Analysis of urban land use efficiency in China based on endogenous directional distance function model
Received date: 2017-01-18
Request revised date: 2017-05-27
Online published: 2017-07-31
Copyright
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 resulted from technical efficiency, and not operation environments. The land use efficiency 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.
WANG Jianlin , ZHAO Jiajia , SONG Malin . Analysis of urban land use efficiency in China based on endogenous directional distance function model[J]. GEOGRAPHICAL RESEARCH, 2017 , 36(7) : 1386 -1398 . DOI: 10.11821/dlyj201707016
Tab. 1 Descriptive statistics表1 描述性统计 |
| 变量 | 样本数 | 均值 | 标准差 | 最小值 | 最大值 | |
|---|---|---|---|---|---|---|
| 超大特大城市 | Labor | 11 | 28.83 | 20.12 | 11.71 | 68.89 |
| Capital | 11 | 55.19 | 17.93 | 30.69 | 99.20 | |
| Land | 11 | 76.15 | 33.60 | 41.30 | 138.6 | |
| GDP | 11 | 11.42 | 6.618 | 3.703 | 23.29 | |
| Green | 11 | 34.11 | 19.88 | 16.59 | 83.73 | |
| Ⅰ型大城市 | Labor | 12 | 13.02 | 10.53 | 5.411 | 44.74 |
| Capital | 12 | 28.72 | 6.698 | 19.32 | 45.11 | |
| Land | 12 | 39.50 | 18.60 | 15.80 | 89 | |
| GDP | 12 | 5.395 | 3.689 | 2.735 | 16.00 | |
| Green | 12 | 16.56 | 8.594 | 6.285 | 40.12 | |
| Ⅱ型大城市 | Labor | 116 | 3.321 | 3.024 | 0.468 | 23.78 |
| Capital | 116 | 8.405 | 6.340 | 1.766 | 35.23 | |
| Land | 116 | 14.96 | 11.28 | 2.400 | 92.20 | |
| GDP | 116 | 1.225 | 1.034 | 0.159 | 5.881 | |
| Green | 116 | 6.056 | 4.938 | 0.689 | 41.73 | |
| 中等城市 | Labor | 98 | 1.277 | 0.624 | 0.208 | 3.324 |
| Capital | 98 | 3.730 | 3.077 | 0.518 | 21.31 | |
| Land | 98 | 6.782 | 2.696 | 1.400 | 17.10 | |
| GDP | 98 | 0.425 | 0.272 | 0.0849 | 2.037 | |
| Green | 98 | 2.631 | 1.332 | 0.0440 | 9.566 | |
| 小城市 | Labor | 46 | 0.739 | 0.368 | 0.239 | 1.750 |
| Capital | 46 | 2.194 | 1.247 | 0.439 | 6.576 | |
| Land | 46 | 4.926 | 2.471 | 1.800 | 11.40 | |
| GDP | 46 | 0.261 | 0.191 | 0.0290 | 0.931 | |
| Green | 46 | 1.943 | 1.167 | 0.586 | 5.715 |
Tab. 2 Descriptive statistics of efficiencies on classifying cities into five groups表2 将城市划分为5类下的共同效率描述性统计 |
| 样本数 | 均值 | 标准差 | 最小值 | 最大值 | 变异系数 | 前沿数 | |
|---|---|---|---|---|---|---|---|
| 所有城市 | 283 | 0.534 | 0.166 | 0.022 | 1 | 0.311 | 12 |
