Cluster identification

Abstract

PurposeThis paper aims to contribute to the approaches based on traditional industry concentration statistics for identifying clusters by complementing them with the techniques of exploratory spatial data analysis (ESDA).Design/methodology/approachUsing a sample with 34,500 observations retrieved from the social information annual report released by Brazil Ministry of Labor and Employment, the methodology was designed to make a comparison between the application of industry concentration statistics and ESDA statistics.FindingsAs the results show, the geographic distribution measures proved to be fundamental for longitudinal studies on regional dynamics and industrial agglomerations, and the local indicator of spatial association statistic tends to overcome the limitation of the industry concentration approach.Research limitations/implicationsIn the period considered, due to economic, structural and circumstantial questions, activities linked to the transformation industry have been losing ground in the value creation process in Brazil. In this sense, the study of other industries may generate other types of insights that should be considered in the process of regional development.Originality/valueThis paper offers a critical analysis of empirical approaches and methodological advances with an emphasis on the treatment of special effects: spatial dependence, spatial heterogeneity and spatial scale. However, the regional dynamic presents a temporal dimension and a spatial dimension. The role of space has increasingly attracted attention in the analysis of economic changes. This work has identified opportunities for incorporating spatial effects in regional analysis over time.