Facing intensified global competition, many countries around the globe, especially developed countries, try to create industrial clusters. Almost all these clusters are designed to accelerate inter-firm networking, but empirical investigations of the impact of clusters on firms’ transaction networks have so far been limited. In a new paper, we examine this issue in the case of Japan, based on firm-level transaction data, and discuss how inter-firm networking is affected by the cluster firm’s relationship with financial institutions (Okubo et al. 2016).
The concept of clusters can be traced back to a series of influential management papers by Michael Porter. In his famous diamond diagram, Porter argues that networking or linkages between suppliers and customers is crucial for creating new businesses. He also regards cluster policy with private sector leadership distinct from traditional industrial policy characterised by targeting and direct subsidisation.
In economics, agglomeration economies can be regarded as a topic closely related with clusters. While it has traditionally been examined for areas within geographical proximity, the externality spills over to remote locations via transaction networks, especially in information or service-intensive sectors. According to Duranton and Puga (2004), there are three types of micro-foundations of Marshallian agglomeration: sharing, matching and learning. All three are obviously connected with networking.
In spite of its importance in agglomeration economics and despite the policy priority of clusters, no empirical evidence has so far been reported on transaction networks, mainly due to the limited availability of transaction data. Expanding transaction networks should be a vital channel not only for increasing transaction volumes and varieties, but also for knowledge spillovers through exchanges of goods and services. This is important, for example, for local firms supplying to multinational enterprises in developing countries and firms in peripheral regions trading with others in agglomerated cores, such as Tokyo in Japan.
By exploiting firm-level transaction data in Japan, we fill a part of this research gap and link this transaction dataset with the list of firms participating in industrial clusters targeted by the government. The Japanese Ministry of Economy, Trade and Industry started the industrial cluster policy in 2001. While the ultimate policy purpose of the Japanese clusters is to create new businesses by fostering innovation, the immediate target of the cluster policy is facilitating inter-firm networking and industry-government-academia collaborations. The government, in the first Industrial Cluster Plan, lists ‘the formation of face-to-face networks’ as top among various policy objectives. The main policy tools for network formations include holding exchange meetings, seminars and exhibitions, dispatching coordinators, developing overseas sales channels, with support from the Japan External Trade Organization (JETRO), facilitating business matching between firms in different sectors, and matching with financial institutions.
The government designates industrial clusters from applications by groups of local firms. Members of the local Chamber of Commerce and Industry often play pivotal roles in preparing these applications. As an important point distinguishing industrial clusters from traditional regional development projects, the clusters in this policy initiative are not tightly defined as geographic spaces with clear boundaries. No minimal or maximal requirement thresholds are imposed on regional characteristics either. This implies that not all firms are automatically entitled to receive subsidies merely by locating in a cluster area. This marks a clear contrast from regional development policies in earlier periods strictly tied to targeted locations, as examined by previous research (e.g. Devereux et al. 2007, Okubo and Tomiura 2012). With the coordination by the branch office of the national government in each regional block, some of the clusters spread across prefecture borders. To accelerate cross-fertilisation of ideas and/or to expand transaction opportunities, the government actually facilitates networking not only within each cluster but also with firms outside of the cluster.
We regressed the growth of a firm’s transaction network on the dummy variable indicating the firm’s participation in the cluster programme, as well as on other control variables, to find that a firm participating in the cluster programme increased its transaction network at a significantly higher speed. Regressions dividing the transaction networks by the location of the transaction counterparts indicate that the effect of cluster programme accelerating the expansion of transaction networks was principally driven by the effect on the networks with firms in Tokyo.
The paper further examined the implication of the relationship with banks on the growth of transaction networks. To explore potential business partners, firms gather information from various sources. The main bank of a firm is valuable and reliable in this regard, as accumulated literature on relationship banking suggests. As most of the firms in clusters are small-sized or young and often located in peripheral regions, professional consulting advice from their main bank based on long-term relationships should be helpful in searching for transaction partners in different regions, especially in distant urban agglomeration.
To see the implication of the relationship with banks, we disaggregated firms depending on the main bank type, and found that firms with expanding networks are mainly financed by regional banks, not by banks with nationwide and global operations. This suggests that information provided from the main bank in the same region appears critical for cluster firms to expand transaction networks with firms in the agglomeration centre in Tokyo.
To handle potential non-randomness in the selection of regions, we also use historical records of other regional development projects in earlier periods as an instrumental variable, and confirm the robustness of our main finding (Okubo et al. (2016). Also, to check the robustness of our regression results, we also use propensity score matching. These analyses confirmed that our results above are robust.
Editors' note: The main research on which this column is based appeared as a Discussion Paper of the Research Institute of Economy, Trade and Industry (RIETI) of Japan.
Devereux, M, R Griffith, and H Simpson (2007), “Firm location decisions, regional grants and agglomeration externalities”, Journal of Public Economics 91, 413-435
Duranton, G, and D Puga (2004), “Micro-foundations of urban agglomeration economies”, chapter 48 in J V Henderson and J F Thisse (eds.), Handbook of Regional and Urban Economics Vol. 4, 2063-2117, Elsevier, Netherland.
Okubo, T, T Okazaki, and E Tomiura (2016), “Industrial cluster policy and transaction networks: Evidence from firm-level data in Japan”, RIETI Discussion Paper 16071.
Okubo, T, and E Tomiura (2012), “Industrial relocation policy, productivity, and heterogeneous plants: Evidence from Japan”, Regional Science and Urban Economics 42, 230-239.