Esta dissertação apresenta um estudo sobre o mercado imobiliário de Porto Alegre que tem como objetivo verificar a formação de preços dos mercados de venda e de locação. Este objetivo é cumprido de duas formas: a) analisando como o tempo que um imóvel está no mercado afeta seu preço e b) comparando como diferentes modelos econométricos preveem os preços de imóveis da capital gaúcha. Para o primeiro tópico, encontram-se evidências de que os proprietários reduzem, em média, 9% o valor de locação para atrair clientes à medida que o tempo passa. Isto não se concretiza da mesma forma no mercado de venda, uma vez que há uma valorização à medida que o imóvel está no mercado. Para o segundo tópico, foram estimados cinco modelos (ingênuo, MQO, SAR, SEM e combinação) para averiguar a performance de previsão fora da amostra dos preços de imóveis. Embora as estatísticas descritivas dos erros não permitam análises conclusivas a respeito dos modelos, a estimação do model confidence set de Hansen et al. (2011) mostra o modelo ingênuo como a melhor série de previsões para os mercados de venda e de locação. / This thesis presents a study on the real estate market of Porto Alegre in order to verify the price formation in both sales and rental markets. This objective is accomplished in two ways: a) analyzing how time on market of a property impacts its price and b) comparing how different econometric models predict real estate prices. For the first topic, I show evidence that owners reduce on average 9% of rental value of available homes in order to attract customers as time on market goes by. This pattern is not present in the sales market, in which an appreciation occurs as time passes, indicating an inflation compensation. For the second topic, five models (naive, OLS, SAR, SEM and a combination) were estimated to determine out-of-sample forecasting accuracy. Although traditional error measures do not allow to choose the best forecasting series, the model confidence set procedure of Hansen et al. (2011) shows the naïve model as being the best forecasting model.

A presente dissertação objetivou analisar o impacto dos royalties do petróleo no PIB per capita dos municípios da região Sudeste brasileira, calculando os efeitos diretos, indiretos e totais das receitas nas localidades beneficiadas. Com efeito, foram empregados modelos econométricos espaciais para controlar as interferências de componentes espaciais nas estimativas, e também para calcular o efeito spillover. Através da análise dos dados e dos testes estatísticos decidiu-se estimar o modelo espacial de Durbin, ou SDM, com a matriz de distância de Haversine inversa com raio de até 200 km. O efeito direto das receitas confirma os indícios de presença da maldição dos recursos naturais nos municípios da região Sudeste. O efeito transbordamento indica que os municípios do Sudeste afetam os seus vizinhos de forma não linear, mas pela média de dependência aos royalties, há indícios de que os efeitos negativos são mais presentes do que os positivos. Restringindo a amostra para os municípios do Rio de Janeiro, onde a dependência é maior, foi encontrado que os royalties não possuem efeitos nem positivos e nem negativos nos municípios beneficiados, porém, o efeito transbordamento é negativo e significativo, indicando, portanto, a presença de uma \"maldição regional fluminense dos royalties do petróleo\". / This paper aimed analyze the oil royalties impact upon Southeastern municipalities per capita GDP, estimating both the direct and the spillover effect of such revenues. The spatial econometric models were estimated to control the spatial interferences in the econometric estimations and estimate the spillover effect as well. The data analysis and the statistical tests indicated the Durbin spatial model as the best, or SDM, likewise the inverse distance Haversine spatial matrix considering a 200-km radio. The direct effect confirmed the evidences of \"natural resource curse\" presence in the Southeastern cities. The spillover effect showed the oil royalties no-linearly impact the neighbors of contemplated cities, however, considering the average oil dependency, there is an evidence of the negative impacts overcome the positive ones. Narrowing the sample to the Rio de Janeiro, where the dependency is worst, it was founded that oil revenues had no effect upon per capita GDP of these municipalities. Nevertheless, the spillover effect was negative and statistically significant, indicating, thus, the presence of a \"oil royalties fluminense curse\".

