exports as a single undifferentiated flow. Ünal and Erkök argue that this flattening hides the very information policymakers need. An export of finished consumer goods and an export of intermediate components that feed a partner’s own industrial production have very different implications for how carbon pricing, border adjustments, or supply chain regulations will shift emissions between countries.

To unpack these differences, the authors deploy an environmentally extended input-output framework built on the Eora Global Supply Chain Database, combined with a technique called the Multi-Stage Hypothetical Extraction Method. Hypothetical extraction is a thought experiment rendered in matrix algebra: the analyst mathematically removes a sector, or a demand channel, from the economy’s input-output structure and measures how much output and, in this case, how much carbon dioxide the rest of the system loses as a result. The difference between the original economy and the extracted one quantifies the sector’s total systemic importance, capturing both its direct emissions and the emissions embedded in all the upstream linkages that depend on it. The multi-stage refinement goes further by decomposing exports into intermediate goods and final goods, and by splitting final demand into household consumption, government expenditure, and investment channels, allowing emissions to be traced along distinct domestic production stages for each destination market.

The study focuses on six major trading partners of Türkiye: France, Germany, Iraq, Italy, the United Kingdom, and the United States. This selection spans the European Union markets that dominate Turkish trade, the transatlantic relationship, and a neighboring energy economy, providing a diverse test bed for the method. For each partner, the researchers calculated the carbon dioxide emissions embodied in Turkish exports and, crucially, a set of efficiency measures that relate those emissions to the economic value delivered. It is the efficiency dimension that produced the study’s most alarming headline result: a systemic deterioration in carbon dioxide efficiency across most destinations, meaning that progressively more emissions are being generated per unit of export value delivered to these markets.

The structural decomposition reveals that intermediate goods typically dominate the emissions profile of Turkish exports. This matters because intermediate goods occupy a particularly awkward position in climate policy. They cross borders as inputs to further production, so their embodied carbon will only be fully accounted for when the final product is eventually sold, often in a third country. Under mechanisms such as the European Union’s Carbon Border Adjustment Mechanism, which is designed to prevent carbon leakage by pricing the embedded emissions of imports, the treatment of intermediate goods remains one of the most contested design questions. The finding that Türkiye’s export emissions concentrate in these upstream flows suggests that the country’s industrial base is deeply integrated into European and global value chains in ways that make it structurally exposed to any tightening of carbon accounting at the border.

The United Kingdom emerges as a distinctive case. Unlike the other partners, where either intermediate or final goods tend to dominate, the UK’s import profile from Türkiye shows a broad reliance on both categories simultaneously. This dual dependence means that British demand pulls carbon out of the Turkish economy through two separate channels at once, amplifying the total systemic liability and complicating any single-lever policy response. As the UK develops its own post-Brexit carbon border policies, the study’s results imply that Turkish exporters will face pressure from both ends of their product mix rather than from a single vulnerable segment.

Perhaps the most striking results concern the role of government demand, which behaves in sharply different ways depending on the partner. In France and Italy, state-procured government expenditures drive intense emissions liabilities, channeled primarily through Türkiye’s petroleum and chemical sectors. In other words, when the French or Italian state buys goods with Turkish content, the production cascade that follows leans heavily on some of the most carbon-intensive links in the Turkish industrial chain. By contrast, in Iraq and the United States, government expenditure acts less as a direct trigger and more as a deep systemic dependency, anchored in the utility and heavy manufacturing sectors. These partner-specific patterns demonstrate that the same category of final demand can produce qualitatively different carbon consequences depending on the industrial structure it activates, a nuance that aggregate trade-emissions statistics completely obscure.

Underlying all of these patterns is a remarkably concentrated industrial structure. Four sector groups account for the dominant shares of the efficiency effects the authors measure: textiles, transport equipment, petroleum and chemical products, and electricity and gas provisions. This concentration is both a vulnerability and an opportunity. It is a vulnerability because it means Türkiye’s export carbon footprint is not broadly distributed across the economy but is instead pinned to a handful of energy-intensive nodes, several of which are precisely the sectors most likely to be targeted by carbon border measures and by competition from lower-carbon producers. It is an opportunity because targeted interventions at a small number of chokepoints, such as decarbonizing the electricity and gas supply that feeds industrial production, or modernizing petroleum and chemical processing, could deliver disproportionate improvements in the carbon efficiency of the entire export apparatus.

The authors draw a clear policy conclusion from this structure: mitigation efforts must be targeted and partner-specific rather than uniform. Because the emissions liabilities attached to French government procurement arise through different sectors than those attached to American systemic dependencies, a single national decarbonization strategy will not address all exposures equally. Instead, the findings argue for industrial policies and technological investments calibrated to the specific production stages and demand channels that generate the largest efficiency losses for each destination. The open availability of the study’s processed datasets and analysis code in a public GitHub repository lowers the barrier for other researchers to extend the multi-stage extraction approach to other economies, and the method itself, which links classical linkage analysis in the tradition of Hirschman and Rasmussen to modern environmental accounting, offers a template for dissecting how carbon flows through the layered stages of globalized production.

As carbon border regulations spread and trading partners begin pricing the emissions embedded in imports, studies of this kind are likely to become essential reading for policymakers in emerging industrial economies. The Turkish case shows that the question is no longer simply how much carbon a country exports, but through which sectors, along which production stages, and in response to whose demand. Answering that question with the granularity that the multi-stage hypothetical extraction method provides may determine which economies adapt smoothly to a carbon-constrained trading system and which find their industrial linkages transformed into liabilities overnight.

Subject of Research: Export-embodied CO2 emissions and carbon leakage in Turkiye's trade analyzed via multi-stage hypothetical extraction

Article Title: Sectoral linkages and carbon leakage: A multi-stage hypothetical extraction analysis of export-embodied emissions

Article References: Ünal, E., & Erkök, B. (2026). Sectoral linkages and carbon leakage: A multi-stage hypothetical extraction analysis of export-embodied emissions. Journal of Industrial Ecology, 30(4), 2189-2207. https://doi.org/10.1007/s44498-026-00149-2

Image Credits: AI Generated

DOI: 10.1007/s44498-026-00149-2

Keywords: carbon leakage, embodied emissions, input-output analysis, hypothetical extraction method, Turkiye, international trade, CO2 efficiency, intermediate goods, carbon border adjustment, industrial policy, global value chains, Eora database