CMStatistics 2017: Start Registration
View Submission - CFE
Title: The quantile-heterogeneous autoregressive model of realized volatility: New evidence from commodity markets Authors:  Konstantin Kuck - University of Hohenheim (Germany) [presenting]
Robert Maderitsch - University of Hohenheim (Germany)
Abstract: A cross-asset perspective is provided on state-dependence in the dynamics of realized volatility in the commodity futures market. Using high-frequency data for futures on Gold, Silver and Light Sweet Crude Oil, covering the period from 2007 to 2016, we estimate various Quantile-Heterogeneous Autoregressive models of daily realized volatility (Q-HAR-RV). The daily volatility is modeled as a linear function of own lags measured over different time resolutions to specifically account for the heterogeneous impact of market participants with different trading motives and investment horizons. Furthermore, using quantile regression, we are able to identify potential state-dependence and asymmetry in the short-, mid- and long-term autoregressive dynamics with respect to different volatility levels. Overall, the daily and monthly volatility aggregates appear more important compared to weekly volatility. However, we also document considerable changes in the relative importance of mid- and long-term volatility components under varying market conditions which appear remarkably similar across the three assets. Specifically, the impact of the weekly volatility increases distinctly from lower to higher quantiles of the conditional volatility distribution while that of daily and monthly volatility decreases. This implies that information generated over the medium-term gains importance in phases of increased uncertainty.