Financial and Sustainability Factors of Supply Chain Resilience: A Hybrid Econometric-ML Model

This paper aims to identify the financial and sustainability factors that determine the resilience of road freight transport firms and, to this end, proposes a hybrid econometricmachine learning (ML) framework for measurement and prediction. Thus, we analysed data from 1,500 Romanian firms over the period 2014-2023 to construct a resilience score using principal component analysis (components retained according to the standard criteria and weighted by the explained variance) and a sustainability score, for which robustness is tested through sensitivity analyses of the weights. The main results emphasise that resilience is associated with a more prudent financial structure and investments in human capital, and the environmental component becomes more important when nonlinearities and interactions with financial factors are allowed. Our analysis clearly shows that ML modelling improved the predictive performance of the models, providing managers with a theoretical and practical basis for prioritising sustainability investments and early diagnosis of operational vulnerabilities in logistics. In essence, the paper contributes to the definition of a resilience score using a replicable combined econometrics-ML methodology for explanation and prediction.