ABSTRACT
In this article, we implement a recently developed statistical method to test asymmetries in cross-asset correlations, focusing in particular on the gold market. Our empirical results provide evidence that gold exhibits asymmetric correlations with stocks and the U.S. dollar, but not with bonds. Furthermore, splitting the sample into three characteristic periods, we find that exceedance correlations exhibit substantial time variation even in similar market tensions for same pairs of assets. Our findings imply that investors and fund managers should take into account the asymmetric dependence structure, which depends on the upside or downside of the market.
Acknowledgements
We are grateful to two anonymous referees for their helpful comments and suggestions.
Disclosure statement
No potential conflict of interest was reported by the authors.
Supplementary Material
Supplemental data for this article can be accessed here.
Notes
1 See also Ang and Bekaert (Citation2002).
2 Miyazaki, Toyoshima, and Hamori (Citation2012) apply the A-DCC model to returns of gold and traditional assets (stock, bond, and foreign exchange).
3 The description of the methodology in this section is fairly rough. Refer to Hong, Tu, and Zhou (Citation2007) for a complete discussion of theoretical developments.
4 The Bartlett kernel function is defined as .
5 SPX is from Datastream. CMBI and USDI are from the Federal Reserve Economic Database of the St. Louis FED.
6 This data set is available at the link provided in the supplementary data set section of this paper.