Sustainable Supply Chains via Data-Driven Risk Analytics

As a result of the outsourcing trend, 80 percent of global trade today flows through multinational corporations and one in five jobs worldwide can be tied to global supply chains. Therefore, global supply chains possess enormous potential to contribute to sustainable development around the world. Yet, they continue to be plagued by persistent social and environmental responsibility violations, such as child and/or forced labor, hazardous working environments, excessive overtime, poor wages, and pollution.

These violations not only pose societal burdens but can also create significant economic inefficiencies for firms in the forms of product recalls, fines, problem corrections, supply interruptions, stock value loss, and brand damage. To make a compelling business case for sustainable supply chains and eventually create them, firms must quantify the financial impacts of the responsibility violations of their suppliers, gain greater visibility into the responsibility-related risks of their suppliers for a more effective allocation of inspection resources, and implement appropriate decision support tools for managers to account for responsibility-related risks in sourcing decisions.

The objective of this research is to couple data-analytics approaches with the newly available datasets on social and/or environmental violations of suppliers to generate the knowledge base needed for firms to move in this direction. This research develops descriptive and predictive statistical models, and behavioral decision support tools that can be deployed by firms to proactively manage responsibility in their supply chains.

This research bridges an important methodology gap in supply chain management by utilizing data-driven approaches for managing supplier responsibility. First, this research utilizes machine learning and language processing approaches to (1) link public supplier environmental violation records with proprietary supply network data, and (2) enable a systematic quantitative evaluation of how suppliers’ responsibility violations affect the financial performance of buying firms for the first time.

Second, statistical models are developed to detect key common characteristics of suppliers that are revealed as environmental violators and to predict “risky” suppliers. These models will enable buying firms, governmental and/or nongovernmental agencies to allocate limited inspection resources according to predicted risk to achieve cost-effectiveness.

Third, behavioral mechanisms are designed and tested through a closed-loop experimental methodology to nudge managers into considering responsibility-related risks in sourcing decisions. Taken together, this research facilitates a shift from “reactive” to “proactive” in the way firms view and manage social and environmental responsibility in their supply chains.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.