Toward Clearer Impact Attribution: Endogenizing System Losses and Coupled Inputs in Input-Output Based Material Flow and Impact Analysis

Material Flow Analyses (MFAs) are typically modeled as directed graphs, where energy and impact intensities (e.g., greenhouse gas emissions) can be assigned to inputs and processes (i.e., the nodes of the graph) and then analyzed using input-output (I/O) techniques to attribute system-wide impacts either to end-use consumption sectors—via demand-driven (Leontief) I/O analysis—or to input sectors—via supply-driven (Ghosh) I/O analysis. However, each attribution method has limitations. In demand-driven analysis, waste flows (e.g., losses) appear as terminal nodes, which can obscure the true drivers of these losses. It will often be insightful to reallocate the impacts assigned to these waste nodes back to the consumption sectors that ultimately generated them. Similarly, in supply-driven analysis, some input sectors are required solely to support the processing of primary materials; e.g., magnesium used as an alloying element in aluminum production. In such cases, it may be useful to reallocate the impacts of these complementary inputs to the primary inputs that determine their necessity. In this article, we explore and define models for such re-allocations for both supply-driven and demand-driven I/O analysis applied to MFA. These techniques are demonstrated through an example, stylized, MFA network as well as a case study on the U.S. steel sector, highlighting the significant effect such reallocation can have on the modeled impact of source nodes or end-use consumption sectors; e.g., the greenhouse gas emissions attributable to the 7 U.S. steel sector consumption nodes increases by an average of 38%. This article’s methodology offers a systematic approach for evaluating environmental impacts attributable to individual nodes, thereby facilitating informed decisions regarding decarbonization strategies.