Abstract: This paper proposes a new dynamic model of realized covariance (RCOV) matrices based on recent work in time-varying Wishart distributions.
The specifications can be linked to returns for a joint multivariate model of returns and covariance dynamics that is both easy to estimate and forecast. Realized covariance matrices are constructed for 5 stocks using high-frequency intraday prices based on positive semi-definite realized kernel estimates. We extend the model to capture the strong persistence properties in RCOV. Out-of-sample performance based on statistical and economic metrics show the
importance of this. We discuss which features of the model are necessary to provide improvements over a traditional multivariate GARCH model that only uses daily returns.
Keywords: eigenvalues, dynamic conditional correlation, predictive likelihoods, MCMC
JEL Classification: C11; C32; C53; G17