O objetivo deste estudo foi analisar a relação dos casos de malária como desmatamento nos municípios da Amazônia Legal, entre os anos de 2003 e 2012. A pesquisa também intenta contribuir para a literatura ao realizar uma análise econométrica espacial, viabilizando capturar efeitos spillovers das variáveis em estudo sobre a incidência de malária na região. Primeiramente, evidenciou-se que os casos de malária na região da Amazônia Legal não são distribuídos aleatoriamente entre os municípios, ou seja, a doença é correlacionada espacialmente. Também foram encontrados indícios de que municípios com altos índices parasitários anuais de malária, estão rodeados de outros municípios que também apresentam alta incidência de paludismo. Em relação às estimações econométricas, foram encontradas evidências de que o desmatamento possui uma correlação estatística com os casos de paludismo, onde, áreas de até 425 Km2 de mata derrubada possuem uma relação positiva com os casos da doença, enquanto áreas desmatadas acima de 425 Km2 possuem relação negativa com contaminação por malária. Não obstante, o estudo também encontrou uma relação positiva dos dispêndios municipais per capita com saúde com os casos da doença. Já os gastos municipais com habitação e o efetivo bovino municipal tiveram efeitos negativos em relação à doença. Ademais, não foram encontradas evidências de que a área de soja e de que de mais culturas agrícolas possam afetar a dinâmica de infecção por malária. Já o Produto Interno Bruto municipal gerou resultados inconclusivos. Para aferir a robustez das estimativas, o estudo também realiza um exercício de comparação dos resultados obtidos a partir de diversos modelos de dados de painel, tanto convencionais, quanto espaciais, atualmente usados pela literatura. / The aim of this study was to analyze the relationship of cases of malaria with deforestation in the municipalities of the Amazon, between 2003 and 2012. The research also intends to contribute to the literature through of spatial econometrics models, enabling capture effects spillovers of several socio economic determinants on incidence of malaria in the region. First, it became clear that the cases of malaria in the Amazon region are not randomly distributed among the municipalities, meaning that the disease is spatially correlated. It was also found evidence that municipalities with higher annual rates of malaria parasites are surrounded by other municipalities that also have a higher incidence of malaria. Regarding to the econometrics estimates, it was found evidence that deforestation is strongly statistically correlated with malaria cases, where are as of up to 425 km2 of loss os forest cover has a positive relationship with the cases of the disease, while deforested areas above 425 km2 have negative relationship with contamination malaria. Notwithstanding, the study also found a positive relationship between health public expenditures with cases of the disease. Housing public expenditures ans cattle heard have shown negative relationships with malaria infection. Moreover, no evidence was found that the soybean area and that other crops can affect the dynamics of malaria infection. Finally, municipal GDP generated inconclusive results. To assess the robustness of the estimates, the study also performed a comparison from multiple panel data models, both conventional well as spatial, currently used in the literature.

