Title: Jointly modeling autoregressive conditional mean and variance of positive valued time series
Authors: Hiroyuki Kawakatsu - Dublin City University (Ireland) [presenting]
Abstract: Observation driven models with conditional mean and variance dynamics are proposed for positive valued time series. The motivation is to relax the strong restriction on the higher order moment dynamics implied by the standard mutiplicative error model that is driven only by the conditional mean dynamics. The empirical fit of the proposed specifications is assessed with daily realized volatility series for a number of stock indices using both in-sample estimates and pseudo out-of-sample prediction densities.