Gold, platinum and the predictability of bubbles in global stock markets

Abstract

This paper examines the predictability of bubbles across global stock markets and whether or not synchronicity in bubble formation can be predicted via metrics of market risk that are readily available. Utilizing the gold to platinum price ratio (LGP) as an easy to implement risk metric and the Log-Periodic Power Law Singularity (LPPLS) model to detect positive and negative bubble formation at different time scales, we document evidence of synchronized boom and bust cycles of the seven developed equity markets in the G7 bloc. More importantly, our analysis shows that bubbles and their comovements are in fact predictable although the predictive relationship is only detectible via models that account for non-linearities in the data. We find that the gold to platinum price ratio serves as a stronger predictor of deeper downward accelerating price formations followed by a rally. The predictability results for the U.S. also carries over to bubble formation in the remaining stock markets of the G7 bloc, to the extent that the gold to platinum price ratio also helps to explain the synchronicity of bubbles across the G7. Our findings provide a valuable opening for market regulators as the results show that readily available metrics of market risk can be used to model and monitor the occurrence of bubbles in financial markets as well as the connectedness of bubbles across the global markets.