O objetivo deste estudo foi analisar a relação dos casos de malária como desmatamento nos municípios da Amazônia Legal, entre os anos de 2003 e 2012. A pesquisa também intenta contribuir para a literatura ao realizar uma análise econométrica espacial, viabilizando capturar efeitos spillovers das variáveis em estudo sobre a incidência de malária na região. Primeiramente, evidenciou-se que os casos de malária na região da Amazônia Legal não são distribuídos aleatoriamente entre os municípios, ou seja, a doença é correlacionada espacialmente. Também foram encontrados indícios de que municípios com altos índices parasitários anuais de malária, estão rodeados de outros municípios que também apresentam alta incidência de paludismo. Em relação às estimações econométricas, foram encontradas evidências de que o desmatamento possui uma correlação estatística com os casos de paludismo, onde, áreas de até 425 Km2 de mata derrubada possuem uma relação positiva com os casos da doença, enquanto áreas desmatadas acima de 425 Km2 possuem relação negativa com contaminação por malária. Não obstante, o estudo também encontrou uma relação positiva dos dispêndios municipais per capita com saúde com os casos da doença. Já os gastos municipais com habitação e o efetivo bovino municipal tiveram efeitos negativos em relação à doença. Ademais, não foram encontradas evidências de que a área de soja e de que de mais culturas agrícolas possam afetar a dinâmica de infecção por malária. Já o Produto Interno Bruto municipal gerou resultados inconclusivos. Para aferir a robustez das estimativas, o estudo também realiza um exercício de comparação dos resultados obtidos a partir de diversos modelos de dados de painel, tanto convencionais, quanto espaciais, atualmente usados pela literatura. / The aim of this study was to analyze the relationship of cases of malaria with deforestation in the municipalities of the Amazon, between 2003 and 2012. The research also intends to contribute to the literature through of spatial econometrics models, enabling capture effects spillovers of several socio economic determinants on incidence of malaria in the region. First, it became clear that the cases of malaria in the Amazon region are not randomly distributed among the municipalities, meaning that the disease is spatially correlated. It was also found evidence that municipalities with higher annual rates of malaria parasites are surrounded by other municipalities that also have a higher incidence of malaria. Regarding to the econometrics estimates, it was found evidence that deforestation is strongly statistically correlated with malaria cases, where are as of up to 425 km2 of loss os forest cover has a positive relationship with the cases of the disease, while deforested areas above 425 km2 have negative relationship with contamination malaria. Notwithstanding, the study also found a positive relationship between health public expenditures with cases of the disease. Housing public expenditures ans cattle heard have shown negative relationships with malaria infection. Moreover, no evidence was found that the soybean area and that other crops can affect the dynamics of malaria infection. Finally, municipal GDP generated inconclusive results. To assess the robustness of the estimates, the study also performed a comparison from multiple panel data models, both conventional well as spatial, currently used in the literature.

Nas últimas quatro décadas, em especial, o meio rural brasileiro vem apresentando uma nova dinâmica em sua população e atividades, o que gera uma nova configuração. Esta nova configuração pode ser observada com relação à ocupação das pessoas no meio rural, com queda do número de pessoas ocupadas em atividades agrícolas com residência no meio rural no século XXI. Observando a nova configuração do meio rural para os estados de São Paulo, Paraná, Mato Grosso do Sul, Mato do Grosso e Goiás, observa-se que os mesmos têm apresentado os fenômenos acima citados. Entretanto, a dinâmica da ocupação do meio rural no Brasil não ocorre de forma homogênea. Isto posto, busca-se, através desta tese, analisar a distribuição e os determinantes da intensidade agrícola no meio rural dos municípios do estado de São Paulo, Paraná, Mato Grosso do Sul, Mato Grosso e Goiás para os anos de 2000 e 2010. Através da análise fatorial foram gerados três indicadores (representados pelos escores fatorais dos fatores extraídos) a saber: o indicador de ruralidade, da agriculturalização (intensidade das atividades agrícolas no meio rural) e da renda não agrícola. Analisando a distribuição espacial destes indicadores, representados pelos escores fatorais dos fatores extraídos, foi constatado que a ruralidade pouco se alterou entre os anos de 2000 e 2010. Também foi observado um aumento na intensidade das atividades agrícolas no meio rural e da renda não agrícola nos municípios pertencentes aos estados do Mato Grosso do Sul e Mato Grosso entre os dois anos analisados. Foram encontradas diferentes associações dos escores