Title: Data cloning estimation for asymmetric stochastic volatility models
Authors: J Miguel Marin - University Carlos III (Spain) [presenting]
Patricia de Zea Bermudez - FCiencias.ID (Portugal)
Helena Veiga - BRU-IUL (Instituto Universitario de Lisboa) (Portugal)
Abstract: The focus is on applying data cloning to the estimation of asymmetric stochastic volatility models with flexible distributions that are able to capture the leptokurtosis and skewness of the distribution of standardized returns. Data cloning is a general technique to compute maximum likelihood estimates, along with their asymptotic variances, by means of the computation of the posterior distributions by using a MCMC methodology. Using an intensive simulation study and high frequency data for two financial time series of returns, the asymmetric stochastic volatility models are estimated and evaluated using the new proposal and a benchmark. Its performance is compared in terms of parameters' estimation, volatility and out-of-sample forecasts with a well-known Bayesian procedure. The results point out gains in efficiency and accuracy of the new estimators of parameters and volatility.