Sectorial neighborhood in the Brazilian manufacturing industry

The innovation in industry occurs due to direct investment of the firm, motivated by economic interests, as profit, market share, economies of scope, etc. However, same these decisions can be influenced by indirect shocks occurring in closest industry, as an innovation in automobile industry motivating changes in its chain production. The aim of this paper is to evaluate the peer effects on the innovation indicators in Brazilian manufacturing industry. For this, we proposed a new way to measure the proximity of the industries. We consider the typical goods produced by an industry in the main subsector and in other sub-sectors. These sub-sectors are considered neighbors because they use the same technological and production bases. The sectorial proximity was building through a detailed “sectorial diversification matrix” of firms producing goods in one or more subsector (of a total 103 subsectors of the Brazilian manufacturing) in 2013. Brazilian Institute of Geography and Statistics (IBGE) provided the data from a specific request, special to this work. The data composed a W-matrix relating one by one sector. As an innovation proxy, we considered the number of engineers employed divided by the total number of 2 employees in each subsector. This indicator is a proxy recognized of the innovative efforts of companies, because innovation and technological intensity are related to technical staff' skills. To test the peer effects hypothesis we considered the Moran's I test, and we used the LM and the LM robust tests to choose a best econometric model that confirms the sectorial spatial dependence. The results suggest the existence of strong sectorial peer effects among subsectors with the same technological level (low and medium-low, and high and medium-high). Moreover, there is a weak neighborhood effect between subsectors of different technological levels. Therefore, the technological level of a sector is a good indicator of their relations of sectorial neighborhood. These results can be useful for policymakers assign public policies focused on subsectors generating greater spillovers effects on the production and innovation structure. We think this work is a contribution to the use of spatial econometric analysis beyond the geographical neighborhood. It also sheds light on a fertile research agenda, which is currently absent in the industrial organization literature.