fatoriais ruralidade, agriculturalização e renda não agrícola. Levando a conclusão de que não existe um único padrão de evolução e associação entre ruralidade, agriculturalização e renda não agrícola nos estados analisados nos anos de 2000 e 2010. Por meio da análise exploratória de dados espaciais, constatou-se a presença de clusters espaciais Alto-Alto e Baixo-Baixo nas fronteiras dos estados e que a intensidade das atividades agrícolas do município i sofre influência das atividades agrícolas dos municípios vizinhos. A partir desse resultado, buscou-se estimar os determinantes da intensidade das atividades agrícolas no meio rural, utilizando os conceitos de variáveis de primeira e de segunda natureza, e empregando a econometria espacial. Os resultados das estimativas indicam a presença do efeito transbordamento, ou seja, um alto valor na variável intensidade das atividades agrícolas nos municípios vizinhos, aumenta o valor dessa variável no município i e que as variáveis valor total do crédito agrícola para custeio, valor da produção agrícola dividido pela área explorada, área explorada dividida pela área total dos estabelecimentos agropecuários, área não agrícola e distância do município até a capital estadual possuem relevância para explicar a intensidade das atividades agrícolas no meio rural dos estados analisados. Entretanto, as variáveis área explorada dividida pela área dos estabelecimentos agropecuários, distância do município até à capital estadual e área não agrícola são as mais importantes para determinar a agriculturalização dos municípios, indicando que a intensidade das atividades agrícolas nos municípios está mais ligada a eficácia do uso da terra. / In the last four decades, in particular, the Brazilian countryside has been showing a new dynamic in its population and activities, which leads to a new configuration. This new setting can be observed with respect to the occupation of the people in rural areas, with fall of the number of persons employed in agricultural activities with rural residence in the 21th century. Noting the new rural setting to the states of São Paulo, Paraná, Mato Grosso do Sul, Mato Grosso and Goiás, noted that they have presented the aforementioned phenomena. However, the dynamics of rural occupation in Brazil does not occur homogeneously. That said, we seek to, through this thesis, analyzing the distribution and determinants of agricultural intensity in rural areas of the municipalities of the State of São Paulo, Paraná, Mato Grosso do Sul, Mato Grosso and Goias for the years 2000 and 2010. Through factor analysis were generated three indicators (represented by fatorais scores of factors extracted) namely: the indicator of rurality, the agriculturalização (intensity of agricultural activities in rural areas) and of non-agricultural income. Analyzing the spatial distribution of these indicators, represented by fatorais scores of factors extracted, it has been found that the rurally little changed between the years 2000 and 2010. Was also observed an increase in the intensity of agricultural activities in rural areas and non-agricultural income in the municipalities belonging to the states of Mato Grosso do Sul and Mato Grosso among the two years analyzed. Different associations have been found of factorials, agriculturalização and rurality scores income non-agricultural.Leading to the conclusion that there is no single pattern of evolution and association between rurality, agriculturalização and non-agricultural income in the states examined in the years 2000 and 2010. Through the exploratory analysis of spatial data, it was found the presence of clusters High-High and Low-Low spatial-bass on borders and states that the intensity of agricultural activities of the municipality i suffer influence of agricultural activities of neighboring municipalities. From this result, sought to estimate the determinants of the intensity of agricultural activities in rural areas, using the concepts of variables and nature, and employing spatial Econometrics. The results of the estimates indicate the presence of spillover effect, i.e., a high value in the variable intensity of agricultural activities in the neighboring municipalities, increases the value of this variable in the municipality i and that the total value of the agricultural credit variables for costing, value of agricultural production divided by the area explored, exploited area divided by the total area of agricultural establishments, non-agricultural area and distance from the municipality to the State capital have relevance to explain the intensity of agricultural activities in rural areas of the states examined. However, the variables explored area divided by the area of agricultural establishments, away from the municipality to the state capital and non-agricultural area are the most important to determine the agriculturalização of the municipalities, indicating that the intensity of agricultural activities in the municipalities ir more on the effectiveness of land use.

