Variable RTS in hierarchical network DEA: Enhancing efficiency in higher education systems

This study presents a novel approach to Network Data Envelopment Analysis (DEA) by introducing “Returns to Scale (RTS) separation” within a hierarchical network DEA framework. Traditional DEA models, which often assume constant RTS, face limitations when analysing complex multi-functional structures. The proposed method, Variable RTS in Hierarchical Network DEA (VRS-HNDEA), addresses these limitations by integrating variable RTS, enabling a detailed efficiency analysis across hierarchical systems with heterogeneous sub-units. By utilising free variables, this model establishes distinct efficiency planes for simultaneous benchmarking of diverse subsystems, yielding a global efficiency frontier through the Minkowski addition of sub-system sets and analysed using an input-oriented enveloped form. Applied specifically to the higher education sector, the VRS-HNDEA model provides insights into the operational efficiency of various academic functions, including teaching, research, and administration. Key findings from this application demonstrate the model's ability to capture efficiency variations across hierarchical levels, supporting nuanced decisions on resource allocation and scale optimization. This framework, with its capability to recognise scale diversity across sub-systems, offers a significant tool for enhancing efficiency assessment in multi-layered public sector contexts, such as higher education, where comprehensive resource management is crucial.