Title: QML estimation of a stochastic volatility with leverage and size effect model
Authors: Paolo Chirico - University of Eastern Piedmont (Italy) [presenting]
Abstract: If the use of the Kalman filter (KF) for simple, univariate and multivariate, stochastic volatility (SV) models is well known, the use of the filter for asymmetric SV model is not simple. As known, the auxiliary model used with the KF, the linear structural model of the log-squared returns, does not allow to consider the correlation between the current return innovation and the innovation on the next return volatility (leverage effect). On the other hand, the disturbances of the auxiliary model are correlated because of the presence of the size effect, the effect on the return volatility due to the magnitude of the previous return. Taking into account that correlation, a SV with leverage and size effect (SV-SLE) model is performed using the KF. The applications of the model to some financial time series show interesting results: in some cases (series), the volatility innovation is not significant so the SV-LSE model is very similar to the EGARCH model; in other cases, the SV-LSE model fits the series better than the EGARCH model.