关键词: 异质性, 非期望超效率SBM－DEA 模型, 物流效率, 物流利用无效率分解, 空间自相关性

Abstract: To study the difference in logistics efficiency between different regions，this paper built a comprehensive evaluation index system for logistics efficiency based on the heterogeneity of production technology in different provinces. Using the meta frontier undesired output super-efficiency SBM-DEA model to measure logistics efficiency under resource and environmental constraints from 2006 to 2017 in China. The measurement results were analyzed based on the inefficiency of logistics utilization decomposition and spatial correlation in the eastern，central，and western regions in China. The results showed that the logistics efficiency values under the meta frontier and the constraints of resources and environment were 0. 801，0. 474 and 0. 352 in the eastern，central and western regions respectively， and there was great potential for improvement among the regions. The technology gap rate in the central and western regions was low，and the logistics efficiency of related provinces was underestimated. There were different degrees of management inefficiency in the eastern， central and western regions in China. Among them， the inefficiency of logistics in the central and western regions was mainly due to technical inefficiency. The groups under the two different frontiers showed a trend of transition from ″strong agglomeration″ to ″weak agglomeration″，which ran counter to the development direction of my country′s current development of a highly concentrated regional central city. Relevant departments need to further improve the management level to reduce the loss of logistics efficiency；improve the level of logistics technology to reduce energy consumption in the logistics industry；promote the construction of regional transportation hubs to improve the development foundation of the logistics industry； innovate regional coordinated development strategies to accelerate the logistics industry upgrade.

Key words: heterogeneity, the undesired super-efficiency SBM-DEA model, logistics efficiency, logistics utilization inefficiency decomposition, space autocorrelation