China's Provincial Eco-Efficiency and Its Driving Factors-Based on Network DEA and PLS-SEM Method

This study aims to estimate the eco-efficiencies of China at provincial levels. The eco-efficiencies of production and treatment stages are disentangled by the network data envelopment analysis (DEA) method. The key driving factors are identified by the integrative use of driving force-pressure-state-impact-response frame model (DPSIR) model and partial least squares structural equation modeling (PLS-SEM) method. This study provides several important findings. In general, the eco-efficiencies of most regions in China are inefficient and show significant regional differences. All DPSIR factors have significant and strong impacts on the eco-efficiency of the treatment stage. The eco-efficiency of the production stage evidently outweighs the eco-efficiency in economically well-developed regions. The originality of this study lies in three aspects. First, using two-stage network DEA, this study dissects the overall eco-efficiency into production efficiency and treatment efficiency. Empirical results provide insights into the root cause of the low efficiency of each province (municipality). Second, on the basis of the DPSIR model, an expanded pool of driving factors is investigated. Third, using the PLS-SEM method to analyze eco-efficiency is more reliable and effective than applying other traditional regression models.