Submitted by Alice Araujo (alice.caraujo@ufpe.br) on 2018-04-12T17:37:26Z No. of bitstreams: 1 DISSERTAÇÃO Edilberto Tiago de Almeida.pdf: 1242121 bytes, checksum: 23245d1f144398ca96d534adc15929f0 (MD5) / Made available in DSpace on 2018-04-12T17:37:26Z (GMT). No. of bitstreams: 1 DISSERTAÇÃO Edilberto Tiago de Almeida.pdf: 1242121 bytes, checksum: 23245d1f144398ca96d534adc15929f0 (MD5) Previous issue date: 2017-02-14 / FACEPE / Esta pesquisa tem como objetivo obter evidências sobre a importância do labor pooling para explicar a aglomeração industrial no Brasil. Com este propósito, a investigação empírica foi dividida em duas etapas. A primeira tem como objetivo quantificar a aglomeração da indústria da transformação nas microrregiões que compõem o país no período 2002-2014, através do índice de Ellison e Glaeser (1997). Na segunda etapa, utiliza-se como base o modelo teórico proposto por Krugman (1991) (Labor Market Pooling Model) com adaptações de Overman e Puga (2010), e a investigação empírica segue a estratégia destes últimos que analisa como as firmas reagem a choques no mercado de trabalho que influenciam a sua produtividade. A ideia é, conforme o modelo teórico, se as firmas se beneficiam da concentração de mão de obra especializada, choques exógenos mais heterogêneos permitem que estas ajustem melhor sua produção e emprego sem incorrer em mudanças significativas nos salários. Com o objetivo de testar esta hipótese, modelos de regressão são estimados regredindo o índice de Ellison e Glaeser (1997) em função de uma variável proxy para o labor pooling, que tenta captar choques exógenos no mercado de trabalho, controlando tanto para características setoriais observadas que variam no tempo quanto pelo efeito fixo setorial. Os resultados são consistentes com uma redução no nível de concentração industrial no período de 2002 a 2014. Quando consideradas trajetórias de setores específicos, foi possível observar que os mais intensivos em tecnologia, em média, são mais concentrados que os menos intensivos em tecnologia, evidências que estão de acordo com outros estudos na literatura. No que concerne ao labor pooling, foi possível obter evidências que corroboram a hipótese teórica do modelo de que este fator local se relaciona positivamente com a aglomeração espacial da indústria, indicando que os setores industriais podem se beneficiar da concentração de trabalhadores. Através destes resultados, as evidências apontam que as firmas tendem a se concentrar espacialmente próximas de onde estão localizados os mercados de trabalho que concentram trabalhadores com as características exigidas em seus respectivos processos produtivos. Intuitivamente tanto as firmas podem considerar a localização da força de trabalho como fator importante em sua localização, quanto os trabalhadores podem se concentrar espacialmente em função da localização das firmas. Nesse aspecto, este estudo se concentrou no primeiro caso e fornece resultados consistentes em relação ao problema da causalidade reversa e demais fontes de endogeneidade. Adicionalmente, as evidências aqui apresentadas podem auxiliar na formulação de políticas norteadoras do direcionamento da atividade produtiva no espaço. / This research aims to obtain evidence about the importance of labor pooling to explain industrial agglomeration in Brazil. For this purpose, the empirical investigation was divided into two stages. The first one aims at quantifying the agglomeration of the transformation industry in the micro-regions that make up the country in the period 2002-2014, through the index of Ellison and Glaeser (1997). In the second step, the theoretical model proposed by Krugman (1991) (Labor Market Pooling Model) with adaptations of Overman and Puga (2010) is used as base, and the empirical investigation follows the latter’s strategy that analyses how firms react to shocks in the labor market that influence their productivity. The idea is, according to the theoretical model, if firms benefit from the concentration of specialized labor, more heterogeneous exogenous shocks allow them to adjust their production and employment better without incurring significant changes in wages. In order to test this hypothesis, the models are estimated by regressing the index of Ellison and Glaeser (1997) as a function of a proxy variable for labor pooling, which attempts to capture exogenous shocks in the labor market , controlling for both observed sectoral characteristics that vary in time as well as the sectoral fixed effect. The results are consistent with a reduction in the level of industrial concentration in the period from 2002 to 2014. When considering specific sector trajectories, it was possible to observe that the most technology intensive ones, on average, are more concentrated than the less technology intensive ones, evidences which are in agreement with other studies in the literature. With regard to labor pooling, it was possible to obtain evidence to support the theoretical hypothesis of the model that this local factor is positively related to the spatial agglomeration of the industry, indicating that the industrial sectors can benefit from the concentration of workers. Through these results, the evidence indicates that firms tend to concentring spatially close to where labor markets are located that concentrate workers with the characteristics required in their respective production processes. Intuitively the firms can consider the location of the workforce as an important factor in their location, as workers can spatially concentrate in function on the location of firms. In this aspect, this study focused on the first case and provides consistent results regarding the problem of reverse causality and other sources of endogeneity. In addition, the evidence presented here may help formulate policies that guide the direction of productive activity in space.

This thesis consists of three papers that examine Marshallian sources of growthand interdependent location of Swedish firms and households. Paper [I] examines the impact of static and dynamic knowledge externalitiesand their impact on Swedish market operating firms growth pattern between1997 and 2005. The three types of externalities are: (i) Marshall-Arrow-Romer(MAR), (ii) Jacobs, and (iii) Porter. My empirical findings for the 40 industriescan briefly be summarized in the following points: (i) static MAR, Jacobsand/or Porter externalities are present in all but nine industries; (ii) except for five cases all industries are exposed to one or more of the MAR, Jacobs and/orPorter type of dynamic externalities; (iii) contrary to previous studies but inline with theoretical predictions, we do find positive and significant effects forstatic as well as dynamic Jacobs externalities. Paper [II] focuses on the presence of agglomeration economies in the form of labor pooling and educational matching and their impact on economic growth in Swedish manufacturing and service industries from 1997 to 2005. To accomplish this I employ a translog production function that enables me to decompose the total agglomeration elasticities into returns that accrue to: direct agglomeration effects, an indirect effect of agglomeration at given input levels, a cross agglomeration effect of matching on labor pooling and vice versa. Household services is the single industry where both the labor pooling and matching hypothesis is supported by our data. Publishing is the sole instance of betterinput usage due to matching consistent with the theoretical claim. Paper [III] studies the interdependent location choices of households and firms expressed as population and employment in Swedish municipalities. Using a model of the Carlino-Mills type to investigate the impact of various location attributes such as differences in public revenue and spending patterns, accessibility to jobs and potential workforce, quality of the labor pool, concentration ofcommercial, private and public services. The findings suggest that fiscal factors significantly alters the impact of housing and accessibility attributes compared to exiting studies on Swedish data. Another finding, in line with previous studies, indicate that there is a significant degree of inertia in household and firm location choices.