Abordagem geoestatística para identificação de potenciais clusters industriais

Abstract

The aim of the present study was to develop a geostatistical approach for the identification of potential industry clusters. It is based on the premise that new spatial concentration indices are fundamental to integrate economics and geography in the formulation of regional development and business competitiveness policies. In the first part of the paper, cluster theory was divided into three strands in which each one has its own form of estimation. In the first case, the Pure Agglomerations are identified by locational indices, in the following view, the Industrial Complexes, the sectorial grouping is defined by means of input-output relations and, finally, clusters of Porter integrate the two previous approaches. Through a bibliometric analysis, it was identified that methods based on indices and spatial statistics, especially in point processes, represent the mainstream in the field of knowledge on the measurement of firm clusters. This generation of studies became prominent because it circumvented the aggregation bias, but it was confirmed that the problem of directional bias (anisotropy) remains neglected, since the researches assume isotropy. Questions of proximity and concentration of firms were examined through geostatistics. This approach was able to meet the principles already consolidated by the mainstream literature, as well as aggregated the directional bias analysis, the zoning of potential industry clusters on maps, and the estimation of firm-level industry concentration. Directional analysis represented better the clustering of firms from a statistical point of view, with a lower level of error, and economic, grouping the firms in regions with a homogeneous profile that tends to facilitate cluster strategic coordination. The geostatistical approach was applied in the roasted coffee industry in Minas Gerais and the potential clusters identified were in the regions known as Matas de Minas, Capelinha and Sul de Minas.