Abstract This study introduces a novel Two-Stage Total Partial Productivity Data Envelopment Analysis (2S-TPP-DEA) model to evaluate the efficiency of the Chinese banking sector. Traditional DEA and network DEA models, while useful for analyzing multi-stage processes, often oversimplify the relationships between inputs, intermediate outputs, and final outputs by aggregating data and assuming fixed functional relationships. Our proposed 2S-TPP-DEA model addresses these limitations by evaluating how effectively each input contributes to outputs at different stages of banking operations, rather than simply aggregating them. This approach offers a clearer, more realistic picture of efficiency—especially across banks that differ in size, structure, and market conditions. In addition, the model incorporates key contextual factors such as market structure, input prices, and financial distress, offering a more comprehensive and realistic understanding of efficiency in diverse banking environments. The findings highlight significant differences in efficiency across ownership structures, with state-owned banks benefiting from government support, while joint-stock banks show higher responsiveness to market forces. These insights provide valuable guidance for policymakers and banking managers in optimizing strategies to improve efficiency and sustain growth in the face of economic pressures. By filling a gap in existing methodologies, this research contributes to the ongoing discourse on efficiency measurement in complex financial systems.
Wänke et al. (Thu,) studied this question.