Rede de Pesquisa - Dissertações

Navegar

Submissões Recentes

Agora exibindo 1 - 9 de 9
  • Item
    Estimando a taxa neutra de juros dos países emergentes e investigando o nível da taxa de juros do Brasil
    (2019) Schulz, Evandro Costa de Oliveira; Muinhos, Marcelo Kfoury
    Este trabalho estima a taxa neutra de juros de um conjunto de 26 países emergentes para períodos entre 1995-2018. Um dos objetivos é verificar se a taxa de juros do Brasil é muito elevada frente ao nível da taxa de juros de outras economias emergentes. Para isso, realizou-se uma revisão da literatura sobre o tema e buscou-se por modelos econométricos que são utilizados na estimação da taxa neutra de juros, como Filtro HP (Hodrick e Prescott), Regra de Taylor, assim como regressão de dados em painel com efeitos fixos. Os resultados obtidos neste trabalho apontam que: (a) a taxa neutra de juros brasileira é uma das mais elevadas entre emergentes; (b) existe uma forte correlação positiva entre o nível da taxa neutra de juros e o nível médio da inflação; (c) a taxa neutra de juros vem diminuído consistentemente ao longo das últimas décadas na maioria dos países emergentes analisados, assim como no Brasil; e (d), no período entre 2015-2018, África do Sul, Brasil, Índia, Indonésia e Rússia apresentaram os maiores níveis de taxa neutra de juros entre economias emergentes, enquanto que (e) economias emergentes na região da Europa Oriental apresentaram níveis menores.
  • Item
    Fish farming and malaria: a model of mosquito population control
    (2019-08-16) Antunes, Felipe José Pinto
    Estudamos dois problemas matemáticos distintos, mas relacionados, alcançando resultados teóricos e aplicados. Estudamos o problema da malária na região do Alto Juruá, no Acre, propondo um sistema de equações diferenciais (incluindo uma equação impulsiva) como modelo epidemiológico levando em conta a relação entre a falta de manutenção dos tanques de piscicultura e a malária. Para isso, foi organizada uma expedição de campo acompanhando uma equipe da Fundação Oswaldo Cruz, revisaram-se artigos sobre a região e fez-se uma revisão histórica da modelagem matemática da malária. Descrevemos o nosso modelo, integrando a dinâmica vector-hospedeira da malária, o ciclo de vida de mosquitos Anopheles darlingi e a limpeza e crescimento da vegetação de borda em tanques de piscicultura. Teoremas sobre o comportamento assintótico do sistema foram deduzidos, e o sistema foi numericamente simulado e analisado. Nossa análise argumenta em favor de empregar campanhas de limpeza de tanques de peixes e campanhas educativas como medidas de controle da malária. Durante a análise, percebeu-se que nosso modelo não satisfazia as condições para aplicar o Teorema de Smith (H. L. Smith. Cooperative systems of differential equations with concave non-linearities. Nonlinear Analysis: Theory, Methods Applications, 1986) sobre o comportamento assintótico de equações diferenciais cooperativas e côncavas. Mais especificamente, o teorema assume a continuidade de tempo da função dinâmica, que nosso modelo não satisfez. A fim de contornar este problema, o background teórico do teorema foi estudado, e o resultado foi estendido para funções com dinâmica apenas mensurável no tempo.
  • Item
    Mining jams into pollution: how Waze data helps estimating air pollution in large cities
    (2019-06) Carabetta, João Luiz Martins
    Air pollution has been a growing worry of international medical organizations and governments due its relation with a large number of respiratory diseases, among other effects. This work proposes an open, crowdsourced, scalable methodology to model spatial air pollution in cities worldwide. We use both Waze and Open Street Maps data to construct a collection of features aimed to model car emissions in (large) cities. Waze data carries information about all jammed road segments of a region for every two minutes and Open Street Maps (OSM) is an open source, detailed, dynamically updated, spatial database of mapped features. Our model is trained using data from a 30 sq km region of Oakland, California in the United States of America. The dependent variables are the annual concentration of fine grained black carbon, nitric oxide, and nitrogen dioxide. The features are aggregated in hexagons with a 173 meters edge. We notice that pollutant concentration between hexagons follows a power law and high concentration is associated with the presence of highways. We estimate four models: simple linear regression where the only feature is the presence of a highway in the hexagon, multiple linear regression, random forest, and XGBoost. The latter yields better results in the validation set for black carbon, NO and NO2. Finally, we extrapolate the model for Montevideo, Uruguay and observe adherence to what is expected in practice.
  • Item
    O nível de consistência das informações financeiras dos municípios brasileiros
    (2019-05-28) Silva, Claudiane Campos da
    Este estudo buscou investigar o nível de consistência das informações apresentadas nos relatórios financeiros dos municípios brasileiros e sua aderência aos princípios da governança aplicada ao setor público. Para esse fim, foram avaliadas as demonstrações financeiras dos municípios brasileiros, considerando o universo dos 5.569 municípios, divulgadas no Siconfi no período compreendido entre 2015 e 2017. Os resultados encontrados mostraram que a grande maioria dos municípios brasileiros apresentou um bom nível de consistência, conseguindo atender em média 75% dos critérios analisados, demonstrando que, apesar da necessidade de ajustes, boa parte dos municípios apresenta processo efetivo de governança pública no que tange às demonstrações financeiras. A análise descritiva do nível de consistência identificou que os municípios de maior porte populacional, as capitais e os situados nas Regiões Sul e Sudeste apresentam dados mais consistentes. As inconsistências mais encontradas estão relacionadas a indicadores que avaliam os saldos das contas, a convergência às novas normas de contabilidade e a comparação entre diferentes relatórios, respectivamente. O resultado da regressão da amostra atestou que o PIB per capita (excluída arrecadação), o IDHM e o porte populacional interferem positivamente para a melhoria da qualidade das informações, mas as últimas duas variáveis se mostraram mais significantes. Contudo, o tamanho de efeito dessas variáveis é reduzido, sendo necessária a análise de outras variáveis em futuras pesquisas.
  • Item
    Previsões de resultados em partidas do campeonato brasileiro de futebol
    (2019-04-29) Santos, João Marcos Amorim dos
