Resumo

Since the evolutionary turn, a great emphasis has been placed on the role of innovation and the circulation of knowledge for the development and survival of clusters. Such importance has directed the studies towards a greater understanding of how ties are formed within networks and their impact on innovation activity

In this article we explore this issue, by analyzing the relation between social networks analysis metrics and the innovation activities of the Brazilian Wine Cluster of Serra Gaúcha. The different metrics of a network have sociological meanings and can assist in the understanding of the collective processes of knowledge diffusion, an essential aspect for the innovation and survival of firms. In this sense, this study aims to study the relationship between innovation and network structure.

The approach of clusters emerged from the administrative sciences, aiming to reconcile the importance of geographic agglomeration for the economic performance of firms (Porter, 1998). The perspective of networks demonstrates that the cluster does not justify its existence only by market forces, but also through non-market interactions that can increase knowledge flow, resources, technologies and business opportunities. In this sense, the social bonds that emerge within the cluster allow the circulation of information and the increase of trust, facilitating the emergence of strong social ties.

We divided our study in three main stages. The first stage has as objective the formation of the Innovation Activity construct. For this analysis we conducted an Exploratory Factor Analysis. The second stage consists of mapping the knowledge exchange networks of wineries present in the WCSG. Such mapping took place through Social Network Analysis (SNA). Finally, the relationship between innovation activity and network metrics was established through different regression techniques and ensemble models.

Based on the results obtained, it is possible to verify that the use of non-linear models resulted in a significant improvement in the coefficient of determination and a reduction in the error statistics. On average, the variables that were most important in the different models were: Authority, Eingen Centrality, Hub and Closeness. The most important metrics highlight the importance of the quality of the ties more than the number of ties that a winery has within the network.

This research demonstrates that the innovation activity of a clustered firm is not necessarily related to the absolute number of connections, or to the hierarchical position in the network. For the acquisition of new knowledge, the position of a firm in the center or on the periphery of the network has a secondary role. The network metrics that showed a better predictive capacity were those related to the qualitative aspects of the relationships. The most irrelevant metrics were Eccentricity, Ego Size and Local Transitivity.

Galaso, P., & Miranda, A. R. (2020). The leading role of support organisations in cluster networks of developing countries. Industry and Innovation, 00(00), 1–30. https://doi.org/10.1080/13662716.2020.1856046 Giuliani, E., & Bell, M. (2005). The micro-determinants of meso-level learning and innovation: Evidence from a Chilean wine cluster. Research Policy, 34(1), 47–68. https://doi.org/10.1016/j.respol.2004.10.008 James, G., Witten, D., Hastie, T., & Tibshirani, R. (2013). An Introduction to Statistical Learning - with Applications in R. Springer. https://doi.org/10.1007/978-1-4614-7138-7