Data-driven Methods for the Risk Analysis of Global Supply Chains

Supply chains have become more global due to the increased interconnectedness of the world’s economy. As a result, companies in the U.S. are seeking more affordable overseas supply alternatives, though doing so may increase risk of domestic impacts when a disruption in the supplier’s country occurs. It is then critical to assess the risks of global disruptive events to identify the cost effectiveness of overseas operations and risk management strategies. This work deploys statistical methods to predict the likelihood of a global supply chain disruption given a particular set of attributes related to the supplier characteristics and demographics, their risk management strategies, and their suppliers. The data analysis shows that the factors contributing to the likelihood of a disruption include the location of suppliers, the supplier resources to manage risk, and their insurance expenditure. Also, the Bayesian kernel model provided a higher accuracy than traditional statistical techniques. The outcome of the research provides decision makers with a flexible and reliable tool to predict the likelihood of disruptions and identify risk management practices that mostly impact the resilience of the company’s supply chain.