Stock, foreign exchange and commodity markets linkages: Implications for risk diversification and portfolio management

We analyze connectedness for a system composed of 111 financial markets from January 3, 2011, to December 29, 2023. Stock, foreign exchange, and commodity markets are included in the sample. Using a two-stage approach based on Principal Component Analysis to remove common global factors affecting financial market returns, we employ a LASSO-VAR model to estimate the global network of financial markets. Our results reveal that financial markets are closely linked. Common global factors intensify spillovers between financial markets. After being removed, financial markets transmit significant idiosyncratic shocks that are not explained by systemic variations. Our results also allow us to accurately identify the markets that are idiosyncratically less vulnerable to liquidity shocks, and those that are most relevant transmitting this kind of disturbances. These findings are relevant for investment decisions, risk management, and financial regulators.