Social Practice Drivers of Circular Consumption: An Empirical Pathway Model Toward Responsible Consumer Behavior

Circular consumption has become a central component of contemporary sustainability transitions, yet cross-national engagement in circular practices remains highly uneven. This study develops an integrated Social Practice Theory (SPT) and Motivation–Opportunity–Ability (MOA) framework to explain how practice competences, material infrastructures, social norms, and motivational drivers jointly shape circular consumption across 31 countries. Using harmonized indicators from OECD, Eurobarometer, and UN global sustainability surveys, we construct country-level measures of practice components and analyze their effects through a hybrid methodological design. Partial Least Squares Structural Equation Modeling (PLS-SEM) tests theoretically derived causal pathways, while explainable machine learning (XGBoost + SHAP) identifies non-linear dynamics, thresholds, and interaction effects. The PLS-SEM results show that motivation for sustainable living and practice competences are the strongest predictors of circular consumption, with material infrastructure and social norms exerting significant but comparatively smaller effects. Motivation partially mediates the impact of social norms, indicating that normative expectations operate primarily by shaping internalized values rather than direct behavioral compliance. Machine learning findings reveal threshold effects for motivation, diminishing returns for infrastructure, and strong complementarities between competences and materials. A four-cluster country segmentation further uncovers heterogeneous national behavioral profiles (Lifestyle Enthusiasts, Practical Minimalists, Anxious Inactives, and Indifferents), highlighting the need for targeted interventions. Overall, the study demonstrates that circular consumption is an emergent social practice dependent on the alignment of psychological, material, and socio-cultural enablers, offering a robust empirical foundation for designing differentiated policy and managerial strategies that accelerate circular economy transitions. With the global economy remaining only marginally circular and cross-national gaps in sustainable consumption persisting despite widespread policy attention, this study addresses a timely and practically urgent gap by combining a cross-national, multi-source design with an explainable machine-learning complement to conventional structural modeling, thereby offering both theoretical clarification and actionable segmentation guidance for policymakers and managers.