Abstract:From the perspective of carbon lock-in, this study constructs an input-output evaluation index system for carbon unlocking in the power industry and analyzes the carbon unlocking efficiency of the power industry based on panel data from 30 provinces, municipalities and autonomous regions in China from 2011 to 2020. By using the static Super-SBM model, it is found that the carbon unlocking efficiency of over 90% of provinces did not reach the frontier level, indicating that most provinces have difficulty in orderly unlocking while maintaining economic benefits. By using the global Malmquist index model for efficiency decomposition and dynamic analysis, it is concluded that there is a close relationship between carbon unlocking efficiency and technological development level, and the lower carbon unlocking level is mainly related to the stagnation of technological efficiency development. Finally, by using the BP neural network model to fit the relationship between TFP index and the rate of thermal power generation exit, data reference is provided for the decision-making of low-carbon transformation path of thermal power generation.