Determinants of the urban spatial network in China: An analysis through the lens of corporate networks within electronic information industry
Received date: 2017-11-01
Request revised date: 2019-02-25
Online published: 2019-05-13
Copyright
The identification of factors underlying the spatial structure of urban network and the analysis of its mechanism is the key to establishing theoretical models of urban network. Based on the ownership linkage model, the urban network in China is specified through the lens of top 100 corporate networks within the electronic information industry in the years of 2005 and 2017, and its structural characteristics are described from three aspects of centrality, linkages and triad census. Then, by using the exponential random graph models (ERGMs), an econometric analysis is conducted to identify the influencing factors, and the micro processes in the spatial growth of urban network are examined. Finally, by combining theories of resource dependence and transaction cost, a conceptual framework for comprehensively understanding the mechanisms driving urban network growth in China is suggested for further discussion. Three main findings are concluded. First, the preference attachment effect and the receiver (GDP) effect constitute the micro basis of centrality pattern of the urban network in China. Outdegree centrality is mainly affected by expansionary effect, and indegree centrality is affected by both convergent effect and receiver (GDP) effect, which has caused urban network growth to be a process of preferential selection. The economic mechanism of this process can be interpreted as the dependence of enterprise on specific assets, market thickness and other specific resources. Second, the influencing factors of city linkages tend to be diversified. While reciprocity remains the important mechanism in the linkages of urban network, network closure mechanism has gradually become an important factor in the relationship between cities. The reciprocity effect and the network closure mechanism constitute the micro basis of coherent subgroups in the urban network in China, which can be explained by the restrictive function of transaction cost. Third, geographical distance does not have a significant effect on the urban structure within the network of electronic information industry. The location of electronic information industry is more flexible, which makes the urban network grow in a “space of flows”. Cities between long distance established linkages from the first beginning of urban network development, promoting the expansion of urban network over a large spatial scale.
SHENG Kerong , ZHANG Hongxia , ZHAO Chaoyue . Determinants of the urban spatial network in China: An analysis through the lens of corporate networks within electronic information industry[J]. GEOGRAPHICAL RESEARCH, 2019 , 38(5) : 1030 -1044 . DOI: 10.11821/dlyj020171009
Tab. 1 The top 10 cities of degree centrality and growth rates of the urban network in China (%)表1 中国城市网络度中心性及增长前10位城市 |
| 序号 | 2005年 | 2017年 | 2005—2017年出度变化 | 2005—2017年入度变化 | |||
|---|---|---|---|---|---|---|---|
| 出度 | 入度 | 出度 | 入度 | ||||
| 1 | 北京(25.15) | 北京(15.27) | 北京(19.62) | 上海(11.11) | 深圳(20.46) | 上海(9.59) | |
| 2 | 深圳(9.28) | 上海(14.67) | 深圳(17.11) | 北京(10.13) | 北京(17.26) | 北京(7.92) | |
| 3 | 杭州(9.27) | 深圳(11.37) | 杭州(5.56) | 深圳(7.52) | 福州(5.75) | 深圳(5.89) | |
| 4 | 青岛(8.08) | 长沙(2.99) | 沈阳(4.48) | 南京(3.49) | 沈阳(4.60) | 南京(3.96) | |
| 5 | 许昌(4.79) | 西安(2.69) | 福州(4.39) | 苏州(3.05) | 惠州(4.60) | 苏州(3.45) | |
| 6 | 沈阳(4.19) | 大连(2.68) | 惠州(3.94) | 成都(3.05) | 上海(4.21) | 成都(3.24) | |
| 7 | 苏州(3.58) | 杭州(2.39) | 上海(3.76) | 广州(2.78) | 南京(3.09) | 广州(3.07) | |
| 8 | 厦门(2.99) | 沈阳(2.39) | 苏州(3.76) | 杭州(2.59) | 杭州(3.96) | 南宁(2.94) | |
| 9 | 上海(2.69) | 南京(2.39) | 南京(3.49) | 西安(2.41) | 石家庄(3.83) | 重庆(2.81) | |
| 10 | 绵阳(2.38) | 武汉(2.39) | 青岛(2.86) | 武汉(2.41) | 苏州(3.83) | 杭州(2.68) | |
