Estimation of Stock Market Crashes on the Basis of Multivariate Skew-Student Density

• Activity Number: 568
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): Lei Wu*+ and Qingbin Meng and Julio Velazquez
• Companies: Nankai University and Renmin University and Delft University of Technology
• Keywords: Multivariate skew-Student density ; TVC-GARCH model ; Stock market crashes ; Crash probability forecast

Abstract

By combining the multivariate skew-Student density with a time-varying correlation GARCH (TVC-GARCH) model, this paper investigates the spread of crashes in the regional stock markets in both bilateral and global environments. This empirical framework provides a more accurate specification for empirical return distributions and helps us to detect the dynamic properties of crashes between regions. By computing the conditional 1-day crash probabilities and evaluating the forecast performance of the TVC-GARCH model with different densities, it is concluded that the multivariate skew-Student density can explain substantial dependence in the occurrence of crashes regions and predict global crashes with better predictive accuracy only depending on the past information set. This empirical framework also gives insight into the further research about contagion and can be used to improve the early-warning systems.