The identification of industrial clusters: Methodical aspects in a multidimensional framework for cluster identification

Regional politicians regard the cluster concept as a key factor for promoting regional economic performance. They refer to a comprehensive set of theoretical and empirical literature which provides the clusters' positive impact on regional development. Recent studies dealing with that topic underline the multi-dimensional character of industrial clusters. However, empirical studies about the existence of positive effects of industrial clusters are mostly based on case-study material. This procedure might be suitable for the description of individual cases. If one demands for a comprehensive identification - for example regional clusters in a national framework - this approach is not capable to produce comparable results. To solve this problem literature provides a widely oriented set of systematic approaches. Previous approaches mostly are limited to the identification of either the horizontal dimension, the spatial concentration of firms of one industrial branch or the vertical dimension, the spatial concentration of firms linked through input-output relations. Empirical cluster research should go along with theoretical progress - this in mind a multi-dimensional framework for cluster identification is needed. This background in mind, our paper contributes to the literature about systematic methodologies for industrial cluster identification. Our paper ties in with a methodological approach which has been undertaken by Titze et al. (2009) combining measures of industries' concentration and input-output-methods to identify clusters from a horizontal and vertical perspective. Extending this methodological framework using innovation interaction matrices we are capable to refer to the important role of regional knowledge networks, institutionalized through formal cooperation projects. We apply this extended framework to the federal state of Saxony in Germany and describe the regional allocation of industry clusters, different regional sources of knowledge and their degree of interaction. One interesting finding lies in the result that most local or regional innovation interactions for knowledge generation do not take place within the traditional boundaries of industrial clusters identified when using concentration measures only. Cross-sectoral cooperation, based on complementary knowledge is the major source of knowledge for industrial clusters. Although the cluster structures show a low degree of potential vertical interaction, no vertically isolated clusters exist in the Saxon cluster network.