Abstract
In this article we investigate a multi-factor version of the Heston
(1993) stochastic volatility model. First, we provide explicit expres-
sions for excess kurtosis and autocorrelation of squared returns and
show that excess kurtosis is smaller than three and squared autocor-
relations are smaller than 0.2 even for a multi-factor model. Then we
discuss a fully Bayesian analysis based on Markov chain Monte Carlo
(MCMC) estimation and data augmentation and improve the perfor-
mance of MCMC estimation by using a partially centered parametriza-
tion of the model. Finally, we apply the multi-factor Heston stochastic
volatility model to simulated as well as to exchange rate data.
(1993) stochastic volatility model. First, we provide explicit expres-
sions for excess kurtosis and autocorrelation of squared returns and
show that excess kurtosis is smaller than three and squared autocor-
relations are smaller than 0.2 even for a multi-factor model. Then we
discuss a fully Bayesian analysis based on Markov chain Monte Carlo
(MCMC) estimation and data augmentation and improve the perfor-
mance of MCMC estimation by using a partially centered parametriza-
tion of the model. Finally, we apply the multi-factor Heston stochastic
volatility model to simulated as well as to exchange rate data.
| Original language | English |
|---|---|
| Pages (from-to) | 5 - 25 |
| Journal | Communications in Dependability and Quality Management |
| Volume | 11 |
| Publication status | Published - 1 Oct 2008 |