Who cultivates the inter-firm networks in regional cluster? -Innovation and Social Networks- Ichiro Sakata

Size: px

Start display at page:

Download "Who cultivates the inter-firm networks in regional cluster? -Innovation and Social Networks- Ichiro Sakata"

Jason Dixon
5 years ago
Views:

1 OECD Workshop on Network approaches to Innovation December 8, 008 Who cultivates the inter-firm networks in regional cluster? -Innovation and Social Networks- Ichiro Sakata Innovation Policy Research Center (IPR-CTR) Graduate School of Engineering University of Tokyo Japan

2 Outline 1. Background. Aim & Scope 3. Data & Methods 4. Results 5. Discussions 6. Conclusion

3 1. Background Regional cluster as a recipe for industrial innovation in region Country Program Time Span Regions Major contents U.S.A. Clusters of Innovation 001 All regions Benchmark of regions Germany BioRegio 1995~ 3 regions Selective funding to the regions InnoRegio 1999~ 3 regions Selective funding to the regions Finland Science Park 198~ 14 regions Japan Industrial cluster 001~ 18 regions Intelligent cluster 00~ 1 regions Industry-Academy relationships Incubation Industry-Academy relationships Incubation Collaboration among organizations intellectual property

4 Definition of regional clusters Clusters are geographic concentrations of interconnected companies and institutions in a particular field. (M.E. Porter, HBR, 1998) agglomeration cluster weakly bounded tightly interconnected

5 Previous works 1 Network as a locus of innovation 1. Information and knowledge dissemination and sharing (ex. Podolny and Page, 1998; Dyer and Nobeoka, SMJ 000).. Resource allocation (ex. Uzzi, AMR 1996) 3. Cooperation (Nishiguchi and Beaudet, SMR 1998) 4. Robustness (Albertand Barabási, 00)

6 Previous works Accumulating evidence on the relationships between network and performance Centrality and firm growth Powell, et al., (ASQ 1996, AJS 005) Centrality and innovativeness Tsai (AMJ 001) Bell (SMJ 005) Soh and Roberts (RP 003, IEEE Eng Mgmt 005).

7 Limitations of previous works Small data set (case study approach) (~500 nodes) Lack of details in network structure (macroscopic) Lack of the networking mechanism Our focuses in this study Comparative study with a large data set (~50,000 nodes in total) From micro-, meso-, to macro-structures of networks Elucidation of networking mechanism

8 . Research Purpose 1. System development to monitor the development of regional clusters and analyze deeper structures of them from macro-, meso-, to micro-levels.. Evaluation of regional clusters in Japan from a network perspective. 3. Elucidation of networking mechanisms by a comparative study.

9 3. Data & Methods Scope Inter-firm networks (supplier-customer relationships) Node Link Network = Nodes + Links Organizational network Node=Organization (Firms, University, Public research institute) Link = Relationships (Customer-Supplier, Alliance, Managerial Board, etc.)

10 Why supplier-customer relationships? the best information channel for Japanese firms. A: not entering a new field. B: entering a new field. Source: Small and Medium Enterprise Agency (006)

18 clusters # Region Main industrial category A Osaka Manufacturing B Chukyo Manufacturing C Kinki Pharmaceutical & Medical D Hiroshima-Okayama Manufacturing E North-Kyushu Manufacturing F Hukuoka

11 18 clusters # Region Main industrial category A Osaka Manufacturing B Chukyo Manufacturing C Kinki Pharmaceutical & Medical D Hiroshima-Okayama Manufacturing E North-Kyushu Manufacturing F Hukuoka Environment G Hokkaido Pharmaceutical & Agriculture H Nagano Manufacturing I Niigata Manufacturing J Sapporo Pharmaceutical & Agriculture K Kyoto Manufacturing L Toyama Pharmaceutical M Sapporo Information Technology N Hamamatsu Manufacturing O Hukuoka Medical Device P Aomori Agriculture Q Yamagata Manufacturing R Okinawa Food

12 3. Data & Methods Methodological overview Data (from NTT & Teikoku DB) I Network construction (extraction of the maximum connected component) III Network Analysis IIIa Density IIIb Path length IIIc Clustering coefficient II Visualization of Networks IV Clustering the network (extraction of modules) VI Module Analysis VIa Inter-module coordination VIb Connector-Hub analysis V Visualization of modules

