Amanhã (14/09) irei apresentar o artigo
- The Multivariate Stochastic Volatility-Double Jump Model: An Application for Oil Assets, realizado em
co-autoria com o Roberto Mauad
(Bacen) e Fernando Lucena (UERJ), no seminário de
EAESP-FGV. O resumo está abaixo.
We propose a new multivariate model to capture the presence of jumps in mean and conditional variance in the returns of oil prices and companies in this sector. The model is based on the presence of common factors associated with jumps in mean and variance, and performs a decomposition of the conditional variance of each asset as the sum of the common factor plus a specific transitory factor in a multivariate stochastic volatility structure, estimated through Bayesian methods using Markov Chain Monte Carlo.
The model allows recovering the changes in prices and volatility patterns observed in this sector, relating the jumps with the events observed in the period 2000-2015. We apply the model to estimate risk management measures, hedging and portfolio allocation and performing a comparison with other multivariate models of conditional volatility.
O seminário é das 9:00 as 10:30, na EAESP.