Previsão da estrutura a termo de taxa de juros aplicando o Filtro de Kalman ao modelo Vasicek: o caso brasileiro

• Autor(a) principal: Hiroki, Marcelo
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional do FGV (FGV Repositório Digital)
• Texto Completo: http://hdl.handle.net/10438/11993
• topic: Curva de juros; Modelo Vasicek; Filtro de Kalman; Vetor autorregressivo; Economia; Investimentos - Administração; Taxas de juros - Brasil; Taxas de juros - Modelos matemáticos

This work aims to test the quality of forecasting of the two factor Vasicek Model coupled with the Kalman filter. Applied to an investment strategy, it runs with a Stop Loss criteria for periods in which the model does not fit the interest rates. Using BMF’s DI future contracts available for the period of 1st of March, 2007 to the 30th of May 2014, we simulated the strategy for different periods with different conditions, aiming on finding the optimal time series window that will lead to the model parameters that best fit the current term structure, also, it will be analyzed how long these parameters optimally estimate interest rate dynamic. The results were compared with the ones obtained from the vector autoregressive model of lag 1, indicating that Vasicek Model underperform compared to this model. The limitation of the two factor model on capturing the term structure dynamics does not allow the model to estimate the term structure at once denigrating the results.