Simultaneity of Tail Events for Dynamic Conditional Distributions of Stock Market Index Returns

The tail events represent a phenomenon long studied in the literature of stock market returns. The dynamical properties of conditional distributions are currently analyzed by means of the first four moments via Gram-Charlier likelihood functions. We propose an analysis of changes in the values of means, volatilities, skewness and kurtosis coefficients for a series of intra-daily frequency of 14 stock market returns to develop a jump detection mechanism based on the estimation of a dynamic threshold that relies on the first four moments of the distribution. Our main objective consists in the estimation of simultaneity of tail values for these moments. We consider the 5% up and 5% down event as jumps in the series of these coefficients and we compare their realizations across the series of different stock markets for simultaneity. Finally we propose an indicator that can show the degree of co-movements in the extreme values of these coefficients for different frequencies.