Files in this item



application/pdfAguinaldo_Maciente.pdf (8MB)
(no description provided)PDF


Title:The determinants of agglomeration in Brazil: input-output, labor and knowledge externalities
Author(s):Maciente, Aguinaldo
Director of Research:Nelson, Charles H.
Doctoral Committee Chair(s):Nelson, Charles H.
Doctoral Committee Member(s):Baylis, Katherine R.; Feser, Edward J.; Garcia, Philip
Department / Program:Agr & Consumer Economics
Discipline:Agr & Consumer Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
natural advantages
Abstract:This research investigates industry agglomeration and coagglomeration patterns in Brazil, and assesses their association with the Marshallian forces that are commonly viewed as sources of agglomeration economies, namely the input-output and labor pooling externalities. Knowledge externalities, the third classic source of Marshallian agglomeration economies, are partially captured through labor-embodied knowledge usage. Industry-specific agglomeration and the coagglomeration of pairs of industries are measured with the use of Ellison and Glaeser (1997), and Moran’s I indices. Direct and indirect input-output linkages are described and measured with the use of direct shares and dual scaling techniques. The ONET database, which contains several skill measures for US occupations, is matched with Brazilian occupations and factor analysis is used to produce a set of ONET skill and knowledge groups. These skills groups are intended to describe the labor profile of industries and regions and constitute the basis for measures of labor and labor-embodied knowledge externalities for pairs of industries. The measures of input-output linkages and labor-use similarity are related to the observed agglomeration and coagglomeration patterns, in order to test for the possible sources of agglomeration economies in Brazil. Results indicate that Brazil has agglomeration levels, as measured by Ellison and Glaeser’s (1997) agglomeration index, that are slightly decreasing over time, but comparable to the international experience. However, the components of the agglomeration index reveal that Gini-type regional employment concentration and plant-level employment concentration are relatively high, despite their decrease from 1994 to 2010. That is, Brazil has most of its employment concentrated into relatively fewer regions and plants, when compared to results found in the literature, for example, for the United States. Agglomeration and coagglomeration patterns indicate higher agglomeration of oil and mining industries, which are based on local natural advantages, as well as of electronics and transportation equipment industries, which are influenced, in Brazil, by the existence of tax incentives in the Amazon region. The “Trucks and buses” industry is the most coagglomerated with other industries, especially with the “Cars, vans and utility vehicles” industry. The coagglomeration of this pair of industries is beyond what could be expect from all types externalities. Since some large firms in this industry pair belong to the same corporations, internal economies of scope may be driving the coagglomeration of these industries. Moran’s I values show that municipal employment displays spatial autocorrelation for several agricultural industries, due to spatially autocorrelated agricultural potential, and for manufacturing industries producing metal products (auto parts and metal forgings, among others) and products intensive in chemical inputs (paints, soaps, rubber products, and plastics). Bivariate Moran indices also indicate the coagglomeration of agricultural production, and of industries that are intensive in the use of metal and chemical inputs. Labor skill and labor-embodied knowledge-use similarities between pairs of industries show that the most collocated industries (agricultural products, and manufacturing based on metal, glass and chemical inputs) also display a high degree of similarity in the use of labor skills and knowledge labor requirements. Overall colocation patterns in Brazil seem to be more associated with labor and labor-embodied knowledge externalities than with input-output externalities Natural advantages, such as agricultural and mining potential and road density are also positively associated with observed coagglomeration. Certain labor skills, particularly those related to manufacturing production are negatively associated with coagglomeration. This is an indication that certain types of skills may be subject to labor poaching, or diseconomies of agglomeration, driving industries to locate far from other sectors that are intensive in the same skills they use. These finding may also suggest the relative scarcity of certain types of manufacturing skills in Brazil. Knowledge externalities are also associated with coagglomeration patterns in general and with manufacturing coagglomeration in particular. Services, however, do not have their coagglomeration associated with knowledge externalities, contrarily to evidence found for the US. This difference may be a consequence of larger differences for services than for manufacturing between Brazil and the US. Tradable good produced by manufacturing are likely to be more influenced by international production and competition than services, which are primarily dependent on each economy’s income and productivity levels.
Issue Date:2013-08-22
Rights Information:Copyright 2013 Aguinaldo N. Maciente
Date Available in IDEALS:2013-08-22
Date Deposited:2013-08

This item appears in the following Collection(s)

Item Statistics