Ensaios sobre economias de aglomeração e MAUP no Brasil: evidências a partir de equações de rendimento

• Ano de defesa: 2015
• Autor(a) principal: Carraro, Anderson Bonetto
• Orientador(a): Jacinto, Paulo de Andrade
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul; Porto Alegre
• Palavras-chave em Português: ECONOMIA - BRASIL; SALÁRIOS - BRASIL; EQUAÇÕES SALARIAIS; ECONOMIA DE AGLOMERAÇÃO
• Link de acesso: http://hdl.handle.net/10923/7399

The purpose of this dissertation consists of two articles, it is to analyze the influence of agglomeration economies and the Problem of Modifiable Area Unit (MAUP) have on the salary level. For this, two articles are needed to better explanation of these data. The first is dedicated to exploit economies of agglomeration from the earnings equations. Through the Exploratory Spatial Data Analysis (ESDA) and the 2010 Census data, there is the existence of spatial associations in the regional aggregates analyzed, such as spatial agglomerations or homogeneous regions (clusters), and atypical observations (outliers). In the second part of the article dedicated to the estimation of the econometric model, use up wage equations with formulation based on the model developed by Fingleton (2003), whose main hypothesis points to a positive relationship between wage differentials and job density. In the second article, we present a case of MAUP with the IBGE Census 2010 data on the average salary of Brazilian workers. It is discussed and demonstrated the need to consider the differences in results arising from the regionalization chosen by the researcher. The analysis focuses on five levels: micro, municipalities, micro, meso and Brazilian states, using for this purpose, earnings equations (Mincer), proposed by Mincer (1974). When analyzing Mincer equations, from the perspective of different regional scales, a first point to Brazil, the role of its variables becomes very important, especially for education. The results show that the conclusions drawn and the policies adopted, from income regressions, are dependent on the choice of geographically analyzed.