Quantile dependencies between discontinuities and time-varying rare disaster risks

Abstract

We study the role of rare disaster risks in discontinuities (jumps) in the US equity market. To this end, we use data from Dow Jones Industrial Average and International Crisis Behavior database (as a proxy for rare disaster risks) over the period January 1918–December 2013. We apply a quantile dependence approach in order to detect directional predictability from rare disaster risks to various types of jumps, realized skewness and realized kurtosis risks at different quantiles and lags. We find an asymmetric relationship between jumps and rare disaster risks, as we report a heterogenous dependency across different quantiles and lag orders. Although rare disaster risks can significantly help in the predictability of jumps, large jumps due to large price movements happened in the market do not associate with rare disasters.