    Predicting football (soccer) results is a problem that has been explored for decades. The results can be seen from two points of view, predict the score or just to predict the result: win, draw or defeat. When we modeling the number of goals from each team, both points of view can be contemplated, score and result. Since 1950, many approaches have been proposed in order to model the number of goals scored by each team in a match. One of the most explored approaches considers the number of goals scored by each team as a variable following a Poisson distribution. From the first works, a underlying hypothesis was that the number of goals scored by the home team and away team was independent. However, some authors have used approaches that consider correlation in the score of the two teams, either through the use of Bivariate Poisson or the adaptation of the independent model. However, the vast majority of these works were limited to the data about the teams playing the matches and the number of goals scored and concede only. This thesis aims to explore the predictive capacity of different Poisson models proposed in the literature to predict the number of goals scored by each of the teams in a match, in addition to making use of more explanatory variables, such as number of shots, number of shots on target, tackles, all those variables coming from Cartola FC. Each one of the explored models was analyzed from the point of view to correct the true scoreboard of the game, as well as to correct the true result of the match, win, draw or defeat.
  • Item
    Investment decisions and capital accumulation: firm-level evidence from Brazil
    (2018-04-27) Camêlo, Felipe Diogo
    Using firm-level data from an administrative Brazilian dataset, I document a few stylized facts regarding capital stock accumulation patterns and investment decisions. Finding evidence largely in favor of micro-level lumpiness of investment as it was found for American firms, I document that there are a few particularities in the behavior of Brazilian firms. First, I document that the distribution of the growth rate of capital is more dispersed, with “fatter” tails. Second, I also show that, as economic activity, the volume of investment and capital stock are more concentrated on a small number of firms, micro-level lumpiness might have a bigger role in understanding aggregate movements. Third, I show that the observable characteristics of Brazilian firms explain a lot more of sudden movements in capital growth when compared to the U.S., after controlling for industry specific characteristics and other variables. Fourth and last, I compute statistical measures related to the investment rate distribution, which show that investment at the firm-level seems to be even “lumpier” in Brazil, with firms investing less on average, while experiencing more episodes of investment spikes and periods of inaction.
  • Item
    Learning in peer-to-peer markets: evidence from Airbnb
    (2016) Wu, Edson An An
    Peer-to-peer markets are highly uncertain environments due to the constant presence of shocks. As a consequence, sellers have to constantly adjust to these shocks. Dynamic Pricing is hard, especially for non-professional sellers. We study it in an accommodation rental marketplace, Airbnb. With scraped data from its website, we: 1) describe pricing patterns consistent with learning; 2) estimate a demand model and use it to simulate a dynamic pricing model. We simulate it under three scenarios: a) with learning; b) without learning; c) with full information. We have found that information is an important feature concerning rental markets. Furthermore, we have found that learning is important for hosts to improve their profits.
  • Item
    Conflito de agência no consórcio DPVAT: uma análise à luz da nova economia institucional sob a perspectiva da teoria da agência
    (2015) Duarte, Danielle Cavalcante
    Different institutional arrangements result in different incentives for performing economic exchanges. Indeed, regulatory frameworks implemented in a particular historical and economic background may produce results different from those originally intended, therefore imposing to the regulator the need for constant monitoring and intervention in order to diagnose and correct or minimize possible distortions in the relationships among actors. Thus, this dissertation is proposed to analyze the operation of the DPVAT Consortium as a connecting mechanism between its various stakeholders. The goal is to check potential conflicts of interest that may arise from the various relationships between these stakeholders – set up by the institutional framework established to the management of the premiums paid by motor vehicle owners for their mandatory motor third-party liability insurance (known as DPVAT insurance in Brazil) – that may foster regulatory intervention in order to avoid such conflicts or, at least, to mitigate them. The survey is based on identifying the expected behavior of economic agents in their own interest, according to the assumptions of the New Institutional Economics from the perspective of the Agency Theory, and on the examination of the main legislative changes in mandatory motor third-party liability insurance in Brazil over the last 50 years. Subsequently, three agency conflicts between DPVAT Consortium stakeholders were identified and analyzed based on theoretical and empirical evidence, arising from: (1) the relationship between the managing body of the DPVAT Consortium (agent) and the insurance companies that are an integral part of the Consortium (principal); (2) the relationship between the managing body of the DPVAT Consortium (agent) and the regulatory body (principal); and (3) the relationship between the insurance company that operates the DPVAT Consortium (agent) and the motor vehicle owners (principal).
  • Item
    Inattention in individual expectations
    (2015-04) Cordeiro, Yara de Almeida Campos
    This paper investigates the expectations formation process of economic agents about inflation rate. Using the Market Expectations System of Central Bank of Brazil, we perceive that agents do not update their forecasts every period and that even agents who update disagree in their predictions. We then focus on the two most popular types of inattention models that have been discussed in the recent literature: sticky-information and noisy-information models. Estimating a hybrid model we find that, although formally fitting the Brazilian data, it happens at the cost of a much higher degree of information rigidity than observed.