Determinantes da estrutura de capital e teoria do trade-off: aplicação de modelos para dados em painel, regressão quantílica e regressão quantílica penalizada com efeitos fixos

• Ano de defesa: 2019
• Autor(a) principal: Guimarães, Paulo Roberto Fonteles
• Orientador(a): Carhuajulca, Jaime José Orrillo
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Católica de Brasília
• Programa de Pós-Graduação: Programa Stricto Sensu em Economia de Empresas
• Departamento: Escola de Gestão e Negócios
• País: Brasil
• Palavras-chave em Português: Dados em painel; Estrutura de capital; Teoria do Trade-Off
• Palavras-chave em Inglês: Trade-Off Theory; Capital structure; Panel data
• Área do conhecimento CNPq: CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA
• Resumo em Inglês: The present work has the main objective of evaluating the main conclusions and results derived from the capital structure trade-off theory, using panel data of 144 brazilian firms in the period ranging between 2013 and 2018. Among all the possible theoretical frameworks that explain firms’ financing decisions and capital structure choices, it was decided to adopt the trade-off framework as there is quite a lot of empirical evidence about its validity. Besides, there is a variety of mathematical theoretical models, both in a statical and in a dynamic view, that build up the foundations of this theory, hence providing some testable results. Based on the empirical data, the main panel data estimators were applied, i.e., fixed effects, random effects, pooled ordinary least squares and first difference estimators, which were compared among themselves with the use of statistical tests and fit measures. In a second moment, with the intention of identifying the behaviour of the relationships among variables under different levels of tax benefits and bankruptcy costs, which are the main elements of the trade-off theory, some quantile regressions were estimated, including the penalized quantile regression with fixed effects model proposed by Koenker (2004). This model allows the estimation of conditional quantile functions, controlling the non-observed heterogeneity in the data by using the shrinkage technique. This last model is considered to be innovative to this kind of work, as it was not found any other study that applied this method to model firms’ capital structure. After estimating all the regressions, empirical evidence in favor of the trade-off theory was found. Finally, as a secondary objective, a broad literary review about the trade-off theory of capital structure was built, ranging since the earliest studies until the more modern ones created nowadays.
• Link de acesso: https://bdtd.ucb.br:8443/jspui/handle/tede/2578

The present work has the main objective of evaluating the main conclusions and results derived from the capital structure trade-off theory, using panel data of 144 brazilian firms in the period ranging between 2013 and 2018. Among all the possible theoretical frameworks that explain firms’ financing decisions and capital structure choices, it was decided to adopt the trade-off framework as there is quite a lot of empirical evidence about its validity. Besides, there is a variety of mathematical theoretical models, both in a statical and in a dynamic view, that build up the foundations of this theory, hence providing some testable results. Based on the empirical data, the main panel data estimators were applied, i.e., fixed effects, random effects, pooled ordinary least squares and first difference estimators, which were compared among themselves with the use of statistical tests and fit measures. In a second moment, with the intention of identifying the behaviour of the relationships among variables under different levels of tax benefits and bankruptcy costs, which are the main elements of the trade-off theory, some quantile regressions were estimated, including the penalized quantile regression with fixed effects model proposed by Koenker (2004). This model allows the estimation of conditional quantile functions, controlling the non-observed heterogeneity in the data by using the shrinkage technique. This last model is considered to be innovative to this kind of work, as it was not found any other study that applied this method to model firms’ capital structure. After estimating all the regressions, empirical evidence in favor of the trade-off theory was found. Finally, as a secondary objective, a broad literary review about the trade-off theory of capital structure was built, ranging since the earliest studies until the more modern ones created nowadays.