技术关联性、复杂性与区域多样化——来自中国地级市的证据
|
马双(1990- ),男,浙江江山人,博士,助理研究员,主要研究方向为区域发展与创新地理。E-mail: ms@sass.org.cn |
收稿日期: 2019-04-08
要求修回日期: 2019-08-02
网络出版日期: 2020-06-24
基金资助
国家自然科学基金项目(41771143)
上海市哲学社会科学规划青年课题(2018EJL002)
上海市软科学研究领域重点项目(19692107400)
版权
Technological relatedness, complexity and regional diversity: Evidence from Chinese cities
Received date: 2019-04-08
Request revised date: 2019-08-02
Online published: 2020-06-24
Copyright
忽视现实基础,盲目追求新兴产业的现象普遍存在于当前中国的一些城市。基于国家知识产权局的专利数据,描绘1987—2016年中国地级市技术关联性和复杂性的时空演化特征,探讨两者对区域技术进入和增长的影响。结果表明:① 新技术的本地关联度越高,就越有可能进入该地区;现有技术的本地关联度越低,就越容易退出该地区;② 中国城市的技术发展总体上呈现路径依赖特点。其中,东部地区的技术关联性和复杂性变化较大,中西部地区变化较小;③ 技术关联性对新技术的进入和增长起到显著的促进作用。在引入复杂性较高的技术时,只有存在较高关联性才会促进区域的技术进步。最后提出的区域多样化发展框架,可为决策者权衡选择技术发展道路提供参考。
马双 , 曾刚 , 张翼鸥 . 技术关联性、复杂性与区域多样化——来自中国地级市的证据[J]. 地理研究, 2020 , 39(4) : 865 -879 . DOI: 10.11821/dlyj020190242
Ignoring the realistic basis and blindly pursuing new industries have become a common phenomenon in Chinese cities. Scholars have generally called for the implementation of regional capacity-based development policies, taking into account regional realities and industrial base, to avoid "one size fits all", repeated construction and vicious competition. Based on the patent data of the State Intellectual Property Office (SIPO), this paper describes the spatial-temporal evolution characteristics of technology relatedness and complexity of Chinese cities from 1987 to 2016, and estimates the impact of technology relatedness and complexity on regional technology introduction and technology growth. The results show that: (1) The dynamic of technology entry and exit in the eastern region is stronger than that of the central and western regions. The correlation analysis between relatedness and technology entry and exit shows that technology with high relatedness is more likely to enter the region. In contrast, technologies are more likely to exit if they do not have strong local technology linkages. (2) From the view of technological relatedness and complexity, technological evolution in eastern coastal areas is more dependent on the technological relatedness, and technological complexity has experienced tremendous growth, while technological relatedness and complexity in the central and western regions have changed little, and technological development in Chinese cities generally presents path-dependent characteristics. (3) Technological relatedness has been playing a significant role in promoting the entry and growth of new technologies. When technology with high complexity is introduced, if it has a high relatedness with the existing technological foundation of the region, it will play a significant role in promoting regional technological progress; on the contrary, even if the blind introduction of complex technology will not have a significant impact on regional technological progress. Based on the relatedness and complexity, this paper divides the technology development paths into four types: The policy of "bright road" means low risk and high benefit, when technologies are expected to exceed average returns under relatively low risk. In contrast, technological development with high risk and low benefit are unlikely to catch and raise value, we therefore refer to such a policy as "dead end". The "road exploration" policy means high risk and high benefit, which aims to develop new and original technology to realize path creation. "Slow Road" policy means driving safely forward in the existing low-level technology path. The framework of regional diversification development can provide a reference for decision makers to weigh and choose the path of technological development.
