Abstract:

In recent years,the arid and semiarid regions in northwest China,including the Xinjiang Uygur Autonomous Region,Inner Mongolia Autonomous Region,Ningxia Hui Autonomous Region,Shaanxi Province,Qinghai Province,and Gansu Province,have undergone rapid development.However,the fragile ecosystems in these areas are facing greater pressure and the conflict between economic development,resources,and the environment is becoming increasingly prominent due to the extensive economic development model that has not yet been fundamentally modified.Eco-efficiency is an effective tool for measuring the degree of coordination of economic,resource,and environmental benefits.In this study,we used the three-stage DEA method to measure eco-efficiency,and thus overcome the effects of external environmental variables and random factors on eco-efficiency to obtain more accurate and reliable eco-efficiency values.Meanwhile,we considered 38 cities in 6 provinces in the arid and semiarid regions of northwest China as research objects and used urban panel data from 2005 to 2016 to analyze the regional eco-efficiency.Therefore,the differences and spatial correlation characteristics of eco-efficiency levels among cities could be depicted and described in a more detailed manner.The results are described as follows: (1) After eliminating the effects of external environmental variables and random factors,the eco-efficiency level of each city changed significantly.The average technical efficiency dropped from 0.65 to 0.57,representing a decrease of 12.3%.The average pure technical efficiency increased from 0.91 to 0.95,showing an increase of 4.4%,and the average scale efficiency decreased from 0.72 to 0.59,indicating a decrease of 18.1%.These results indicate that external environmental variables and random factors have a great effect on eco-efficiency and that it is reasonable and necessary to use the three-stage DEA method to analyze eco-efficiency.(2) In 38 cities,those with medium efficiency and greater accounted for 39% of the total,including 15 cities such as Urumqi,Karamay,and Ordos,while those with low efficiency accounted for 61% of the total,including 23 cities such as Pingliang and Guyuan.These results show that the level of eco-efficiency in the arid and semiarid areas of northwest China is generally low and that there are significant urban differences and imbalances.(3) The global Moran’s I index was positive and increased yearly,thus indicating that the spatial distribution of eco-efficiency exhibits increasingly significant positive autocorrelations.The local Getis Ord Gi^* results show that the distribution areas of cold spots and hot spots of eco-efficiency were mostly the same as those of low efficiency,medium efficiency,and above.These results indicate that urban eco-efficiency has an external driving effect and cities with prominent geographical advantages and perfect ecological management levels have high eco-efficiency,which can also promote the improvement of eco-efficiency of neighboring cities and vice versa.This study compensates for the relative shortage of studies of eco-efficiency for cities in the arid and semiarid regions of northwest China.The accurate measurements of the value of eco-efficiency and detailed analyses of the differences and the spatial correlation characteristics of eco-efficiency of the cities in this region enhance the objective understanding of eco-efficiency in the arid and semiarid regions of northwest China,and thus provide new empirical evidence for general improvement in regional eco-efficiency.

Key words: arid and semiarid regions in northwest China, ecological efficiency, three-stage DEA, exploratory spatial data analysis