As indústrias da transformação são concentradas espacialmente? Um teste empírico para o Brasil (2002-2014)

Abstract

portuguesResumo O artigo apresenta novas evidencias da concentracao industrial no Brasil para o periodo de 2002 a 2014. A partir de Ellison e Glaeser (1994; 1997) e usando os microdados dos estabelecimentos da RAIS-ME, por meio de dois testes estatisticos, verificam-se quais divisoes da industria da transformacao sao concentradas. A analise e desagregada setorialmente, e explora-se a dinâmica do indicador de concentracao (EGI). Os resultados indicaram que as industrias sao mais concentradas do que seriam se as escolhas locacionais das firmas fossem aleatorias, e que a maioria das industrias tenderam a se desconcentrar ou o EGI ficou relativamente estavel, a excecao da industria de “Equipamentos/Material Eletronico”, que e de alta tecnologia. Porem, a tradicional “industria de Artefatos de Couro”, intensiva em mao de obra, tambem esteve entre as primeiras colocadas no ranking do EGI, sugerindo que o labor pooling, entre outros fatores, e uma importante economia de aglomeracao da industria. EnglishAbstract The article presents new evidence of industrial concentration in Brazil for the period from 2002 to 2014. Based on Ellison and Glaeser (1994, 1997) and using the microdata of the RAIS-ME establishments, it is verified which manufacturing industries are concentrated through two statistical tests. The analysis is disaggregated by sector, and the dynamics of the concentration indicator is explored. The results indicated that industries are more concentrated than they would be if the firms’ locational choices were random. And that most industries tended to deconcentrate or the EGI was relatively stable, with the exception of the industry Equipment/Electronic Material, intensive in technology. However, a traditional labor-intensive “Leather Industry” was among the first placed in the EGI ranking, indicating that the labor pooling is an important agglomeration economy to the industry.