Key words: regional diversification; technological relatedness; complexity; cities; China
表1 2007—2016年中国各地区技术进入和退出情况Tab. 1 Technology entry and exit in China during 2007-2016 |
| 年份 | 技术进入 | 东部 | 中部 | 西部 | 技术退出 | 东部 | 中部 | 西部 |
|---|---|---|---|---|---|---|---|---|
| 2007 | 13.71 | 18.72 | 16.57 | 7.56 | 8.80 | 12.31 | 10.10 | 5.05 |
| 2008 | 8.36 | 13.61 | 8.12 | 4.49 | 11.18 | 11.82 | 13.90 | 8.44 |
| 2009 | 15.50 | 21.93 | 16.28 | 9.93 | 10.07 | 16.22 | 10.55 | 4.95 |
| 2010 | 10.93 | 16.49 | 12.07 | 5.73 | 8.43 | 12.19 | 9.49 | 4.68 |
| 2011 | 17.30 | 25.30 | 19.98 | 8.99 | 12.71 | 16.22 | 16.35 | 7.08 |
| 2012 | 25.96 | 40.16 | 27.42 | 13.85 | 14.71 | 20.63 | 18.32 | 7.24 |
| 2013 | 11.62 | 14.12 | 15.74 | 6.39 | 8.65 | 10.84 | 10.92 | 5.14 |
| 2014 | 11.83 | 15.34 | 14.62 | 6.89 | 5.40 | 6.06 | 6.50 | 3.53 |
| 2015 | 11.97 | 14.09 | 15.26 | 7.70 | 4.97 | 6.02 | 5.91 | 3.40 |
| 2016 | 18.22 | 21.82 | 24.94 | 10.05 | 16.64 | 31.23 | 12.99 | 8.31 |
表2 各变量的描述性统计Tab. 2 Descriptive statistics of variables |
| N | Mean | St.dev. | Min | Max | |
|---|---|---|---|---|---|
| Entry | 807 923 | 0.14 | 0.35 | 0 | 1 |
| Growth | 1 126 547 | 34.99 | 236.61 | -100 | 1000 |
| Relatedness_Density | 980 741 | 22.16 | 18.38 | 0 | 100 |
| Technology_Complexity | 987 825 | 36.78 | 24.61 | 0 | 100 |
| Population | 1 148 735 | 501.25 | 346.39 | 0.04 | 3 371.84 |
| GDP per capita | 1 148 735 | 8.58 | 3.87 | 2.79 | 23.46 |
| Population Density | 1 148 735 | 471.03 | 607.88 | 0.67 | 10 268 |
| Technological stock | 1 229 840 | 238 513 | 6388 | 467 | 875 137 |
| Technological size | 1 229 840 | 238 761 | 3189 | 327 | 741 265 |
数据来源:1988—2017年《中国城市统计年鉴》;国家知识产权局网站。 |
表3 1997—2016年全样本的技术进入模型Tab. 3 Entry models of full sample during 1997-2016 |
| 模型I | 模型II | 模型III | 模型IV | 模型V | |
|---|---|---|---|---|---|
| Constant | 0.153 079 7*** | 0.153 187 6*** | 0.151 287 5*** | 0.162 842 0*** | -0.012 034 7 |
| (0.000 523 7) | (0.000 553 9) | (0.000 517 9) | (0.000 564 1) | (0.025 665 3) | |
| Relatedness Density | 0.005 249 7*** | 0.004 429 4*** | 0.004 078 3*** | 0.003 876 6*** | |
| (0.000 037 2) | (0.000 037 5) | (0.000 040 2) | (0.000 049 9) | ||
| Technology Complexity | 0.000 046 3* | 0.000 035 4 | -0.000 060 8** | ||
| (0.000 017 8) | (0.000 019 8) | (0.000 022 1) | |||
| Population | 0.031 575 2*** | 0.016 234 9*** | -0.115 539 8*** | ||
| (0.000 798 6) | (0.000 798 3) | (0.014 987 2) | |||
| GDP per cap. | 0.000 002 0*** | 0.000 000 5*** | 0.000 001 7*** | ||
| (0.000 000 1) | (0.000 000 1) | (0.000 000 3) | |||
| Population Density | -0.000 000 86*** | -0.000 003 0*** | 0.000 018 6 | ||
| (0.000 006 5) | (0.000 006 7) | (0.000 011 8) | |||
| Tech. stock | -0.000 002 5*** | -0.000 003 1*** | -0.000 002 3*** | ||
| (0.000 000 1) | (0.000 000 1) | (0.000 000 2) | |||
| Tech. size | 0.000 000 4** | 0.000 000 05 | 0.000 001 3*** | ||