| 超大特大城市 | 11 | 0.722 | 0.186 | 0.509 | 1 | 0.258 | 2 |
| Ⅰ型大城市 | 12 | 0.672 | 0.175 | 0.456 | 1 | 0.260 | 2 |
| Ⅱ型大城市 | 116 | 0.542 | 0.146 | 0.088 | 1 | 0.269 | 3 |
| 中等城市 | 98 | 0.502 | 0.159 | 0.022 | 1 | 0.317 | 3 |
| 小城市 | 46 | 0.500 | 0.178 | 0.136 | 1 | 0.356 | 2 |
Fig. 1 Comparison between Qingyuan and similar cities图1 清远市与类似规模城市的比较 |
Tab. 3 Kolmogorov-Smirnov test表3 Kolmogorov - Smirnov检验 |
| KS统计量 | P水平 | 结果 | |
|---|---|---|---|
| 划分为5类 | |||
| 超大特大城市vs Ⅰ型大城市 | 0.212 | 0.929 | 不能拒绝 |
| 超大特大城市 vs Ⅱ型大城市 | 0.495 | 0.009 | 拒绝 |
| 超大特大城市vs中等城市 | 0.622 | 4.332e-04 | 拒绝 |
| 超大特大城市 vs小城市 | 0.652 | 4.644e-04 | 拒绝 |
| Ⅰ型大城市vs Ⅱ型大城市 | 0.532 | 0.002 | 拒绝 |
| Ⅰ型大城市vs 中等城市 | 0.588 | 5.976e-04 | 拒绝 |
| Ⅰ型大城市vs 小城市 | 0.616 | 6.924e-04 | 拒绝 |
| Ⅱ型大城市vs中等城市 | 0.228 | 0.007 | 拒绝 |
| Ⅱ型大城市vs小城市 | 0.341 | 6.447e-04 | 拒绝 |
| 中等城市vs小城市 | 0.167 | 0.316 | 不能拒绝 |
| 划分为3类 | |||
| A类城市 vs B类城市 | 0.490 | 1.073e-04 | 拒绝 |
| A类城市 vs C类城市 | 0.590 | 8.112e-07 | 拒绝 |
| B类城市 vs C类城市 | 0.239 | 0.001 | 拒绝 |
注:原假设为两个样本来自同一分布。 |
Tab. 4 Descriptive statistics of meta-efficiency on classifying cites into three groups表4 将城市划分为3类下的共同效率描述性统计 |
| 样本数 | 均值 | 标准差 | 最小值 | 最大值 | 变异系数 | 前沿数 | |
|---|---|---|---|---|---|---|---|
| 所有城市 | 283 | 0.534 | 0.166 | 0.022 | 1 | 0.311 | 12 |
| A类城市 | 23 | 0.696 | 0.178 | 0.456 | 1 | 0.256 | 4 |
| B类城市 | 116 | 0.542 | 0.146 | 0.088 | 1 | 0.269 | 3 |
| C类城市 | 144 | 0.501 | 0.165 | 0.0220 | 1 | 0.329 | 5 |
Tab. 5 Descriptive statistics of group efficiencies on classifying cities into three groups表5 将城市划分为3类下的组群效率描述性统计 |
| 样本数 | 均值 | 标准差 | 最小值 | 最大值 | 变异系数 | 前沿数 | |
|---|---|---|---|---|---|---|---|
| 所有城市 | 283 | 0.654 | 0.197 | 0.0437 | 1 | 0.301 | 29 |
| A类城市 | 23 | 0.878 | 0.113 | 0.646 | 1 | 0.129 | 8 |
| B类城市 | 116 | 0.741 | 0.147 | 0.121 | 1 | 0.198 | 12 |
| C类城市 | 144 | 0.548 | 0.178 | 0.0437 | 1 | 0.325 | 9 |
Tab. 6 Descriptive statistics of meta-technical ratio on classifying cities into three groups表6 将城市划分为3类下的共同技术率描述性统计 |
| 样本数 | 均值 | 标准差 | 最小值 | 最大值 | 变异系数 | |
|---|---|---|---|---|---|---|
| 所有城市 | 283 | 0.826 | 0.123 | 0.452 | 1 | 0.149 |
| A类城市 | 23 | 0.785 | 0.126 | 0.607 | 1 | 0.161 |
| B类城市 | 116 | 0.728 | 0.0863 | 0.469 | 1 | 0.119 |
