Trade in the telecoupling framework: evidence from the metals industry

As a conceptual framework for understanding contemporary sustainability challenges, telecoupling emphasizes the importance of socioeconomic and environmental interactions over long distances. These long-distance interactions can occur through multiple human activities. We focus on international trade, a major channel of telecoupling flows, and in particular on the international trade of metals. We use the data of physical products and embedded greenhouse gas (GHG) emissions trade in the World Input-Output Database (WIOD) to quantitatively examine how countries contribute to both economic and environmental flows through the trade of metals, but also how that contribution varies depending on their position in the global value chain (GVC) of contemporary international trade. This analysis is built on previously developed techniques for decomposing gross exports of products, which we apply to examine embedded GHG emissions. We make comparisons between countries' contributions to flows of economic value versus embedded GHG emissions, but also examine contributions beyond total volumes of trade and bilateral trade. Specifically, we quantify the economic and environmental spillover effects that occur in contemporary international trade because of the GVC in which flows of intermediate goods form components in other subsequently traded goods. We interpret differences between countries' contributions to the flows of economic value versus embedded GHG emissions as being related to the intensity and efficiency of resource use during production. In turn, differences in contributions to direct trade flows versus spillover flows are related to their positions in the GVC. Subsequently, we discuss other elements of the telecoupling framework in trade, i.e., agents, causes, and effects. Quantitatively incorporating these telecoupling framework elements alongside spillover flows will enable investigation of dynamics and relationships that traditional trade theories, data, and models do not currently account for well.