Classificação setorial das empresas brasileiras: uma aplicação da análise de cluster

Abstract

The decision-making in a company requires knowledge of the external environment in which it operates. The group of companies, according to an activity or with the behavior of certain variables regardless of the type of product or service, serves as a basis for understanding the relationships and characteristics of businesses in markets that sell or buy. This thesis seeks to evaluate the efficiency of the sector classification from the cluster analysis method. The use of rating sector in studies in the administration area has been based on the assumption that companies within the same industry are similar. So if the classification sector is efficient, it is expected that companies within the same industry form a cluster of industries and companies that are grouped into different clusters mixed. The results, in general, show a high heterogeneity in sectoral clusters or when the number of clusters is suggested by the rule of stopping or BOVESPA criterion. And this heterogeneity is stable, as it remained throughout the study period. Therefore, the results point to the inefficiency of sectoral analysis. The study group of companies from financial and economic data served to identify the lack of homogeneity of the sectors in terms of the indicators when subjected to a classification based solely on the type of goods or services produced. For more precise decisions taken in those sectors can be restructured to incorporate in each group, companies that can be reasonably similar.