CMStatistics 2016: Start Registration
View Submission - CFE
Title: Stylized facts for extended HEAVY models: The importance of asymmetries, power transformations and long memory Authors:  Menelaos Karanasos - Brunel University (United Kingdom) [presenting]
Starvoula Yfanti - Queen Mary University of London (United Kingdom)
Abstract: The High frequency bAsed VolatilitY (HEAVY) model is studied and extended. Our main contribution is the enrichment of the model with asymmetries, power transformations and long memory (fractionally integrated or hyperbolic) through the hyperbolic (double) asymmetric power (HYDAP) formulation. The conclusion that the lagged realized measure does all the work at moving around the conditional variance of stock returns while it holds in the benchmark specification it does not hold once we allow for asymmetric, power and long memory effects. Our main findings are as follows. First, the power transformed conditional means (of the squared returns and the realized measure) are significantly affected by both the lagged power transformed realized measure and absolute negative returns. Second, fractional integration fits the HEAVY model of the returns better, whereas hyperbolic long memory is more suitable for modelling the conditional mean of therealized measure. Third, the overnight trading activity indicator affects positively the daily volatility and lowers its persistence, whereas it has a trivial impact on the intra-daily conditional variance. Fourth, the HEAVY framework applied to the Garman Klass (GK) volatility gives results very similar to the ones from the stock returns.