Title: A self-exciting hurdle model for extreme returns in financial markets
Authors: Katarzyna Bien-Barkowska - Warsaw School of Economics (Poland) [presenting]
Abstract: Forecasting the occurrence of extreme returns is at the forefront of modern financial econometrics and allows for effective management of financial risk. We propose an application of the discrete variable hurdle model for the occurrence of extreme losses in financial markets which allows us to explain apparent bursts of volatility and clustering of extreme returns. The conditional probability that the loss exceeds a large threshold on a given day is modeled dynamically exhibiting the self-exciting nature, where recently observed days with extreme returns increase the likelihood of experiencing further large losses. We show that the model outperforms the standard peak-over-threshold methods for forecasting value at risk and expected shortfall.