The short- and long-run cyclical variation of the cross-asset nexus: Mixed-frequency evidence on financial and ‘financialised’ assets

Abstract

We study the dynamic interdependence between stocks, a risky and financial ‘by definition’ asset class, and the ‘financialised’ assets from the real estate and commodity markets. We first introduce a new multivariate corrected Dynamic Conditional Correlations Mixed-Data Sampling (cDCC-MIDAS) model through which we analyse short- and long-run time-varying correlation dynamics among stocks, real estate, and five commodity types with direct implications for risk management and portfolio optimisation. The correlation analysis identifies short- and long-run hedging properties and interdependence types and concludes on strong countercyclical cross-asset interlinkages, highly dependent on the state of the economy in most cases (contagion effects) and weak procyclical connectedness for certain safe-haven assets (flight-to-quality). We further investigate the macro-relevance and crisis-vulnerability of the correlations’ evolution by unveiling the macro-determinants of asset co-movements. The economic environment plays a key role as a contagion or flight-to-quality transmitter, outweighing the effects of economic linkages among assets, while the uncertainty channel intensifies the macro impact on the cross-asset nexus.