Improving the Asymmetric Stochastic Volatility Model with Ex-Post Volatility : The Identification of the Return-Volatility Correlation

Simulation studies prove that the asymmetry stochastic volatility (ASV) models infer erroneous return-volatility correlation, due to the unobservability of the true volatility. We propose to incorporate the ex-post volatility in the ASV framework for the identification of the correlation coefficient. Comparing to the standard ASV model, we obtain a significantly smaller magnitude of the estimated return-volatility correlation on major U.S. equity market indexes. Out-of-sample return density forecasts demonstrate superior performance when jointly estimating the return and the ex-post volatility processes. The corrected returnvolatility correlations by estimating proposed ASV models with subsample data further document the time-varying leverage effect