| (0.000 000 2) | (0.000 000 2) | (0.000 000 2) | |||
| 区域固定效应 | 否 | 否 | 否 | 否 | 是 |
| 时间固定效应 | 否 | 否 | 否 | 否 | 是 |
| 调整R2 | 0.030 309 2 | 0.030 316 2 | 0.0 397 627 | 0.030 675 8 | 0.036 628 1 |
注:如果区域r在对应的10年时间窗口期间新增某项技术的相对技术优势,则Entry为1,否则为0。所有自变量都是平均数且滞后一年。显著性水平:*P<0.05,**P<0.01,***P<0.001,括号内为标准误。 |
表4 1997—2016年不同关联度下的技术进入模型Tab. 4 Entry models by level of technology relatedness during 1997-2016 |
| 高关联度 | 低关联度 | 高关联度 | 低关联度 | 高关联度 | 低关联度 | |
|---|---|---|---|---|---|---|
| 模型I | 模型II | 模型III | 模型IV | 模型V | 模型VI | |
| Constant | 0.363 782 9*** | 0.030 949 3*** | 0.361 429 8*** | 0.040 398 1*** | 0.229 875 1 | 0.090 400 2** |
| (0.002 457 6) | (0.000 638 9) | (0.002 666 7) | (0.000 892 7) | (0.183 652 0) | (0.032 772 3) | |
| Technology Complexity | 0.000 472 6*** | -0.000 037 2 | 0.000 267 1* | -0.000 005 9 | 0.000 243 6* | -0.000 036 1 |
| (0.000 100 6) | (0.000 026 3) | (0.000 127) | (0.000 037 6) | (0.000 121 4) | (0.000 042 0) | |
| Population | 0.043 298 7*** | 0.021 379 7*** | -0.065 751 6 | 0.048 836 2* | ||
| (0.004 487 6) | (0.001 398 70) | (0.093 571 6) | (0.020 876 2) | |||
| GDP per cap. | 0.000 000 4 | 0.000 001 5*** | 0.000 001 6 | 0.000 000 2 | ||
| (0.000 000 4) | (0.000 000 1) | (0.000 001 6) | (0.000 000 5) | |||
| Population Density | 0.000 001 6 | -0.000 005 7*** | 0.000 025 2 | -0.000 020 2 | ||
| (0.000 003 4) | (0.000 001 5) | (0.000 056 9) | (0.000 028 1) | |||
| Tech. stock | -0.000 002 6*** | 0.000 000 2*** | -0.000 003 6*** | 0.000 000 3 | ||
| (0.000 000 4) | (0.000 000 2) | (0.000 000 7) | (0.000 000 4) | |||
| Tech. size | 0.000 008 7** | 0.000 002 1** | 0.000 013 9*** | 0.000 001 8* | ||
| (0.000 001 2) | (0.000 000 7) | (0.000 001 2) | (0.000 000 7) | |||
| 区域固定效应 | 否 | 否 | 否 | 否 | 是 | 是 |
| 时间固定效应 | 否 | 否 | 否 | 否 | 是 | 是 |
| 调整R2 | 0.000 501 9 | 0.000 014 5 | 0.005 209 1 | 0.012 607 8 | 0.052 028 4 | 0.030 293 2 |
注:高关联度模型仅包括前10%的区域观测数据,低关联度模型仅包括后10%的区域观测数据;显著性水平:*P<0.05,**P<0.01,***P<0.001,括号内为标准误。 |
表5 1997—2016年全样本的技术增长模型Tab. 5 Growth models of full sample during 1997-2016 |
| 模型I | 模型II | 模型III | 模型IV | 模型V | |
|---|---|---|---|---|---|
| Constant | 14.023 762 2*** | 13.698 542 2*** | 13.338 297 3*** | 13.610 300*** | 73.798 271 8*** |
| (0.169 873 6) | (0.163 629 0) | (0.169 764 3) | (0.175 274 93) | (7.422 473 6) | |
| Relatedness Density | 0.473 682 9*** | 0.457 389 28*** | 0.349 039 2*** | 0.198 273 6*** | |
| (0.010 019 2) | (0.010 087 4) | (0.008 252 1) | (0.012 093 0) | ||
| Technology Complexity | 0.207 529 3*** | 0.180 982 3*** | 0.125 472 63*** | ||
| (0.007 765 3) | (0.008 192 8) | (0.008 023 1) | |||
| Population | 15.709 382*** | 13.789 283 7*** | 56.982 910 2*** | ||
| (0.301 938 2) | (0.308 124 2) | (4.512 839 2) | |||
| GDP per cap. | 0.000 452 8*** | 0.000 342 7*** | 0.000 006 0 | ||
