基于内生方向距离函数的中国城市土地利用效率分析

Analysis of urban land use efficiency in China based on endogenous directional distance function model

中国特有的土地财政制度激励着地方政府不断扩大城市建设用地。为了对城市的土地使用效率进行衡量与比较,在内生方向距离函数模型基础上,提出一种新的测度效率的模型,对中国283个地级以上城市的土地使用情况进行了测算。结果显示:土地使用效率在不同规模城市之间存在显著差异,大城市的土地效率显著高于中小城市。进一步meta-frontier分析显示,不同规模城市的土地使用效率差别主要来源于自身技术效率差异,而非城市运作环境的差异。基于上述发现,提出若干关于提升中国城市土地使用效率的政策建议。

The Chinese government＇s pursuit of expanding municipal areas results in the speed of land urbanization exceeding that of population urbanization, which not only affects urban land use efficiency, but also threatens China＇s grain security. In this study, we measured urban land use efficiency under the DEA framework, compared the gap among different groups of cities, and presented relevant policy suggestions. We proposed an endogenous directional distance function model for measuring efficiency, which has many advantages compared with existing models. First, our model can decrease inputs and increase outputs simultaneously, and the measure is closer to the general definition of urban land use efficiency. Most studies have employed classic input-oriented models to measure land use efficiency. Our new model can choose endogenous directional vectors according to slack values, which can avoid the directional distance function model＇s problem of overestimating efficiency. In addition, compared with previous models, our model makes better economic sense. Using the new model, the urban land use efficiencies of 283 prefecture-level cities were evaluated. The results showed the average efficiency of all cities was about 0.53, meaning China＇s overall urban land use level is low and has substantial potential to be exploited. Moreover, large efficiency gaps were found among cities of different scales, and the average land use efficiencies of large-scale cities were higher than those of small-scale cities. Through efficiency decomposition, we found that the gaps among cities of different scales operation environments. The land use efficiency resulted from technical efficiency, and not of the small and medium-sized cities with a population less than 1 million was the lowest. However, their group-frontier was very close to the meta-frontier of all the samples. Overall, the poor land use performance was mainly due to the low technical efficiency of the urban individual. Further analysis showed that the deeper reason was the existence of land over-input. Based on the analysis results above, we presented some suggestions including changing government functions, strengthening the construction of land use institution, and paying attention to the revamp of the stock of land.