Modelo Nelson-Siegel dinâmico da estrutura a termo da taxa de juros com fatores exógenos macroeconômicos: uma aplicação ao mercado brasileiro

• Ano de defesa: 2014
• Autor(a) principal: Bernz, Bruno Müller
• Orientador(a): Ruilova Terán, Juan Carlos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Diversas aplicações; Variáveis exógenas; Métodos de espaço de estados
• Link de acesso: http://hdl.handle.net/10438/12023

The traditional representation of the term structure of interest rates in three latent factors (level, slope and curvature) had its original formulation developed by Charles R. Nelson and Andrew F. Siegel in 1987. Since then, several applications have been developed by academics and practitioners based on this class of models, mainly with the intention of anticipating movements in yield curves. At the same time, recently published papers as Diebold, Piazzesi and Rudebusch (2010), Diebold, Rudebusch and Aruoba (2006), Pooter, Ravazallo and van Dijk (2010) and Li, Niu and Zeng (2012) suggest that incorporating macroeconomic information to interest rates models can provide higher predictive power. In this study, the dynamic version of the Nelson-Siegel model, as proposed by Diebold and Li (2006), was compared to a similar model in which exogenous macroeconomic variables are included, for Brazilian data. In parallel, two different methods of parameter estimation were tested: the traditional two-step approach, and another, with the use of an Extended Kalman Filter, which allows parameters to be estimated recursively, every time a new information is added to the system. Regarding the models tested, the results were shown to be inconclusive, indicating only a marginal improvement in the estimates both in-sample and out-of-sample when exogenous variables are included. Nonetheless, the use of the Extended Kalman Filter showed more consistent results compared to the two-step method for almost all horizons studied. Keywords: Nelson-Siegel model, Term-Strucutre of Interest Rates, Macro-Finance, StateSpace Models, Brazil.