Tracing contagion between bitcoin and traditional markets

Abstract

Purpose We investigate the presence of contagion between Bitcoin and four traditional assets (stocks, bonds, gold and the US dollar exchange rate) over the period 2015–2024. Design/methodology/approach We implement a framework that combines the DCC-GARCH specification and a time-varying causal inference methodology. Findings Our findings support that Bitcoin remains weakly connected to the global financial markets. Contagion is limited, appearing sporadically from S&P 500 to Bitcoin and from Bitcoin to the US dollar index. However, when we impose a stricter definition of extreme correlation or a multivariate VAR specification, the contagion results vanish, indicating no systematic contagion between Bitcoin and traditional assets. Practical implications Our evidence implies that Bitcoin may be used as a useful portfolio diversification instrument. Originality/value We deploy a recently developed novel methodology that combines the DCC-GARCH model and a recent time-varying Granger causality procedure to distinguish between extreme high correlation and contagion and find no evidence of systematic contagion effects of Bitcoin with conventional asset classes.