Dynamics of co-bubble networks across commodity futures prices and portfolio performance

This paper introduces a price co-bubble network model then investigates the transmission of co-bubbles across 36 commodity futures prices, enabling a comprehensive understanding of the dynamic changes in co-bubble dependencies at both the overall and sector levels and during crisis periods. The analysis highlights several key results. Firstly, commodity futures within the Energy group demonstrate a higher propensity for experiencing co-bubbles compared to the Metals, Agricultural, and Grain and Cereal groups. Secondly, co-bubble occurrence is heterogeneous among pairs of commodity futures prices within different groups. Thirdly, the influence of co-bubble centrality is time-varying, reflecting its sensitivity to crisis periods. During the COVID-19 pandemic, Metals, particularly Gold and Palladium, exhibit a gradual improvement in their co-bubble network centrality ranking. During the Russo-Ukrainian war, Carbon Emissions maintain a dominant position in the centrality ranking. Fourthly, a portfolio analysis demonstrates that commodity portfolios constructed based on the results of the proposed co-bubble network outperform the baseline strategy, yielding higher Sharpe ratios and cumulative returns. These results extend our understanding on the dynamics of co-bubble dependencies in the commodity markets, and highlight the potential benefits of incorporating co-bubble dynamics into the portfolio construction.