13 3. Data & Methods Measures of network structure Density (DEN) Closure of regional economy Def DEN = number of links per node Path length ratio (PLR) Efficiency of resource allocation Def PLR = average path length (random network)/ average path length (real network) Clustering coefficient ratio (CCR) Robustness of regional economy Def CCR = clustering coefficient (real network)/ average coefficient (random network) Inter-module coordination (IMC) Def IMC = 1/modularity Density of the connections among modules

14 4. Results #: Cluster ID, n: number of nodes, K: number of links # Region Main industrial category n K A Osaka Manufacturing 8,834 43,09 B Chukyo Manufacturing 7,914 34,16 C Kinki Pharmaceutical & Medical 5,437 5,310 D Hiroshima-Okayama Manufacturing 3,553 13,77 E North-Kyushu Manufacturing 3,75 13,40 F Hukuoka Environment 3,7 14,6 G Hokkaido Pharmaceutical & Agriculture,038 7,740 H Nagano Manufacturing 1,933 10,018 I Niigata Manufacturing 1,898 8,46 J Sapporo Pharmaceutical & Agriculture 1,871 6,086 K Kyoto Manufacturing 1,798 7,36 L Toyama Pharmaceutical 1,397 5,364 M Sapporo Information Technology 1,113 3,80 N Hamamatsu Manufacturing 1,049 4,080 O Hukuoka Medical Device 931,70 P Aomori Agriculture 673,164 Q Yamagata Manufacturing 65,078 R Okinawa Food 57,594

15 PRL CCR DEN IMC 4. Results 6 5 DEN: Density, IMC: Inter-module coordination PRL: Path length ratio, CCR: Clustering coefficient ratio #node #node #node #node

16 Summary on the size effect of regional network As cluster sizes increases, PRL and CCR strongly increases. Efficiency of resource allocation Robustness of regional economy DEN and IMC slightly increases. Closure of regional economy In small clusters, A large variance in network structures. Connection among modules There is a room improving regional networking by political support.

17 Okinawa Yamagata Aomori Hukuoka Hamamatsu Sapporo Toyama Kyoto Sapporo Niigata Nagano Hokkaido North-Kyushu Hukuoka Hiroshima-Okayama Kinki Chukyo Osaka DEN IMC 5. Discussions What determines the variance in small clusters? Well structured and not-well structured networks DEN IMC Well-structured: Nagano, Hamamatsu, Okinawa Not-structured: Hiroshima-Okayama, Sapporo, Hukuoka

18 Hub (Leader in the module) Connector-Hub analysis Nagano Hamamatsu Okinawa Hiroshima- Okayama Sapporo Hukuoka Connector (linking modules)

19 Hub (Leader in the module) Connector-Hub analysis Provincial Hub4 Connector 1 Hub Nagano Hamamatsu Okinawa Hiroshima- Okayama Sapporo Hukuoka Connector (linking modules)

20 Comparison of modular structures Ex. Nagano 1,930 firms Kyorin Chinon-Tech Number of firms in each module Strength of ties between modules Nihon-Dennetsu Nihon-Densan-Sankyo, Tamagawa-Seiki Mitsubishi-electronics Yuasa-Shoji Nihon-Densan-Kopal Fujitsu, Toshiba, NEC Shinko-Denki, Nisse-Jushi Techno-Excel Kitz Seiko-Epson

21 Comparison of modular structures Ex. Hamamatsu 1,049 firms Enshu-Seiko Sanga Yamaha Hamamatsu Photonics Number of firms in each module Strength of ties between modules Suzuki-Yuso-Konpo Yamaso Honda Yamaha Motor Hayashi-Kikou Enkei Denkosha Suzuki Asahi-Denso Daiwabo-Information-System Shizuoka-Ricoh

22 Comparison of modular structures Ex. Okinawa 57 firms Kinki-Nihon-Tourist Coca-Cola-Bottling Number of firms in each module Strength of ties between modules Nettai-Shigen-Syokubutsu-Kenkyujo San-Bio Okinawa-Mitsukoshi Zima, Nanto-Syuhan, Orion-beer Kanehide-Bio Hokugan Okinawa-Nougyou- Kyoudou-Kumiai Kanehide-Shoji Ryukyu-Jasco Okinawa-Kyoudou-Seika Mitsui-Bussan Okinawaken-Bussan-Kousya Daimon-Shouji Mitsubishi-Shoji, Itoutyu

