Manufacturing location patterns in Brazil

Abstract

In this paper, we present a detailed and comprehensive analysis of the location patterns of manufacturing industries in Brazil using a distance‐based measure and micro‐geographic panel data for a 10‐year period. Our results show that 89.9% and 91% of manufacturing at the 3‐digit level have statistically significant localization for 2006 and 2015, respectively, and that these patterns remain high when we consider 4‐digit classification. High‐tech industries have location patterns at short distances, being localized mainly in large urban areas, while low‐tech industries are also localized at long distances. We also present evidence that agglomeration economies affect the degree of spatial concentration.

The spatial scope of agglomeration economies in Brazil

“…population, as revealed by Haddad and Barufi (2017), since they suggested that a longer distance to jobs may make it harder to benefit from more spatially located establishments' agglomeration gains. In addition, our set of evidence also helps shed light on the high level of spatial concentration of manufacturing industries at short distances recently pointed out by Almeida et al (2022) using distance-based measures.…”

“…In this regard, favoring our investigation about the effect of agglomeration gains on new entrants, we find that more than 90% of new establishments each year represented a new company in the period 2007-2014. 3 Reflecting the general location pattern of manufacturing establishments (see, e.g., Almeida & Rocha, 2018;Almeida et al, 2022;Lautert & Araújo, 2007;Rocha et al, 2019;Silveira Neto, 2005), the new enterprises are concentrated mainly in the Southeast and South regions (as seen in Table A1 in Appendix A). Both in 2007-2008 (75.18%) and 2013-2014 (71.66%) the two regions concentrated more than 70% of the new establishments.…”

“…The higher internal mobility has made economic activities in Brazil more market‐oriented and spatially concentrated. Using distance‐based measures of concentration, for example, Almeida et al (2022) recently presented evidence indicating that manufacturing activity in Brazil is more spatially concentrated than in other developing countries, such as China and Russia, and much more concentrated than in developed countries. The formation of this specific economic environment certainly would be impossible without the presence of strong urban agglomeration gains.…”

“…Technically, as will become clear in the next section, some industries have few new establishments annually, which poses a limitation on the use of distance-based measures(Duranton & Overman, 2005Klier & McMillen, 2008). In addition, to some extent, the 2-year period attenuates the influence of cyclical or more circumstantial variations in the number of establishments.5 As shown byAlmeida et al (2022) using the metric ofDuranton and Overman (2005), the establishments in these two industries are localized relative to overall manufacturing. Therefore, when evaluating the spatial distribution of the entrants relative to existing establishments for these specific industries, it is important to consider that the establishments in the latter group are already localized, so if the entrants present a location pattern similar to the existing one, they will still be localized relative to overall manufacturing.6 According to the technological classification of manufacturing industries into high-technology, medium-high-technology, mediumlow-technology and low-technology groups proposed byHatzichronoglou (1997) for OECD countries, and recent revisions (e.g., ISIC, 2011), adapted for Brazil by Cavalcante (2014) through compatibility with CNAE 2.0 (seeAlmeida et al, 2022).de ALMEIDA ET AL.…”

Convergence heterogeneity at the local level in sub‐Saharan Africa

“…https://ucdp.uu.se/downloads/index.html#ged_global.24 These two variables are constructed by taking the distance from the centroid of the AE in (100) km.25 The main cities are as follows: Bamako, Abidjan, Accra, Cotonou, Kigali, Daressalaam, Blantyre, Maputo, Lusaka and Johannesburg.26 de Almeida et al (2022) look at the location of high-tech and low-tech industries in Brazil using a distance base measure.27 Ruggedness is taken from Nunn and Puga (2012) and includes a latitude adjustment effect. Agricultural land suitability is taken from the Atlas of the Biosphere https://nelson.wisc.edu/sage.…”