Estimação de modelos de estrutura a termo: aplicação para o Brasil dos modelos afim de Vasicek e CIR 2009 / Estimation of term structure models: application to Brazil of affine models Vasicek and CIR. 2009
AUTOR(ES)
Luiz Fernando de Gênova
DATA DE PUBLICAÇÃO
2009
RESUMO
In this work, we studied the Brazilian term structure of interest rate for the recent period, from January 2004 to March 2009, exploring the characteristics implied in the swap rates and ID contracts. Based on the theoretical framework of affine models, we analyze the models of Vasicek (1977) and Cox, Ingersoll and Ross (1985), two particular cases of affine models, related to a single risk factor. To estimate these models we adopted two distinct approaches, first, using the exact maximum likelihood, following the framework of Chen and Scott (1993), and second, a method with daily estimation based on criteria of error minimization. We compare the results obtained in both methodologies and models, as well as the quality of the adjustments for several measures and sub-samples. The results suggest a good performance of the two approaches, however, the use of the method of daily calibration, particularly for the model of Vasicek, was superior and capable to better adjust the dynamic of the structure term of interest rate
ASSUNTO(S)
estrutura a termo de taxa de juros maximum likelihood affine models cir calibration máxima verossimilhança structure term of interest rate calibração vasicek vasicek economia cir modelos afim
ACESSO AO ARTIGO
http://tede.ibmecsp.edu.br/tde_busca/arquivo.php?codArquivo=149Documentos Relacionados
- Estimação da oferta de trabalho com modelos coletivos: uma aplicação para o Brasil
- Comparison models of estimation of the term structure of interest rates: an exploratory study of the Brazilian market
- Fitness estimation models applied to genetic algorithms
- Modelos estatísticos de segmentação da estrutura a Termo: testes empíricos de ajustes e previsões
- UMA ANÁLISE EMPÍRICA PARA A ESTRUTURA A TERMO DA TAXA DE JUROS BRASILEIRA: USANDO O ALGORITMO DO FILTRO DE KALMAN PARA ESTIMAR OS MODELOS DE VASICEK E COX, INGERSOLL E ROSS