23 Comparison of modular structures Ex. Hiroshima-Okayama 3,553 firms Daiwabo-Information-System Number of firms in each module Strength of ties between modules Eagle-Kougyou Umetoku J.S.T.-Connectors Marugo-Gomu-Kougyou JFE-Steel IHI Ichikawa-Bussan Miyaji Nishikawa-Gomu-Kougyou Fujitoku-Bussan Mitsubishi-Chemical Mitsubishi-Heavy-Industry Naganoka-Syouji Teral Manakku Maeda-Kikou

24 Comparison of modular structures Ex. Sapporo 1,871 firms Fujitsu Houkensyo Number of firms in each module Strength of ties between modules Air-Water Higashi-Nihon-House Kaneshime-Takahashi-Suisan NEC Mitsui-Bussan Hitachi Oryx Toppan Suzuken Docon Tesk Kuwazawa-Kougyou Tanbaya Hokuren Visual-Systems

25 Comparison of modular structures Ex. Hukuoka 931 firms Itou-Tyouonpa Arata Number of firms in each module Strength of ties between modules Sumitomo-Shouji Organo Sanix Suzuken Nishitele-Information-Service Mitsui-Bussan Nihon-Arcon Kyudenko Hitachi Toshiba-Medical Nakagawa-Kenzai Toshiba Cannon Nihon-Kayaku Syoukou

26 Propositions Regional networking can be improved by the existence of; 1. Connector-hub firms. Headquarters of them in the region.

27 Number of headquarters of public firms Rank Prefecture # Rank Prefecture # Rank Prefecture # 1 Tokyo Gunma 7 33 Wakayama 7 Osaka Ishikawa 7 34 Ehime 7 3 Aichi Okayama 0 35 Kochi 7 4 Kanagawa 07 0 Kagawa 0 36 Aomori 6 5 Hyogo 11 1 Tochigi Oita 6 6 Fukuoka 81 Mie Okinawa 6 7 Saitama 79 3 Toyama Tokushima 5 8 Kyoto 7 4 Fukui Kumamoto 5 9 Shizuoka 64 5 Ibaraki Miyazaki 5 10 Hokkaido 56 6 Yamaguchi 14 4 Nara 4 11 Chiba 56 7 Shiga Tottori 4 1 Hiroshima 47 8 Yamanashi Saga 4 13 Niigata 41 9 Kagoshima Akita 3 14 Nagano Fukushima Nagasaki 3 15 Miyagi 9 31 Yamagata 9 47 Shimane 16 Gifu 8 3 Iwate 7

28 # of headquarters / # of worker population Normalized number of headquarters of public firms 本社所在上場企業数 / 労働力人口 ( 十万人 ) Tokyo Osaka Hukuoka Kyoto Hyogo Shizuoka Aichi Kanagawa Saitama Hokkaido Chiba # of worker population 労働力人口 ( 十万人 )

29 6. Conclusion 1. System development: Our tool includes indices (density, clustering coefficient, path length, Inter-module coordination, connector-hub), and visualization (network & module).. Evaluation of network structures of 18 regional clusters in Japan. 3. Elucidation of networking mechanism: network structures are dependent of cluster size (# of firms), especially in the nascent stage of development, and they are determined by the existence of connector-hub firms not but provincial hub firms, and headquarters of hub firms in the region. 4. Implication: Policy for networking is effective for small clusters. To develop dense networks, rearing regional firm or attracting headquarters from the other regions especially Tokyo is are necessary.

30 Limitations and future works Limited Scope: Supplier-Customer relationships among firms Organization Extended network analysis Other regions Other countries visible layer Contract (Transaction, Joint research, etc.) invisible layer Human Technology Multiple layer structures of network consisting of various types and scopes of nodes and links

31 Thank you for your listening! Any questions, comments, suggestions, and proposals are welcome! Innovation Policy Research Center (IPR-CTR) University of Tokyo, Japan 31

Monthly Report of Tuberculosis Surveillance, Japan January, 2016

Surveillance Center, RIT, JATA Monthly Report of Tuberculosis Surveillance, Japan January, 216 The reporting and recording of tuberculosis () and Latent Infection (LI) are managed at public health centers