This paper explores the relationships between the US dollar, crude oil, gold, and bitcoin by taking into account the higher-moment linkages. Specifically, we construct robust estimators for the realized volatility, realized skewness, realized kurtosis, and jump, and study the causalities between the estimators through the Granger causality test. A generalized impulse response analysis identified by our quad-variate VAR specification is further implemented to uncover the lead-lag spillover effect across the variables of interest. We utilize high-frequency data for the chosen assets from January 3, 2016, to June 23, 2022, and observe various patterns of cross-market interconnection related to higher-order moments. These findings suggest that systematic risk factors must be considered while jointly modeling market linkages. Practical implications for investors and market regulators are also discussed.
- higher distribution moments
- market linkage