Measuring and predicting supply chain resilience under disaster and pandemic disruptions: Evidence from Japan’s automotive sector

This study investigates the supply chain resilience (SCR) of Japanese automotive manufacturers by integrating perspectives from disaster risk and operational disruption. A hybrid analytical framework that combines a logistic regression (LR) scorecard and a long short-term memory (LSTM) model is employed. This framework uses publicly available data from 2015 to 2024, which consists of information from seven major automakers. Resilience is assessed across internal and external dimensions in response to disruptions such as the COVID-19 pandemic and natural disasters. Explanatory variables include production volume, market concentration level, employee tenure, recall rate, exchange rate, and disaster indices. The LR scorecard identifies leading indicators of disaster resilience by estimating the probability of achieving post-disruption peak profitability (AUROC = 0.83), while the LSTM model captures time-series predictors of profit recovery (test MSE = 0.0194). Results show that long employee tenure, stable production, and strong quality control are the most critical enablers of supply chain continuity. The findings highlight the importance of operational resilience as a foundation for reducing disaster risk in Japan’s automotive sector. Future research should extend the analysis to global supply chains and explore causal inference approaches to better guide policy and strategic decision-making. • Assesses supply chain resilience in Japanese automotive manufacturers using a hybrid approach combining logistic regression scorecards and LSTM neural networks. • Finds long employee tenure, production stability, and strong quality control are the most critical factors for post-disruption profitability. • Machine learning models achieve high predictive accuracy, enabling robust estimation and forecasting of resilience. • Highlights the importance of operational factors over market concentration or short-term HR measures. • Recommends expanding analysis to global supply chains and applying causal inference for broader insights.