Abstract
Large corporations contribute to waste proliferation and global environmental degradation; however, more research is needed to understand these trends. Unsupervised machine learning can provide unique insights into how firms cluster on different sustainability indicators with relevance for the circular economy (CE) framework (i.e., metrics related to waste reduction and promotion of reuse). This study uses data from 2010 – 2020 for 2,485 global companies to explore how firms cluster together on sustainability metrics and whether a firm’s headquarter location and industry determine its clustering. Using the k-means unsupervised machine learning algorithm, one-way ANOVA, and Chi-square tests for key variables (e.g., industry, location of headquarters, total environmental cost and progress toward the UN Sustainable Development Goals 12.2, 14.1 and 14.2) across clusters, I found that the majority of firm-year observations cluster together (n = 13,313), with a small minority of firm-years (n = 39) with the most negative CE and environmental impacts clustered together, as well. The key findings indicated that the k-means algorithm grouped firms into four distinct clusters. Firms headquartered in the EU were not more likely to be in the most sustainable cluster, while firms in extractive industries (e.g., fossil fuel and mining) were more likely to be in the least sustainable cluster. These results can help policymakers identify key factors that could influence firms to adopt business practices aligned with circular economy goals worldwide while also enhancing the understanding of complex interfaces between corporate sustainability and public policy.
Presenters
Gloria SchmitzStudent, PhD Candidate, M.A., Graduate Certificate in Data Analytics, Northeastern University, Massachusetts, United States
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
2025 Special Focus—Sustainable Development for a Dynamic Planet: Lessons, Priorities, and Solutions
KEYWORDS
Machine Learning, Corporate Social Responsibility, Circular Economy, Sustainable Development Goals
