Efficient Bayesian Inference for Stochastic Volatility Models

Aktivität: VortragWissenschaftlicher Vortrag (Science-to-Science)


This talk considers Bayesian inference for stochastic volatility (SV) models using efficient MCMC inference. Our method is based on the popular approximation of the log $\chi^2$-distribuion by a mixture of 10 normal distribution which allows to sample the latent volatilities simultaneously, however, we introduce several improvements. First, rather than using standard forward-filtering-backward sampling to draw the volatilities, we apply a sparse Cholesky factor algorithm to the high-dimensional joint density of all volatilities. This reduces computing time considerably because it allows joint sampling without running a filter. Second, we consider various reparameterizations of the augmented SV model. Under the standard parametrization, augmented MCMC estimation turns out to be inefficient, in particular for the volatility of volatility parameter in the latent state equation. By considering a non-centered version of the SV model, this parameter is moved to the observation equation. Using MCMC estimation for this transformed model reduces the inefficiency factor in particular for volatility of volatility parameter considerably.
Zeitraum10 Dez. 201012 Dez. 2010
Ereignistitel4th CSDA International Conference on Computational and Financial Econometrics (CFE'10)
VeranstaltungstypKeine Angaben

Österreichische Systematik der Wissenschaftszweige (ÖFOS)

  • 101026 Zeitreihenanalyse
  • 502025 Ökonometrie
  • 101018 Statistik
  • 102022 Softwareentwicklung