| C类城市 | 144 | 0.913 | 0.0758 | 0.452 | 1 | 0.083 |
Fig. 2 Histogram of operation environment potential图2 运作环境潜力的直方图 |
Fig. 3 Cities whose operation environmental potential exceeds 0.3图3 运作环境潜力大于0.3的城市 |
Tab. 7 Land excess and output short of C-style cities (small and medium-sized cities)表7 C类城市(中小城市)的土地冗余与产出不足 |
| 城市 | 土地冗余 | 产出不足 | 城市 | 土地冗余 | 产出不足 | 城市 | 土地冗余 | 产出不足 | ||
|---|---|---|---|---|---|---|---|---|---|---|
| 秦皇岛 | 0 | 0 | 舟山 | 1.618 | 0 | 阳江 | 2.426 | 0 | ||
| 邢台 | 5.365 | 0 | 丽水 | 1.082 | 0 | 云浮 | 0 | 0 | ||
| 张家口 | 4.077 | 0 | 马鞍山 | 2.899 | 0 | 桂林 | 3.121 | 0 | ||
| 承德 | 6.310 | 0.089 | 铜陵 | 2.309 | 0 | 梧州 | 1.972 | 0 | ||
| 沧州 | 2.706 | 0 | 安庆 | 4.238 | 0 | 北海 | 1.117 | 0 | ||
| 阳泉 | 2.504 | 0 | 黄山 | 3.024 | 0 | 防城港 | 0.000 | 0 | ||
| 吕梁 | 1.370 | 0.003 | 宁德 | 1.011 | 0 | 眉山 | 2.024 | 0 | ||
| 乌海 | 2.054 | 0 | 景德镇 | 3.300 | 0 | 雅安 | 1.602 | 0 | ||
| 通辽 | 0 | 0 | 萍乡 | 1.725 | 0 | 六盘水 | 1.903 | 0 | ||
| 鄂尔多斯 | 0 | 0 | 九江 | 2.056 | 0 | 遵义 | 4.696 | 0 | ||
| 呼伦贝尔 | 9.460 | 0 | 新余 | 0.612 | 0 | 安顺 | 2.783 | 0 | ||
| 巴彦淖尔 | 2.997 | 0 | 鹰潭 | 1.577 | 0 | 铜仁 | 2.834 | 0 | ||
| 乌兰察布 | 0 | 0 | 吉安 | 2.658 | 0 | 曲靖 | 3.102 | 0 | ||
| 本溪 | 3.731 | 0 | 上饶 | 2.499 | 0.152 | 玉溪 | 0.000 | 0 | ||
| 丹东 | 2.849 | 0 | 东营 | 0 | 0 | 保山 | 1.875 | 0 | ||
| 盘锦 | 1.590 | 0 | 许昌 | 5.171 | 0.011 | 铜川 | 1.852 | 0 | ||
| 铁岭 | 2.575 | 0.018 | 三门峡 | 1.316 | 0 | 咸阳 | 4.018 | 0 | ||
| 朝阳 | 3.728 | 0 | 周口 | 3.487 | 0.033 | 渭南 | 4.869 | 0 | ||
| 鸡西 | 4.521 | 0.032 | 湘潭 | 2.184 | 0 | 张掖 | 4.894 | 0 | ||
| 鹤岗 | 2.680 | 0 | 衡阳 | 9.462 | 0 | 平凉 | 2.183 | 0 | ||
| 双鸭山 | 0 | 0 | 邵阳 | 3.101 | 0 | 酒泉 | 2.770 | 0 | ||
| 伊春 | 16.353 | 0 | 张家界 | 1.518 | 0 | 庆阳 | 1.203 | 0 | ||
| 黑河 | 1.269 | 0.016 | 韶关 | 4.243 | 0 | 石嘴山 | 4.602 | 0 |
注:限于篇幅,本文仅展示部分城市的结果,全部城市的结果可向作者邮件索取。 |
The authors have declared that no competing interests exist.
| [1] |
|
| [2] |
[
|
| [3] |
[
|
| [4] |
[
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
[
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
|
| [19] |
|
| [20] |
|
| [21] |
|
| [22] |
|
| [23] |
Lee C Y. Johnson A.
|
| [24] |
|
| [25] |
|
| [26] |
|
| [27] |
|
| [28] |
|
| [29] |
|
| [30] |
|
| [31] |
|
| [32] |
|
| [33] |
[
|
| [34] |
[
|
/
| 〈 |
|
〉 |