Jian Chen (ICMA)– “Modelling Price and Volatility Jump Clustering by Marked Hawkes Process” – PhD Seminar

Abstract

This paper studies clustering behaviours of price and volatility jumps using high-frequency data, modelled using a Marked Hawkes Process embedded in a bivariate jump-diffusion model. Under de-periodisation, we find evidence showing self-excitation behaviours of jumps in both individual stocks and an index. Also, considering positive, negative price jumps and volatility jumps, the impact that an occurrence of a jump in one dimension has on that in another dimension is shown to be asymmetry. More importantly, the extent of this impact is shown empirically to be positively correlated with jump size. We also formalise the self-excitement and self-freeze properties of durations between two jumps. More self-freeze behaviours have been found in empirical studies. We estimate model parameters using Bayesian inference by Markov Chains Monte Carlo.