<p>While competition among companies tends to be beneficial for the general public, this is not necessarily the case for competition among governments. Key in the fiscal competition theory is that the mobility of firms and households yields incentives for governments to aim to improve their relative position through successive undercutting of tax rates and welfare state arrangements. This mechanism has the potential to work as a disciplining device because it ensures that no jurisdiction is allowed to be grossly inefficient, because if it were grossly inefficient, mobile factors of production would move away. The main concern in the theoretical fiscal competition literature, however, has been that fiscal competition lowers government spending below their efficient levels. Another concern related to fiscal competition is that household mobility is likely to undermine attempts by governments to redistribute income. Empirical evaluation of both the existence and consequences of fiscal competition is the central topic of the thesis “An Empirical Analysis of Decentralization, Fiscal Competition and Welfare Policy”. </p><p>A particular focus of this thesis is on fiscal competition in welfare policy. With decentralized responsibility for the welfare benefit system in Norway, theory predicts that local governments will behave strategically in setting their welfare policy in order to avoid becoming ‘welfare magnets’. The key finding in Chapter 2 of this thesis is that Norwegian local governments in fact engage in such a ‘welfare game’. A local government will respond with reducing their welfare benefits when neighboring local governments reduce their welfare benefits. Encouraged by the finding in Chapter 2, Chapter 3 seeks to answer the question: Does Welfare Policy Affect Residential Choices? The analysis shows that Norwegian welfare recipients respond to changes in welfare policy by migrating. Local politicians concern about being to generous compared to their peers seem warranted. The analysis in Chapter 4 evaluates whether strategic interaction among Norwegian local governments in property tax decisions occurs. With limited mobility of the tax base and politically highly visible decisions, we interpret the strategic interaction found to be driven by yardstick competition, rather than competition for a mobile tax base. The final chapter differs from the rest in that it utilizes data from 18 OECD countries. The essay analyzes the effects of decentralization of government on the size and composition of government spending. Since jurisdictions with limited geographic scope (such as local governments) are, in general, more likely to face greater competitive pressures than larger ones (such as countries), it follows that the more fiscally decentralized countries are expected to experience stronger fiscal competition. One of the key findings is that decentralization of taxing powers is associated with less transfer spending, but unrelated to government consumption.</p> / Paper I reprinted with kind permission of Elsevier, Sciencedirect.com

While competition among companies tends to be beneficial for the general public, this is not necessarily the case for competition among governments. Key in the fiscal competition theory is that the mobility of firms and households yields incentives for governments to aim to improve their relative position through successive undercutting of tax rates and welfare state arrangements. This mechanism has the potential to work as a disciplining device because it ensures that no jurisdiction is allowed to be grossly inefficient, because if it were grossly inefficient, mobile factors of production would move away. The main concern in the theoretical fiscal competition literature, however, has been that fiscal competition lowers government spending below their efficient levels. Another concern related to fiscal competition is that household mobility is likely to undermine attempts by governments to redistribute income. Empirical evaluation of both the existence and consequences of fiscal competition is the central topic of the thesis “An Empirical Analysis of Decentralization, Fiscal Competition and Welfare Policy”. A particular focus of this thesis is on fiscal competition in welfare policy. With decentralized responsibility for the welfare benefit system in Norway, theory predicts that local governments will behave strategically in setting their welfare policy in order to avoid becoming ‘welfare magnets’. The key finding in Chapter 2 of this thesis is that Norwegian local governments in fact engage in such a ‘welfare game’. A local government will respond with reducing their welfare benefits when neighboring local governments reduce their welfare benefits. Encouraged by the finding in Chapter 2, Chapter 3 seeks to answer the question: Does Welfare Policy Affect Residential Choices? The analysis shows that Norwegian welfare recipients respond to changes in welfare policy by migrating. Local politicians concern about being to generous compared to their peers seem warranted. The analysis in Chapter 4 evaluates whether strategic interaction among Norwegian local governments in property tax decisions occurs. With limited mobility of the tax base and politically highly visible decisions, we interpret the strategic interaction found to be driven by yardstick competition, rather than competition for a mobile tax base. The final chapter differs from the rest in that it utilizes data from 18 OECD countries. The essay analyzes the effects of decentralization of government on the size and composition of government spending. Since jurisdictions with limited geographic scope (such as local governments) are, in general, more likely to face greater competitive pressures than larger ones (such as countries), it follows that the more fiscally decentralized countries are expected to experience stronger fiscal competition. One of the key findings is that decentralization of taxing powers is associated with less transfer spending, but unrelated to government consumption. / Paper I reprinted with kind permission of Elsevier, Sciencedirect.com

