1: \begin{abstract} 

2: The  stochastic volatility   model  is a popular tool for modeling the volatility of assets. 

3: The model is  a nonlinear and non-Gaussian state space model, and consequently is difficult to fit. 

4: Many approaches, both classical and Bayesian, have been developed that rely on numerically

5: intensive techniques such as quasi-maximum likelihood estimation and Markov chain Monte Carlo (MCMC).

6: Convergence and mixing  problems still plague MCMC algorithms when drawing samples

7: sequentially from the posterior distributions.

8: While particle Gibbs methods have been successful  when applied to nonlinear or non-Gaussian

9: state space models in general, slow convergence still haunts the technique when applied specifically

10: to stochastic volatility models. 

11: We present an approach  that couples particle Gibbs with ancestral sampling

12: and joint parameter sampling

13: that ameliorates the slow convergence and mixing problems when fitting both univariate and multivariate

14:  stochastic volatility models.  We demonstrate the enhanced method on various numerical examples.

15: \end{abstract}

16: