A Bayesian Approach to Heterocedastic Models of Time Series and its Application in the Volatility of Financial Assets

• Autores: Flores Montoya, Edwin Antero, Bravo Quiroz, Antonio
• Formato: artículo
• Fecha de Publicación: 2021
• Institución: Universidad Nacional Mayor de San Marcos
• Repositorio: Revistas - Universidad Nacional Mayor de San Marcos
• Lenguaje: español
• OAI Identifier: oai:ojs.csi.unmsm:article/21152
• Enlace del recurso: https://revistasinvestigacion.unmsm.edu.pe/index.php/matema/article/view/21152
• Nivel de acceso: acceso abierto
• Materia: DCC - GARCH heteroscedastic models; MCMC methodology; Modelos heterocedasticos DCC - GARCH; metodología MCMC

In this work, the modeling of the volatility of financial assets is studied using a Bayesian approach. DCC - GARCH models are used, for the errors of these models asymmetric and leptokurtic probability distributions are considered, which are parameterized according to the asymmetry and the weight of the tails, therefore these parameters are also estimated. The estimation of the model parameters was performed using the MCMC methodology Metropolis - Hastings random walk algorithm using the software R package bayesDccGarch, daily data from 04/01/2015 - 01/31/2020 of the stock indices of: Frankfurt are considered (DAX), Tokyo (NIKKEI225), Paris (CAC40), and Lima (BVL). The Bayesian approach to estimating the model parameters facilitates interpretation and provides the ability to insert a priori information for the parameters.