Sustainability performance of SMEs: a causal inference approach to the firm strategy and interval-scale DEA

We examine the sustainability performance of Smalland Medium-Sized Enterprises (SMEs) across four European countries, focusing on operational, economic, social, and environmental practices. Data were collected through a Likert-scale questionnaire. While Data Envelopment Analysis (DEA) has been widely used to assess SMEs' sustainability, most studies rely on conventional DEA models. This study adopts an interval-scale DEA model to better accommodate interval-scale data. We then apply dimensionality reduction and visualization techniques to explore intra- and inter-country differences at both aggregate and SME levels. Finally, we introduce a novel HHI-GPSM-LASSO framework to assess the causal effect of SMEs’ resource allocation strategies on efficiency. This framework integrates a Herfindahl-Hirschman-like index, Generalized Propensity Score Matching, and weighted Lasso regression to uncover the causal relationships between SMEs' resource allocation strategies and efficiencies. While visualization enhances interpretability for practitioners, the causal framework supports strategic and policy decisions. To our knowledge, this is the first DEA study that combines interval-scale DEA, advanced visualization, and causal inference to inform sustainability benchmarking and policy.