An application of Bayesian vector heterogeneous autoregressions to study network interlinkages of the crude oil and gold, stock, and cryptocurrency markets during the COVID-19 outbreak

Abstract

We investigate fat tails and network interconnections of crude oil, gold, stock, and cryptocurrency using seven Bayesian vector heterogeneous autoregression fashions. In this paper, we incorporate parameter uncertainty by using Bayesian VAR models for estimation. To make rational investment decisions, we decompose a network of financial assets and commodity prices into various time horizons to obtain essential insight and knowledge. During the short, medium, and long run, this paper differentiates dynamically between network interlinkages between these markets. We found some noteworthy results in our study. In the first place, network interlinkages exhibit remarkable differences over time. Interlinkages between networks are increased in the short term, medium term, and long term due to transient events occurring in markets during the study period. As a result of the ongoing COVID-19 epidemic, the long-term ties within the system are significantly impacted. Additionally, based on net directional linkages, each market’s role shifts (from sending to receiving shock and vice versa) before the pre-COVID-19 pandemic course, whereas they remain persistent during COVID-19. Observations of short- and medium-term trends reveal that three markets, namely, crude oil, gold, and stock, receive shocks, which are transmitted to these markets by the cryptocurrency market. In terms of long-horizon measures, the results indicate that the gold and cryptocurrency markets persist as shock transmitters. Our findings are critical since policymakers can also design appropriate policies to reduce the vulnerabilities of such markets and prevent risk spread and instability.