<論説>産業集積の進化と近接性のダイナミクス : 知識学習とネットワークの視点から (特集 : 学びのネットワーク)

Abstract

現在の経済地理学において知識学習とネットワークが重要な鍵概念となっている。産業が集積することは、地理的に近接した主体間での知識の学習を通じて、集積地域の発展をもたらすとされてきた。しかし、それは時系列でみると永続的なものではない。産業集積が特定の産業に特化することは、しばしば負のロックインの状態をもたらし、時間の経過とともに成熟から衰退の経路をたどりうる。負のロックインを防ぐためには、地域の既存産業から技術的に関連ある産業を分岐させて産業の多様性を維持することが必要となる。また、主体間のネットワークにおいて知識の学習を決定するのは地理的近接性だけではなく、制度的や組織的、社会的など様々な次元での近接性もまた学習を促す。新しい知識を探索するためにはこれらの諸次元で適度な近接性を保つことが求められる。地域発展を考える際には、ネットワークと近接性の両面について動態的にとらえる視点が必要となる。

This paper aims to explore the dynamic understanding of learning and networks in industrial clusters. In recent years, learning and networks have been key concepts in economic geography. Local agglomerations of industry have been thought to facilitate regional development through learning among geographically proximate actors. However, this localization of economies in clusters is not necessarily permanent. We need dynamic perspectives on learning and networks in clusters. Additionally, learning does not necessarily require geographical proximity, and geographical proximity cannot be a sufficient condition for learning. For a better understanding of learning in clusters, the concept of proximity needs to be reconsidered, and proximity dynamics should be explored. In the first half of this paper, the author considers recent arguments focused on cluster evolution. Specialized clusters often gradually mature and decline over time. As a cluster matures over time, the specialization of the cluster increases and the heterogeneity of accessible knowledge in the cluster decreases. This can lead the cluster to negative lock-in situations, which make it vulnerable to rapid changes in external economic environments and lead it to decline. The decline of clusters is likely to be caused by factors that were advantages in the past. In the cluster lifecycle approach, clusters are seen as following a kind of life cycle with stages from emergence to growth, maturity, and decline. However, this is not a predetermined process. It is true that some clusters gradually decline because of negative lock-in processes. But other clusters escape negative lock-in situations and create new growth paths. We have to discuss how clusters gain the adaptability to avoid negative lock-ins and create new paths. Several possibilities enable clusters to create new paths. One of the possibilities is the existence of redundancies and heterogeneities of accessible knowledge and capacities in clusters. Diversity in clusters promotes innovative activities by the recombination of existing varieties. Especially, related variety is likely to enable actors more mutual learning than unrelated variety. Another possibility lies in diversifying from existing industries or technologies in a cluster into related industries or technologies. This can increase the diversity of the cluster, and this diversity can, in turn, enable the cluster to diversify further. We have to understand various types of cluster evolution, and not be limited to the life-cycle model.

In the latter half of the paper, the author argues that the dynamic perspective of both proximity and networks is necessary for considering regional development. Effective knowledge learning does not necessarily presuppose geographical proximity between actors. Other dimensions of proximity (such as institutional, organizational, social, and cognitive dimensions) also enhance learning. While geographical proximity acts as a substitute for other dimensions of proximity, other dimensions of proximity can bridge geographical distance. We have to explore the interrelationship of these various dimensions of proximity. Furthermore, while too little proximity between actors prevents mutual understanding, too much proximity is unfavorable for knowledge creation because too- proximate actors often lack novel knowledge. To explore novel knowledge and to learn it effectively, an optimal level of proximity is required in these dimensions. To take the fact that proximity can change over time into consideration, it is necessary to rewire knowledge networks to keep an optimal level of proximity. We need a dynamic perspective on network relationships within and outside clusters for considering proximity dynamics. New network formations can be enhanced by cognitive proximity, which many social network studies have described as "homophily." Moreover, geographical proximity also contributes to forming a new connection to other actors because geographical closeness increases opportunities for new-tie formation. In a specialized cluster, the density of network relations can increase over time by geographical proximity combined with homophily. Although such dense network relations facilitate local learning within a cluster, heterogeneity of accessible knowledge in the cluster is likely to decrease. This leads the cluster to the stage of decline in the cluster lifecycle. To avoid this, various actors in clusters need to rewire their knowledge networks to connect geographically and cognitively distant actors who were previously unconnected.