Volatility Shock Causing Incidents

Abstract

The identification of the latent data generation process (DGP) of each series of volatility estimates, also determines the forecasting object, which in this thesis is the changing of the parameters of the Markov switching mixture model discussed in Section 4.3.2. Note, that the task is not to forecast the statistical control parameters, but to identify the latent process explaining the estimated volatility and to test for structural breaks in the identified process that can be explained with the appearance of unexpected news.