CMStatistics 2018: Start Registration
View Submission - CFE
A1124
Title: Estimating threshold stochastic volatility models using integrated nested Laplace approximations Authors:  J Miguel Marin - University Carlos III (Spain)
Haavard Rue - KAUST (Saudi Arabia)
Helena Veiga - BRU-IUL (Instituto Universitario de Lisboa) (Portugal)
Patricia de Zea Bermudez - FCiencias.ID (Portugal) [presenting]
Abstract: Volatility of financial time series is well captured by stochastic volatility models. These models have been extended to be able to represent the asymmetric response of volatility to negative and positive shocks of the same magnitude. In the literature, there are two well-known extensions. One of these extensions is the threshold stochastic volatility (TSV) model which considers the asymmetric response of the volatility by allowing the parameters of the log-volatility equation to be different depending on the sign of lagged returns. The aim is to apply integrated nested Laplace approximations (INLA) in this framework. INLA is very fast and efficient. It replaces MCMC simulations with accurate deterministic approximations. Proper although not very informative priors are used, as well as PC priors. Finally, an application will be presented using a financial data set.