注:出度(入度)后面括号中的数字为城市出度(入度)统计值占整个城市网络出度(入度)统计值总和的比重,2005—2017年出度(入度)变化后面括号中的数字为城市出度(入度)增加值占城市网络出度(入度)增加值之和的比重。 |
Tab. 2 The number of triad motifs (个) |
| 三方组类型 | 2005年 | 2017年 | 三方组类型 | 2005年 | 2017年 |
|---|---|---|---|---|---|
| A, B, C | 497062 | 517033 | A→B←C, A→C | 23 | 186 |
| A→B, C | 17450 | 44429 | A←B←C, A→C | 0 | 14 |
| A↔B, C | 2521 | 4490 | A↔B↔C | 83 | 184 |
| A←B→C | 432 | 3286 | A←B→C, A↔C | 16 | 51 |
| A→B←C | 165 | 412 | A→B←C, A↔C | 22 | 183 |
| A→B→C | 323 | 1455 | A→B→C, A↔C | 14 | 48 |
| A↔B←C | 164 | 356 | A→B↔C, A↔C | 28 | 125 |
| A↔B→C | 355 | 1520 | A↔B↔C, A↔C | 7 | 28 |
注:→代表前者向后者发送关系;←代表前者接收后者发送关系;↔代表互惠性关系。 |
Tab. 3 Variables and hypotheses in ERGMs表3 ERGMs变量设定及其假设检验 |
![]() |
Tab. 4 Results of ERGMs analysis表4 ERGMs回归结果 |
| 参数[PNet中的代码] | 2005年 | 2017年 | |||||
|---|---|---|---|---|---|---|---|
| 估计值 | 标准误 | t-比率 | 估计值 | 标准误 | t-比率 | ||
| 弧[Arc] | -4.43* | 0.28 | -0.01 | -4.16* | 0.31 | 0.01 | |
| 互惠性[Reciprocity] | 2.63* | 0.68 | -0.04 | 1.58* | 0.20 | -0.07 | |
| 交替-出-星[AoutS] | 1.33* | 0.16 | 0.07 | 1.67* | 0.15 | -0.03 | |
| 交替-入-星[AinS] | 0.83* | 0.18 | -0.02 | 0.63* | 0.18 | 0.02 | |
| 循环三方组[T10] | 0.68 | 0.40 | -0.08 | 0.54* | 0.23 | 0.01 | |
| 发送者(GDP)[sender] | 1.11 | 0.69 | -0.02 | 0.22 | 0.44 | -0.04 | |
| 接收者(GDP)[receiver] | 1.90* | 0.73 | -0.02 | 1.95* | 0.47 | 0.06 | |
| 趋异性(GDP)[difference] | -0.44 | 0.81 | -0.04 | -0.81 | 0.56 | 0.01 | |
| 空间距离[Covariate Arc] | -0.20 | 0.15 | -0.03 | -0.02 | 0.03 | -0.01 | |
注:(1)t-比率=(观察值-预测值)/标准差;(2)2005年模型的乘数因子(multiplication factor)为20,2017年的乘数因子为10;(3)“*”表示参数在统计上是显著的。 |
Tab. 5 Results of goodness of fit analysis表5 拟合优度(GOF)检验结果 |
| 参数[PNet代码] | 2005年 | 2017年 | |||||
|---|---|---|---|---|---|---|---|
| 观测值 | 预测值 | t-比率 | 观测值 | 预测值 | t-比率 | ||
| 弧[Arc] | 149.000 | 149.833 | 0.096 | 342.000 | 341.992 | -0.000 | |
| 互惠性[Reciprocity] | 23.000 | 22.748 | 0.148 | 49.000 | 49.077 | -0.036 | |
| 2-路径[2-path] | 1079.000 | 1015.896 | 2.083 | 3576.000 | 3541.574 | 0.873 | |
| 2-入-星[in-2-star] | 571.000 | 520.065 | 0.600 | 1685.000 | 1313.611 | 0.112 | |
| 2-出-星[out-2-star] | 691.000 | 684.714 | 0.740 | 3013.000 | 3089.823 | -0.449 | |
| 交替-入-星[AinS] | 178.953 | 177.798 | 0.060 | 486.270 | 485.872 | 0.013 | |
| 交替-出-星[AoutS] | 195.782 | 190.481 | 0.065 | 536.852 | 515.453 | 0.869 | |
| 传递三方组[T9] | 232.000 | 215.354 | 1.660 | 1181.000 | 1135.781 | 3.276 | |
| 循环三方组[T10] | 56.000 | 53.612 | 0.137 | 243.000 | 245.574 | 0.059 | |
| 循环闭合[AT-C] | 101.500 | 100.256 | 0.162 | 341.984 | 342.938 | -0.052 | |
| 发送者(GDP)[sender] | 17.530 | 17.170 | 0.086 | 40.750 | 40.289 | 0.038 | |
| 接收者(GDP)[receiver] | 15.410 | 14.882 | 0.149 | 29.300 | 28.079 | 0.270 | |
| 趋异性(GDP)[difference] | 19.040 | 16.624 | 0.104 | 36.510 | 36.385 | 0.024 | |
| 空间距离[Covariate Arc] | 231.000 | 227.504 | 0.085 | 548.000 | 546.337 | 0.024 | |
| 入度标准差 | 3.978 | 3.763 | 0.968 | 5.205 | 5.113 | 0.466 | |
| 入度分布偏度 | 3.097 | 3.015 | 2.690 | 2.748 | 2.504 | 0.874 | |
| 出度标准差 | 4.382 | 4.985 | 0.945 | 8.367 | 7.909 | 0.564 | |
| 出度分布偏度 | 3.309 | 3.202 | 0.847 | 2.348 | 2.873 | 0.737 | |
注:t-比率=(观察值-预测值)/标准差;在拟合度检验中的变量共56项,篇幅的关系表中仅列出14项。 |
Fig. 1 The linkage patterns of urban network in China图1 中国城市网络关联格局 |
Fig. 2 Path dependence of degree centrality (152 cities, valued network)图2 城市中心度的路径依赖(152个城市,多值网络) |
Fig. 3 The relationship between urban centrality and GDP in 2017 (152 cities excluding Beijing and Shenzhen, valued network)图3 2017年城市中心性与GDP的关系(152个城市,剔除北京和深圳,多值网络) |
Fig. 4 The relationship between the number of linkages and their geographical distance (152 cities, valued network)图4 城市间链接数量与空间距离的关系(152个城市,多值网络) |
Fig. 5 A conceptual framework for understanding the mechanisms of urban network growth图5 城市网络发育机理分析框架 |
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] |
|
| [24] |
|
| [25] |
|
| [26] |
|
| [27] |
|
| [28] |
|
| [29] |
|
| [30] |
|
| [31] |
|
| [32] |
|
| [33] |
[
|
| [34] |
|
| [35] |
|
| [36] |
|
| [37] |
|
| [38] |
|
/
| 〈 |
|
〉 |