| (0.000 020 9) | (0.000 019 8) | (0.000 090 2) | |||
| Population Density | -0.003 873 6*** | -0.003 401 8*** | -0.003 372 9 | ||
| (0.000 213 2) | (0.000 231 5) | (0.003 300 1) | |||
| Tech. stock | -0.000 502 93*** | -0.000 558*** | -0.005 098 2*** | ||
| (0.000 033 0) | (0.000 031 9) | (0.000 098 3) | |||
| Tech. size | 0.001 076 0*** | 0.000 714 2*** | 0.001 726 32*** | ||
| (0.000 056 0) | (0.000 054 82) | (0.000 060 3) | |||
| 区域固定效应 | 否 | 否 | 否 | 否 | 是 |
| 时间固定效应 | 否 | 否 | 否 | 否 | 是 |
| 调整R2 | 0.003 969 0 | 0.00560981 | 0.00702039 | 0.0102983 | 0.0672812 |
注:因变量为技术增长,即技术i在区域r中从周期t到周期t+1的专利数量的增长率;显著性水平:*P<0.05,**P<0.01,***P<0.001,括号内为标准误, |
表6 1997—2016年不同关联度下的技术增长模型Tab. 6 Growth models by level of technology relatedness during 1997-2016 |
| 高关联度 | 低关联度 | 高关联度 | 低关联度 | 高关联度 | 低关联度 | |
|---|---|---|---|---|---|---|
| 模型I | 模型II | 模型III | 模型IV | 模型V | 模型VI | |
| Constant | 54.092 830 2*** | -5.872 819 2*** | 50.378 291 2*** | -7.123 161 4*** | -29.746 290 5 | -9.172 839 2 |
| (0.759 829 38) | (0.109 283 72) | (0.784 958 1) | (0.199 048 3) | (57.287 292 5) | (6.172 635 1) | |
| Technology Complexity | 0.312 837 20*** | -0.017 625 30 | 0.259 215 1*** | -0.010 800 1 | 0.219 847 29*** | -0.029 174 83* |
| (0.034 001 8) | (0.006 982 4) | (0.036 112 3) | (0.010 407 5) | (0.034 488 0) | (0.010 118 7) | |
| Population | 27.267 418 2*** | -3.418 293 0*** | -27.102 930 2 | -2.709 583 9 | ||
| (1.501 928 4) | (0.320 003 2) | (30.126 487 5) | (3.980 945 5) | |||
| GDP per cap. | 0.000 110 76 | -0.000 358 8*** | 0.003 800 1*** | 0.000 130 0 | ||
| (0.000 107 3) | (0.000 023 1) | (0.000 498 5) | (0.000 121 0) | |||
| Population Density | -0.0050001*** | 0.000 658 7* | -0.074 011 0*** | -0.005 102 93 | ||
| (0.0009090) | (0.000 287 4) | (0.013 989 8) | (0.004 766 7) | |||
| Tech. stock | -0.0020102*** | -0.000 201 1* | -0.008 099 3*** | -0.001 500 1*** | ||
| (0.0001001) | (0.000 089 7) | (0.000 299 7) | (0.000 267 3) | |||
| Tech. size | -0.0012563*** | -0.001 299 8*** | 0.000 200 9 | -0.001 105 4*** | ||
| (0.0001260) | (0.000 270 9) | (0.000 117 3) | (0.000 270 1) | |||
| 区域固定效应 | 否 | 否 | 否 | 否 | 是 | 是 |
| 时间固定效应 | 否 | 否 | 否 | 否 | 是 | 是 |
| 调整R2 | 0.001 740 1 | 0.000 022 6 | 0.011 501 9 | 0.015 398 7 | 0.130 921 5 | 0.046 001 3 |
注:因变量为技术增长,即技术i在区域r中从周期t到周期t+1的专利数量的增长率;高关联度模型仅包括前10%的区域观测数据,低关联度模型仅包括后10%的区域观测数据;显著性水平:*P<0.05,**P<0.01,***P<0.001,括号内为标准误。 |
评审专家对本研究的模型优化、行文规范和文字表述方面提出客观、准确、详实的审稿意见,特致以诚挚感谢。
| [1] |
闻一言 . 新兴产业“浮躁症”背后是“政绩投机”. 学习时报, 2010-05-31(4).
[
|
| [2] |
萧函 . 光伏产业需谨防“大跃进”式建设. 太阳能, 2014,34(7):63-64.
[
|
| [3] |
陈建军 . 长江三角洲地区的产业同构及产业定位. 中国工业经济, 2004,20(2):19-26.
[
|
| [4] |
苏红键, 赵坚 . 相关多样化、不相关多样化与区域工业发展: 基于中国省级工业面板数据. 产业经济研究, 2012,92(2):26-32.
[
|
| [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] |
贺灿飞 . 区域产业发展演化: 路径依赖还是路径创造. 地理研究, 2018,37(7):1253-1267.
[
|
| [37] |
|
/
| 〈 |
|
〉 |