Organized Industrial Zones (OIZs) are used as a main and important industry policy tool in Turkey. In 2012, the number of OIZs is 263 with 148 active and 115 planned ones. Network between the firms reveal the knowledge spillovers which is inevitable for economic growth of a country for neo-classical economists. In this thesis, existence of intra-OIZ and intra-industry knowledge spillovers in Ankara and Istanbul is tested by the help of an export decision function. As it considers the spatial dependence between the regional units the spatial econometric method is preferred for the analysis. The data set is taken from the &ldquo / Field Research Survey&rdquo / of Small and Medium Enterprises Development Organization (SMEDO). It consists of 62,137 firm level observations from 24 manufacturing industries in 81 provinces between 2004 and 2007. After the cleaning process of the data, 1545 and 1172 observations are left for Ankara and Istanbul, respectively. The results show that the size of the firm (which is proxied by logarithm of total labor), technology (which is proxied by computer usage), organizational proximity and foreign language knowledge of the administrator are the common determinants of export decision for Ankara and Istanbul for both intra-IOZ and intra-industry relations when spatial dependence is not ignored. Besides these variables, in Ankara percentage of high skilled labor is significant while in Istanbul cluster proximity is significant. Moreover, for Ankara while for intra-OIZ relations the spatial effect is one third of the total effect, it is one fourth of the total effect for intra-industry relation. For Istanbul one fourth of the total effect is from spatial effects for both intra-OIZ and intra-industry relations.

The thesis consists of an introductory part and three self-contained papers. Paper [I] studies the determinants of the differences in expenditure on services for functionally impaired individuals among municipalities in Sweden. A spatial autoregressive model is used in order to test whether the decisions on the expenditure level in a neighboring municipality affect the municipality’s own expenditure. The results show of spatial interaction among neighbors, possible due to mimicking. However, when controlling for differences among counties there is no evidence of spatial interaction. Therefore, the positive interaction first found can be interpreted either as a result of differences in the way county councils diagnose individuals or due to interaction among the neighbors in the same county. Paper [II] takes advantage of a new intergovernmental grant in two ways. First, the grant is used to study the effect on municipal spending related to the grant. Second, the grant is used to test a hypothesis of spatial interaction among municipalities due to mimicking behavior. The data used pertains to the periods before and after the introduction of the grant. A fixed-effects spatial lag model is used to study the spatial interactions among municipalities. The results show that before the grant, municipalities interact with their neighbors when setting the expenditure level, while there is no evidence of interaction in the second period. This would support the hypothesis that the grants provide information to the municipalities and the need for mimicking diminishes with the grant. Paper [III] examines whether local public expenditures on services to functionally impaired individuals crowd out other local public expenditures in Sweden. The hypothesis is tested on five different spending areas using a two-stage least squares (2SLS) fixed-effects model. While the results give no support for crowding out in the areas of social assistance, culture &amp; leisure, and childcare &amp; preschool, a negative relationship on spending for elderly &amp; disabled care and on spending for education is found, suggesting that crowding out indeed occurs within the municipal sector. The negative relationships are significant both in a statistical and an economic sense.

The empirical literature in economic geography has recently dealt with two fundamental questions, whether growth or innovation is spatially bounded or not and how far the economic growth is determined by knowledge. In this thesis, relations between economic growth and knowledge relation is discussed from spatial spillovers perspective with the help of spatial econometric techniques. Adding city-region discussion to the existing literature, the thesis aims to evaluate the economic growth and knowledge spillovers from a broader perspective. The selected cases are two dynamic and rapidly transforming centers, namely izmir and Antalya City-regions and the results suggest the strong relevance of proximity effects and spillovers in both of the cases. Moreover, the analyses show that growth and knowledge spillovers operate in opposite directions rather than a parallel pattern as expected in Knowledge Based Economy discussions.